“As soon as a candidate who wins in the primaries loses the race for the presidency of the Republic, primaries may very well cease to be one of their party’s priorities”.
1The open primary system is associated in France with an unappealing “Americanization” of political life, and importing it has long been seen as unthinkable. Primaries were thought to be contrary to French political culture and the “spirit” of the Fifth Republic’s institutions, and against the interests of the political parties, which have typically exerted control over the nomination of candidates. Over a few weeks in 2016 and 2017, however, both Les Républicains (LR) and the Parti Socialiste (PS) chose their candidate for the presidential election using open primaries. Such a selection method was meant to reaffirm the parties’ importance on the electoral playing field, to offer an undisputed candidate, and to unite their camps. But, in François Hollande’s words, “nothing went as planned”.  In a strikingly unpredictable campaign, the primaries defied all expectations. They had been used to nominate Hollande in 2011; now, they were a crucial factor in his decision not to stand again. On both right and left, the candidates chosen seemed more “radicalized” than their party lines and, once chosen, they struggled to broaden the electorate beyond those who had been mobilized during the primary phase. When challenged, François Fillon used the legitimacy of the primary to block the emergence of an alternative candidate. Candidates who did not go through such filters—Emmanuel Macron, Marine Le Pen, and Jean-Luc Mélenchon—outperformed those from the two “primary” parties.
2Primaries had been presented in 2011 as “a road to the modernization of French democracy”.  But the prevailing image of them was now inverted. They were meant to combine reform and efficiency, to peacefully resolve leadership questions, to produce unity—and so were meant to be, as Arnaud Montebourg said, “a weapon of mass construction”.  But instead they handicapped the candidates who won them. Worse, they proved to be a difficult-to-master source of instability. They were introduced to stave off division and dissidence,  but failed to stop platforms from fragmenting, or individual candidacies emerging outside the party. They were designed to make the organization of the contest more stable, but their organizers lost control of them, demonstrating the limits of a hybrid—partisan and electoral—voting system.
3Would the primaries be a first round “that didn’t count?”  Would the winning formula from 2011 become an engine of defeat in 2017? Is there a danger of attributing too much meaning to an electoral process whose value was determined on the basis of Hollande’s victory in 2012, and whose “effects” are analyzed mechanistically? By inventing this new selection method, the Parti Socialiste elevated the 2012 election beyond its expected “normality”,  and redefined how it was organized. By outsourcing the candidate selection process to its supporters, it brought about a paradoxical rehabilitation of the party and its resources, which were now viewed as something to be fought over (a monopoly on party approval remains the stake of the primary). This was especially true because the main candidates went on to become ministers. Christophe Béchu (LR), the mayor of Angers, described “a very unfortunate precedent, in which 100% of those competing won: the winner became president, and all the losers became ministers at some point or other”. 
4But we must not judge primaries solely by the aims assigned to them, or by their effects. That would be too teleological by far. This article examines what led the Union for a Popular Movement (UMP) (who became Les Républicans) and the PS to open up their selection process, and so lose control of candidate selection. We show here that the primaries of the PS and LR obeyed multiple rationalities. They were “the” solution to multiple problems.  Structurally, open primaries were a result of weakening political parties and shrinking activist bases, both of which meant closed internal primaries were less valuable.  Institutionally, they were a result of the parties adapting to presidentialization. In each case, they were triggered by party “crises” which arose during periods of defeat and leadership vacuum. (For the PS, this was the Reims congress, and for the UMP it was the clash between Copé and Fillon.)  But we show that, on both right and left, open primaries were above all the result of something more circumstantial: Hollande’s double electoral victory in 2011–12, which became the founding myth of the primaries, and its spillover and contagion effects. This selection process—which entered the French party system “by force” —was adapted for the organizational culture of the UMP (later LR) because of a belief in its electoral effectiveness (which had been demonstrated in 2011), and by mimesis.  Open primaries were the object of intense conflict within the party from 2012 onwards, and for a long time it was unclear that they would be adopted, until they were eventually implemented in a planned, reasoned way. The PS returned to the project of 2011, but in a highly confused, improvisatory, and “path dependent” way, in that it was decided not to automatically select the outgoing president as candidate.
5The 2011 primary served in both cases as a decisive “precedent”.  But the statutory adoption of primaries did not mean they were in fact automatically introduced. This article’s second aim is to observe the process of bricolage and the struggles which primaries gave rise to in both parties. The PS chose primaries as their statutory method of presidential candidate selection in October 2012, at the Toulouse congress. The UMP adopted them in June 2013, after a vote by its members. In both cases, however, primaries were implemented after an erratic process. They were a matter of private party law, which contained a substantial number of uncertainties.  The unpredictability of primaries is principally a result of the uses they are put to in particular circumstances, and of contextual properties and situational logics.  Primaries are still only weakly institutionalized, and therefore poorly controlled and unstable. As a technology and a procedure, they are still being mastered.
6We begin by reviewing theoretical models of open primary adoption, supplementing these discussions with a look at the “2011 Socialist precedent” set in France. We then analyze the mimetic dynamics that led the right to adopt primaries in turn, and the process of planning and implementing them in the context of a predicted victory in the presidential election. Finally, we examine the chaotic bricolage process which led the PS—a prisoner of its own “invention”—to formalize the primary of the Belle Alliance Populaire. 
The Adoption of Primaries: Structural Logics and the Weight of the 2011 Precedent
7Open primaries have become the international standard of selection in America and beyond, and there has been much discussion of their development. Such work centers on the controversial question of their effects—the “primary bonus” versus the “primary penalty”. But researchers have also examined the logics which determine their adoption. The French case allows us to enrich these models by taking interparty mimesis phenomena into account.
8Generally, open primaries first arose in response to new challenges and crises of legitimacy parties were facing. Primaries are part of a process of party “democratization”.  This initially leads parties to give members more power, and then to extend candidate nomination rights to their supporters or to voters in general. The first step towards open primaries tends to be a closed primary. (This was the case both for the PS and the UMP.)  As the “participatory imperative” grows stronger, parties use the democratic innovation represented by primaries to project an image of openness and modernity, breaking with the self-absorbed “sterility” of internal struggles. Primaries seem all the more “democratic” because, during a membership “crisis”, the position of parties within societies is substantially weakened. Given the increasingly narrow portion of society each party represents, should their candidates be chosen solely by their members? We can see primaries as the product of weakening political parties and their relegitimization strategies. The PS adopted primaries after its failed attempts to revitalize the party.  In general, primaries are a way of strengthening links between parties and voters by involving the latter in candidate selection.  The strengthening of the electoral dimension of parties—a crucial development in the cartel party model —has led them to outsource the task of selecting candidates. In this sense, primaries demonstrate the growing “electoralization” of parties, even if only the dominant parties have the resources to hold them. They allow parties to optimize candidate choice, to increase their public legitimacy and credibility, to expand (hitherto limited) voter choice, to mobilize voters prior to the election, and to reinforce their sense of participating in a common political project.  Many Anglophone studies have shown the electoral benefits which can be expected from primaries. They allow parties to select the “best” candidate for the contest, or to increase the candidate’s value and better target the average voter’s interests.  They thereby obey an “informational rationale”. Finally, primaries can be treated as an “adaptation” to the presidentialization of institutions and the increasing personalization of political life—phenomena which they exacerbate in turn. Primaries are a result of the individualization of the political field and the primacy of personalities, as gauged by polling.  In France, primaries are the final stage of (hyper-)presidentialization, reinforced by the five-year term and the inversion of the election calendar in 2002.
9But these general factors should be contextualized relative to specific configurations and circumstances. Not all political parties adopt primaries. As well as the long term of structural transformation, we must consider the short term of the moment of adoption itself. Situational elements (like crises and repeated defeats) and other factors (reformist activism, leaders’ strategies, or regulations imposed by internal struggles) often determine whether and how primaries are adopted. Primaries are introduced in situations of (emerging) crisis triggered by defeat and a subsequent leadership vacuum, an inability to respond appropriately using existing tools, and the mobilization of actors to pursue specific strategies.
10There are strong structural similarities between the situations in which the PS and the UMP adopted primaries. In both cases, defeat had produced a leadership vacuum which intensified competition for top positions. The parties became absorbed by the sort of internal struggles that existing procedures like congresses and internal votes were supposed to be able to settle, with the party leadership responsible for nominating a presidential candidate. These conflicts could not be resolved by traditional arbitration instruments, however. In this context, primaries were presented as a way of regulating internal struggles, a procedural solution that would use a codified, negotiated framework to civilize the disputes. Montebourg’s “coopetition” approach would allow the parties to move beyond the brutal primaries going on behind closed doors, and to settle differences out in the open. Candidates might clash, but they would co-operate at the same time, because the final goal was their side’s victory. 
11These factors mostly explain why the PS adopted primaries in 2009.  To a lesser extent, they explain why the UMP adopted them in 2013. But they only partly explain why they were introduced in 2016–17.
François Hollande’s Double Victory: The Founding Myth of French Primaries
12The dissemination and legitimization of open primaries owes much to the “success” of 2011, and to a dynamic of contagion and institutionalization which it brought with it. The primaries of October 2011 were, without question, a political risk for the PS. They could exacerbate and publicize partisan differences just a few months away from the presidential election; there could be logistical failures; they might not find an audience. But they were ultimately presented, unanimously, as a “democratic” success and vaunted as a crucial factor in Hollande’s victory in the 2012 presidential election.
13A few weeks after the vote, the Fondation Terra Nova wrote in its report:
The surprise was complete: none of these fears materialized, and every goal was met. Primaries have shown themselves to be a formidable element in democratic modernization, and a political asset for the left. 
15Self-promotion aside, most journalists and political commentators made the same observation—as did leaders on the right, whose admiration for the primary was unanimous. The Socialist primaries were a democratic success, a logistical success, and a media success.
16The primaries were an organizational feat that revitalized party activism for almost a year and renewed a sort of activist pride—even if it didn’t last. The media praised the good conduct, discipline, and quality of the debates, in spite of the fractious campaigning between the two rounds of voting. Even without a benchmark, voter turnout was considered very strong. The PS had carefully set itself the attainable goal of a million participants. In the end, 2,661,231 people voted in the first round, and 2,860,000 in the second. As newspapers reported:
What we witnessed was not a deadly cacophony but a polyphony, in which the differences did not put the overall harmony at risk.
It’s a political revolution. No other party can boast of having mobilized so many participants outside an official election.
19The PS successfully mobilized around 30% of its target electorate. The primaries created pre-election mobilization momentum, all the more intense because the organizing party enjoyed an undeniable media boost. The primary process saturated the media landscape, with its leaders and their platforms being heavily exposed for seven weeks. Between July and October 2011, the PS benefited from an average air time 80% higher than that of the majority.  There was widespread surprise when the first debate attracted five million viewers. The convention was broadcast live, showing the candidates rallying to Hollande and the Socialists achieving newfound unity. The PS had been focused on internal struggles since 2002, but it now presented an image of openness, of “renovation”, and of democratic modernity.
20All in all, the procedure gave Hollande the undisputed legitimacy he had lacked in 2007. Fewer than 3% of those surveyed less than a year before the primary said that they would vote for the Socialist candidate. Would Hollande have beaten Sarkozy without this first phase of the campaign? Impossible to tell, but his victory de facto validated and confirmed the usefulness of open primaries. Our central hypothesis is that the unanimous sense that the primaries had been a success helped the model to spread. The right adopted it because of electoral pragmatism, and the PS because it was difficult to challenge the precedent set in 2011.
The Right’s Pragmatic Conversion to Primaries
21Primaries go against what leaders on the right call the Bonapartist “culture of the leader”, which had until then “naturally” led them to choose the head of the party as their candidate. In June 2013, however, the UMP adopted open primaries for future presidential candidates into its statutes. After much uncertainty, this procedure was confirmed in April 2015 with the validation of a primary charter. The right’s adoption of primaries followed multiple rationales related to the structural and situational factors described above. But it seems as though the precedent set in 2011, and the strength of that example, were decisive. The 2011 primary stunned leaders on the right. They may officially have challenged its principles, but they were struck by its electoral effectiveness. (A number of them demanded that the primaries be canceled by the Constitutional Council, or referred them to the CNIL.) Many saw primaries as a major competitive advantage.
Adoption: Partisan Crises and Mimesis
22On first observation, the adoption of primaries by the right was a response to the crisis brought on by the “leadership vacuum” that followed Sarkozy’s defeat in 2012, and by the inability of internal procedures to deal with this. Primaries were first adopted as a result of a compromise to resolve the crisis caused by the clash between Jean-François Copé and François Fillon over the leadership of the party in November 2012, and by Copé’s controversial election. Large-scale fraud during internal voting discredited the principle that presidential candidates should be selected internally. (Similarly, irregularities in internal voting during the Reims congress had led to the adoption of primaries, like an admission of powerlessness on the part of the PS.) This internal vote plunged the UMP into an unprecedented crisis. The results on the evening of 18 November were particularly close, and that night the two camps both proclaimed victory within twenty minutes of each other.
23Fillon presented open primaries as the main condition for restoring peace within the party. At the outset, Fillon called insistently for a new election for the UMP presidency. He ultimately abandoned this demand, but in exchange won from Copé a promise of open primaries for presidential candidates. Fillon could thus “turn towards the French people”, wagering that control of the party machinery would become less important once primaries were introduced, and that the role and resources of the UMP’s president would be discredited to some extent. The primaries arose from a situational logic, and from a compromise which was linked to internal party struggle and aimed to get the party out of a crisis.  Primaries were adopted by statute in June 2013, with 92% of members voting for them in an online poll.
24The rise of the Front National (FN)—which many commentators labeled the “first party of France” in 2014–15—was another situational factor, creating pressure towards unity from the first round. Beginning in 2014, a new argument for primaries emerged: that the UMP risked absence from the second round of the presidential election because of an FN victory. The FN’s performance in the intervening municipal, European, departmental, and regional elections created the threat on the right of “another April 21”. The press spoke increasingly of “tripartitism”. In this context, primaries seemed like a “guarantee of unity”.
25But open primaries would not have seemed a viable solution if the method had not already been proven by the PS, and had not been presented as the winning formula for the presidential election. Primaries only became a legitimate option after Hollande’s double victory in 2011–12. The basis of the shift towards primaries was mimetic. Since 2007, the UMP and the PS had been involved in a collective logic of democratic emulation.  Occupying symmetrical positions on the political playing field, the two parties in this duopoly were locked in a competitive relationship that encouraged them to observe and imitate each other, although the effects of this “convergence” are still difficult to fully assess. 
26Bernard Accoyer, the former president of the Assemblée Nationale, was one of the first to support primaries, even before the results of the Socialist primary were known:
We are living through a development in the Fifth Republic which will eventually become established. 
28Fillon also supported primaries, seeing them as “a major political event that will affect the future of our country”, and as “a modern process, suitable for the right and the left and for every election”.  Bernard Debré, député for Paris, called on the UMP to free itself from its “culture of the leader”: “This is the direction history is taking. We’ll need open primaries for the right and the center”.  The UMP’s main leaders, however, were less attracted by the philosophy of primaries than by the belief that their effectiveness explained Hollande’s victory. Édouard Balladur made this clear in May 2013:
The primary was the reason for François Hollande’s success in 2012. Three million voters made for a powerful launchpad. 
30Maël de Calan, the head of the think tank La Boîte à Idées, which promoted this selection method on the right very early on, emphasized:
The leaders of the UMP were very struck by the second round of the Socialist primary, and particularly by how it was picked up in the media, and by the unity which the Socialists achieved and displayed […] The leadership had to face up to facts: primaries had shown themselves to be particularly powerful tools. 
32Jérôme Chartier, a close ally of Fillon, took the same point of view:
We were amazed by how successful the primaries had been, by how they stirred things up, by how electorally effective they were, and by how successful they were with the media. They’re an excellent way of managing and moving past infighting, and building unity methodically and procedurally. 
34Primaries are one option available in the marketplace of selection methods. Within the UMP, which leadership crises had made more receptive to the introduction of mimetic dynamics, they gradually established themselves as the path to take. The homology or symmetry between the positions of the UMP and the PS within the political field also supports the hypothesis of an institutional isomorphism. This has been suggested, in particular, by Paul DiMaggio and Walter Powell.  The mimesis went a long way: the rules of the game the right adopted in 2015 were very similar to those used by the PS in 2011—a physical vote, in two rounds, in the fall. The only difference was in endorsement conditions, with the Républicains choosing conditions closer to those of the general election.
Uncertainty and the Securitization of the Process
35But the adoption of primaries by the right was not entirely free from difficulties. For a long time, the process remained uncertain. Sarkozy’s “return” compromised the procedure for a spell. The ex-president followed a legitimation approach which predated that of primaries, trying to guarantee his candidacy by winning leadership of the party. According to a number of people close to him, he also wanted to challenge the entire primary process.  On the sideline of his rally in Nice on 21 October 2014, Sarkozy announced that he had changed his mind on the primary, and that he was opposed to an independent authority organizing it.  As with the PS in 2011, he wagered that the procedure would be abandoned and that “selection by polling” would take place instead, or else that there would be a “ratification primary” without any real competition. Claude Bartolone had made a very similar comment in connection with Dominique Strauss-Kahn’s candidacy. Outside Sarkozyist circles, the UMP leadership strongly believed that the presence of primaries in the statutes did not guarantee that they would be introduced, underscoring the flexible relationship between the party’s leaders and its rules. 
36A coalition to defend primaries was formed in January 2014, containing Xavier Bertrand, Bruno Le Maire, Juppé, and later Fillon. Beyond their differences and their own interests, these aspirants were collectively interested in securing the position of primaries. They viewed doing so as a way of stopping Sarkozy’s return, or of establishing a balance of power by forcing him to go through this pre-selection stage.  Juppé warned that, if there was no primary, he would still stand as a candidate.  The defenders of primaries exploited the media, which accused Sarkozy of trying to scuttle the process by neglecting its logistical organization or by challenging the very idea of it.
37In an illuminating interview, Benoist Apparu, a Juppéist député, remarked:
It’s obvious that we Juppéists are happy about this attitude in the media; we tried to influence them on this question by making them afraid, by making threats. There were lots of fantasies about the primaries not taking place which rather suited us[…] 
39Although he had been elected leader of the UMP in the first round on 29 November 2014, with 64.5% of the 155,801 votes cast, Sarkozy could not oppose the process once it was underway. The député Thierry Solère described an exchange between his ally Le Maire and Sarkozy, in the wake of the latter’s election as leader. 
This was the moment the fate of the primary was decided, was sealed. Up until then, nothing had really been set in stone. It wasn’t serious. Nobody read the first primary charter–not even me. It was patched together, there was nothing interesting about it. So Sarkozy met with Le Maire, who was polling at 30%, and said: “What do you want? Do you want to be secretary general? What do you want for those close to you?” He said he didn’t want anything, but he gave him a warning. He said that the right wing would explode if there wasn’t a primary. So he asked Sarkozy to structure the primary around a working group which would be completely independent from the party in order to safeguard it. Sarkozy said no. He didn’t want that. He said: “Right, and you don’t want my job, either”. He said: “We’ll have primaries, but on my terms”. But the next day he accepted. What happened? I think he said to himself, there’s so much mistrust that I don’t have a choice, the party might explode. And suddenly he decided, I’m going to be the one to pacify it, and his tone changed to: “Love me, everyone”. He had to have a primary. Maybe Sarkozy regrets primaries, but after that he couldn’t oppose them. 
41A primary organization committee was set up, chaired by Solère. A new primary charter was produced, and an organizational architecture was defined which was more formal than the one the PS had established. On “legal” grounds, the rules of the primary were given their own text, distinct from the statutes.  Unlike in the PS, organization and control of the primary was outsourced to two entirely autonomous entities: the Comité d’Organisation de la Primaire (Primary Organization Committee), headed by Solère, and the Haute Autorité de la Primaire (High Authority of the Primary), chaired by Anne Levade, a lawyer. The second of these was “strictly independent of all political parties”, and oversaw the planning of the primary. This marked a new stage in the development of primaries since 2011 into extra-partisan affairs. It was precisely because the primary did not just concern LR that it was established as an independent procedure. The Haute Autorité, which had its own budget, was seen as a guarantor of the impartiality and depoliticization of the process.
42Levade returned to this question in an interview:
Compared to the PS, the new thing we offered was a strongly independent Haute Autorité and a less internal method of organization. There had always been the idea of an independent organ. As a lawyer I played a role in the matter, but not necessarily a decisive one. We organized the primary for the parties involved.  The Haute Autorité had the premises, an accountant, substantial resources of its own, and we were able to negotiate a payment of five million euros. 
44Two other factors explain the emphatically formal nature of the process: the desire to avoid any misconduct and fraud, given the climate of general mistrust in the party, and a confident expectation that the right would win the presidential election. In meetings on the primaries held in October 2016 in the northern federation, Gérald Darmanin, the deputy mayor of Tourcoing, spoke of the winning candidate in the primaries as the next president of the Republic.
45Apparu, a close associate of Juppé, extended this line of thought in an interview:
The party was seen as rotten to its core, completely discredited. Confidence was very low, both inside and outside. In one move, we outsourced things to people who don’t know anything about this, and to external processes. We felt that we came out on top with organizing the primary, because we achieved a number of things during the process which were irreversible. The guarantee was that the party didn’t organize it, which came from the fact that it’s the primary of the right and center. There’s the party on the one hand and the Haute Autorité on the other. That’s all decided. We managed to create a barrier between the party and the organizing authorities, and that was a proper guarantee. At the beginning there was lots of distrust. The Haute Autorité was supposed to get rid of that. Little by little, Nicolas Sarkozy was forced to play it straight. If he attacked the primaries or if they went to hell, he’d be the one responsible, he’d be jinxed by the failure. 
47Paradoxically, the real or imagined threats which hung over the primaries and their organization helped to institutionalize them. The reasons for their consolidation lie in the struggles to control them. 
The PS, a Prisoner of Its Own “Invention”
48While the UMP/LR largely followed the precedent set in 2011, the PS remained a prisoner of its own “invention”. For the president and party leaders, the primary system became “a band-aid sticking to the fingers of the PS leadership”.  Instrumentalized by “dissidents” who exploited the ratchet effect of the first open primary in France and the irreversible precedent it set, primaries gradually developed into a “necessary evil” after 2014.  Their opponents eventually conceded, and primaries were adopted in order to preserve the unity of a party which had been very badly served by the exercise of power and by poor discipline when in government. 
The Temptation to Abandon Primaries
49The 2011 primary was not regulated by statute. It was ratified by party documents, produced in response to its success, at the Toulouse Congress in October 2012. According to these, the Socialist presidential candidate was to be chosen through an open primary whose form and schedule were to be set at least a year before the presidential election (article 5.3.1). Unlike in the UMP, there was no specific clause addressing the candidature of an outgoing president. 
50In spite of being inscribed in the statutes, the primary question did not reappear on the Socialist agenda until 2014. Several factors explain this delay. On the one hand, Hollande’s supporters tried to put off any debate on primaries once it seemed they were directed against the French president. Worse, primaries were even presented as impossible for a sitting president to organize. According to Bruno Le Roux, the president of the Socialist group in the Assemblée Nationale, “the time for primaries hasn’t yet arrived. I don’t know if it will one day”.  As Jean-Christophe Cambadélis said, “primaries are not currently on the agenda for the PS”.  Bartolone, the president of the Assemblée Nationale, was even more direct: “If François Hollande is candidate, I don’t see how a primary could be held, if only for practical reasons”.  Moreover, relatively few in the party leadership expressed support for primaries. Only Julien Dray and Thierry Mandon did so, starting in November 2014. They were later followed by the left wing of the PS.
51But the publication in January 2016 of an opinion piece in Libération calling for “primaries for the left and Ecologists” upset matters. The appeal was signed by intellectuals and leaders from the Ecologists, and it forced the PS to engage with its own “invention” and the “democratic advance” it had made—even if the coalition primary model proposed was very different from the one used in 2011. Above all, the project catalyzed a series of tactical maneuvers in the run-up to the presidential election. By appealing directly to the “grassroots left”,  the signatories were trying not just to short-circuit the parties but to make the PS “face up to its responsibilities”.
52Initially, the proposal met with a degree of approval. But second thoughts quickly intruded. Cambadélis explained the situation:
I had two face-to-face meetings with the president just about primaries in 2016. The first was in January, when Libération published its op-ed. In our conversation, we said that it was possible, but only if Mélenchon wasn’t in it. I said to him: we can start talking with the different partners, but from the outset I’m doubtful about how feasible this is, because the Communists and the Ecologists won’t even think about the outgoing president standing again. François was in favor of the primary. If people decided on a primary, he didn’t think he’d be able to escape it. I proposed a range from Macron to Mélenchon, which was unlikely. And so in his first interview he also mentioned the idea of a Belle Alliance Populaire primary to go beyond the immediate confines of the PS. 
54The first secretary of the PS requested a “Macron to Mélenchon” primary as the only way of guaranteeing the left a presence in the second round. For the other left-wing parties, the prospect of a primary with such broad scope simply looked like the best way to compromise its success. Pierre Laurent, the national secretary of the French Communist Party (PCF), said the “door” of his party was “open”, and Europe Ecology—The Greens (EELV) accepted the proposal in principle. But these two parties very quickly set conditions, demanding that “a base of common values” was agreed on and that the sitting president would not run. Ultimately, only Mélenchon categorically rejected what he had always described as a “political horse race”.
55The leadership of the PS was convinced that the only goal of this initiative was to damage the “natural” candidate. It therefore used Mélenchon’s refusal as a way to shift responsibility for the coalition primary over to its partners and, paradoxically, to emphasize its active support for the project as it gradually failed. As the main target of the call for a broad primary, the PS emerged temporarily strengthened from this attempt at destabilization, having demonstrated that primaries were, more than ever, a “party affair”. Only the PS appeared in a position to organize such a contest.
56The abandonment of the project only put off the decision over a selection method. The statutes did not address the case in which an outgoing candidate stood for re-election, and so the Socialist leaders had to adapt their own internal electoral rules. On 9 April 2016, the national council of the PS voted on a resolution which offered “the president of the Republic, François Hollande, the opportunity to stand if he wishes”. This was presented as a protective measure which could accommodate the constraints of the president’s calendar, and offered the Socialist leadership legal support in their efforts to delay the implementation of their statutes for as long as possible. As Cambadélis summed it up, “up to now, it’s not the PS that dictates orders to the president of the Republic. He’ll stand in the primary if he wants to. I hope he does. But it’s up to him. That’s in the rules. If you read the rules of the PS, they say: ’ıf the president of the Republic wants’”. 
57But this first experiment by the Socialist leadership was not enough. Caught between the president’s allies and Socialist “dissidents”, Cambadélis used a judgment by the president of the Tribunal de Grande Instance de Paris, on 15 June 2016, to clarify the party’s position.  Emphasizing that the statutes of the PS “appear contradictory, or at least ambiguous”, and that “given these ambiguities, no performance obligation can be attached to these statutory provisions”, this judgment confirmed the PS leadership’s freedom to act and, more generally, the distinctiveness of the “corporate law” which governed the internal workings of political parties. 
58In public debate, Cambadélis tested a number of hypotheses. During the meeting of the national bureau on 16 June 2016, he discussed “the possibility of abandoning the primary” by convening an extraordinary Congress which could “designate” Hollande, as well as a number of other “possibilities”, including a primary involving the entire left, or the PS alone. The representatives of the minority motions B and D, which challenged government policy, described this as a “denial of democracy” and a “coup de force”.
The Unforeseen Belle Alliance Populaire
59Against expectations, Cambadélis held a vote on a Belle Alliance Populaire primary during the national council of the PS on 18 June.  This was in order to avoid the fragmentation of a group which, as first secretary, he was required to hold together.
60Cambadélis discussed this crucial series of events in an interview:
I saw the president for a second time ten days before the June national council, on June 8. The ecologists and the communists had said no to the big primary. The president thought, wrongly, that Macron still wasn’t a really serious candidate–he underestimated him and made a serious mistake. We both thought that the governing left could still impose a candidate, and we thought that victory would be possible because Mélenchon was very far to the left and the right was going to be very far to the right. I had the president’s agreement on the primary, in spite of hostility on the part of Valls and some of Hollande’s supporters who weren’t convinced. President Hollande thought that he would have to go through with it if he was to relaunch himself. There was no longer any choice. Our reasoning went like this: the cost to pay if we bypass the primary is huge. If the primary doesn’t happen, we free up a space for the left of the PS, which could put forward a candidate outside the party, or shift over to Mélenchon. We need everyone in the party for the presidential election. Both of us could foresee a defeat in the primary, there was a very real risk, but the primary was the only way of relegitimizing him. But note that the president also thought that the primary might not take place, and that the political situation might change. Valls was completely opposed to it because it yielded too much to the dissidents, it gave in to them. Le Foll was opposed to it because he thought the president was sure to get beaten.
62Rachid Temal, the national secretary of the PS, worked closely with Cambadélis. He confirmed and fleshed out his account:
Cambadélis had to find a solution, before the summer, to the troublesome question of the primary, which we couldn’t get away from. The pressure was huge. It was becoming difficult to take control. People had become convinced that the party needed a primary so that it didn’t explode, but also because we had to relaunch Hollande, and this would give him a way of presenting himself to the left. And people assured him that the primary could go off alright. The fact that the dissidents had several candidates and so were divided also played a role. 
64The decision took the entire party by surprise, because the prospect of a primary on the left in 2017 had been seeming less and less likely. But it allowed the PS to regain control over its “invention”, which the right had now embraced; it reassured the left wing of the PS, which had made the abandonment of primaries a casus belli; and it demonstrated to the president’s allies that any attempt to change the statutes by force would fail.  The desire to reconcile the different factions of the party, and so to preserve unity, was also a key factor.
65This was calculated risk-taking by a first secretary who was “impartial but not neutral”.  These primaries of “relegitimization”  had to be adapted to the interests of the sitting president, both in terms of scheduling (they must be held late) and format (there should be only a short campaign, after the Christmas holidays). As primaries of “reassurance and discussion”, they also had to be instruments of pre-election mobilization. Cambadélis proposed an election held over the last two weekends of January 2017 to fit the schedule which the president had fixed beforehand, and limited to the Belle Alliance alone. Doing so, he proposed a minimal preselection method, and even warned that “there will not be as many voters” in January because of the “haste” in which the primary had been organized. 
66But the unanimous vote on 18 June was only a facade, both among the dissidents and the president’s supporters.  On 21 August, Montebourg, who had been one of the main proponents of the 2011 primary, made his candidature for the presidency official, without saying whether or not he would stand in the PS’s primary. (He wanted to obtain guarantees that it would be organized fairly and trustworthily.) Lacking organizational or financial resources, Montebourg finally acceded and recognized the legitimacy of the party framework which the procedure maintained.  On 27 September, a poll placed Montebourg level with Hollande.  On 20 October, a tell-all book about the president appeared.  This was the point of no return in a primary campaign during which the former first secretary of the PS described his party committing “hara-kiri”.
67To counter the dissidents’ offensive, but also to confront external attacks by Macron (who had resigned from the government), and above all internal attacks by Manuel Valls (whom Bartolone had publicly implicated in the book by Davet and Lhomme), Hollande’s allies once again challenged the idea of holding a primary. They would continue to do so until candidacies officially opened, in order to avoid, in Stéphane Le Foll’s words, “rewriting the history of the quinquennat”.  They hoped to stop the primary turning into an execution—at the risk, however, of a political impasse: either the increasingly uncertain outcome of the procedure would humiliate the outgoing president, or they could bypass such a vote and court charges of a “denial of democracy”.
68On 1 December 2016, Hollande finally announced that he would not seek a new mandate, becoming the first president since 1958 not to stand for a second term.  By making his announcement on the first day candidates could register for the Belle Alliance Populaire primary, he also emphasized how much his withdrawal owed to the obstacle posed by the primary. The 2011 primary aided Hollande’s success greatly; the 2017 primary contributed to his early ouster.
69Unlike the primary of the right and the center, which was designed to strengthen the candidate’s position in the second round, the issues in the Belle Alliance Populaire primary tended to revolve around internal party dynamics. It was no longer a question of reproducing the precedent set in 2011, and even less of the sort of successful vote which LR had held. Instead, it was a matter of transforming a procedure designed to relegitimize the outgoing president into an “early Socialist party congress”, or simply into revenge for the last Socialist congress in Poitiers—the different leadership hopefuls (especially on the left wing of the PS) having for the most part accepted their predicted defeat.
70By withdrawing, Hollande deprived the Socialist candidates of the opportunity to assess his first term. With no outgoing president to oppose, Montebourg lost his main foil. Helped by the president’s withdrawal and his main rival’s destabilization, Benoît Hamon was able to present himself as leader of a “post-defeat” left which would be social and ecologist, and as the only candidate able to adjust his profile to a political situation he himself had helped define and validate. 
71* * *
72Structural factors, logics of imitation, circumstances, political struggles, and the reactivation and adaptation of cultural repertoires came together in a context of declining party legitimacy to produce a primarization process which remains uncertain. Primaries have been dominant within the agenda of the UMP and LR since 2014, even if they initially faced opposition, and have had a significant effect on shaping and institutionalizing the selection process. They were planned with particular care because their organizers, opinion surveys, and political commentators all saw the primary victor as the most likely candidate for the presidency. For the PS, the process occurred far later, and was far less controlled. Initially, party leaders tried to save Hollande from a primary. They then agreed to hastily organize one in January 2017, both in order to relegitimize the outgoing president, and to avoid the party falling apart because of its inability to check internal dissent.  In both cases, Hollande’s victory in the 2011 primaries helped fix the dominant image of the procedure: on the right, it confirmed the idea that primaries were electorally valuable; on the left, it served as a precedent which made the shift to primaries irreversible. In this context, the “democratization” of candidate selection seems secondary. The introduction and development of the primaries should be treated, in a “realist and political” way, as “the result of struggles between actors with different goals fighting to conquer or maintain their power”.  As in other situations, demands for open primaries in France “can be seen as a tool to challenge leaders using the rhetoric of democracy”. 
73The decision to adopt primaries was not a definitive or clear-cut one for either the PS or LR. In spite of the path dependence phenomena we have discussed, the primarization of political life which began in 2011 appeared distinctly threatened by the end of the last electoral cycle. The hypothesis, or prophecy, that primaries would become generalized as part of an irreversible “democratic” process ultimately seems overly mechanistic and deterministic. Much as Hollande’s double victory in 2011–12 secured a place for primaries, the failure of Fillon and Hamon in the second round of the presidential election is today helping to delegitimize the procedure. This supports Balladur’s remark that “as soon as a candidate who wins in the primaries loses the race for the presidency of the Republic, primaries may very well cease to be one of their party’s priorities”. 
74Primaries may have been a response to partisan crises, but in fact amplified them.  It was thought that a competitive but procedural method would in the end guarantee unity; instead, primaries fragmented parties and exacerbated a shift away from their centers. In the PS, primaries were a “tool” to excise the intra-party political differences which had built up since 2012, but they did not reunify the party, avoid schisms, or reduce dissent.  On the right, a considerable turnout (four million people voted in the second round) produced an uncontested but very “right-leaning” candidate, who appealed to the legitimacy of the procedure to protect himself after the Canard enchaîné published a story about his wife’s job in January 2017—a reminder that, unlike in America, the primary phase had not been able to test candidates’ eligibility or produce the one “best suited” to run.  Neither Hamon nor Fillon were able to unify their camp, or to impose the cardinal rule of primaries: support for the winner.
75The period when primaries were ascribed “magic” powers—renewing democratic life and restoring governing parties—has passed, and we have reached a time of disenchantment. In September 2017, a survey found that 70% of LR members rejected primaries, and wanted them to be abandoned.  In a “legacy interview” with Le Point on 13 April 2017, Hollande declared that:
There must be no more primaries for governing parties. Otherwise, there will soon be no more governing parties at all.
77Should primaries be abandoned? For as long as they have existed, particularly in the United States, their development has fueled debates and disagreements about their expected effects: whether they offer an electoral bonus or a handicap, whether they are an “engine of defeat” or a “launchpad”, and so on. Causal beliefs in their (shifting) effects change with experience. Primaries do indeed produce structural effects, insofar as they enlarge the circle of those involved in selecting a candidate—from party members to supporters more broadly or, with the development of strategies which target voters far removed from the organizing party, to an even more vaguely defined electoral body. But they also produce uncertainty, and analysis of them must be set within a particular context. Those participating in the primary are not necessarily representative of the party members organizing it, or of voters as a whole. Furthermore, primaries increase campaign time, and expose the candidate to the dangerous business of “moving into the deep end”, in Hamon’s words. Finally, the process makes it all the more difficult to manage time during the presidential race, because two very different contests follow on from each other in very quick succession. These are the paradoxical lessons of the 2016–17 primaries. Political actors have not yet successfully mastered this manyfaceted unpredictability, and are still taken aback by the perverse effects of a process they introduced because it seemingly offered transparency, stability, and effectiveness.
Quoted by Olivier Duhamel, Les primaires pour les nuls, Paris, Éditions First, 2016, 10.
François Bazin, Rien ne s’est passé comme prévu: Les cinq années qui ont fait Macron, Paris, Laffont, 2017.
“Les primaires: une voie de modernisation pour la démocratie française”, Fondation Terra Nova, November 2011.
Le Monde, 30 November 2016.
Rafael Hortala-Vallve and Hannes Mueller, “Primary: the unifying force”, Working Paper, Barcelona Economics Working Paper Series, 496, September 2010, 1-31.
Jean-Luc Parodi, “Les règles du scrutin majoritaire”, Projet, 122, 1978, 191-200.
Pascal Perrineau (ed.), Le vote normal: Les élections présidentielle et législatives d’avril, mai, juin 2012, Paris, Presses de Sciences Po, 2013.
Le Figaro, 4 May 2017.
Rémi Lefebvre and Éric Treille (eds), Les primaires ouvertes: Adoption, codification, mobilisation, Rennes, Presses Universitaires de Rennes, 2016.
Giulia Sandri, Antonella Seddone, and Fulvio Venturino (eds), Party Primaries in Comparative Perspective, Farnham, Ashgate, 2015.
See Shlomit Barnea and Gideon Rahat, “Reforming candidate selection methods: a three-level approach”, Party Politics, 13(3), 2007, 375-94. The authors discuss three levels which explain the success or failure of attempts to introduce primaries in a country: the political system, the partisan system, and intra-party rivalry.Online
In the words of Matthias Fekl, “Les primaires de 2014: bilan et leçons pour la démocratie libérale”, Pouvoirs, 154, 2015, 81-8.
On the effects of interparty contagion, see particularly the comparison of Britain and Belgium in Bram Wauters, “Democratising party leadership selection in Belgium: motivations and decision makers”, Political Studies, 62(1), 2014, 61-80.
Dominique Damamme, “Les précédents: L’enjeu de la qualification”, Politix, 20, 1992, 35-53.
Recall that the process of codifying primaries is just as decisive as their formal adoption, because primaries are “private” processes whose rules parties themselves define.
Michel Dobry, “Ce dont sont faites les logiques de situation” in Pierre Favre, Olivier Fillieule, and Fabien Jobard (eds), L’atelier du politiste: Théories, actions, représentations, Paris, La Découverte, 2007, pp. 119-48.
This article is based on an analysis of documents (press articles and records of party activities), and on a series of interviews with the main actors in the process of adopting and organizing primaries, carried out during and after the period when these new methods of candidate selection were codified. We have decided not to privilege direct observation and ethnographic analysis, but rather to rely on official documents and to reflect the a posteriori reconstructions offered by the actors we interviewed. The article extends the line of thought broached in an earlier piece of research, which was published in September 2016 and consequently did not deal with the last two presidential primaries (Lefebvre and Treille, Les primaires ouvertes).
For a critical interpretation, see Rémi Lefebvre and Antoine Roger (eds), Les partis politiques à l’épreuve des procédures délibératives, Rennes, Presses Universitaires de Rennes, 2009.
The PS held its first closed primary in 1995, although it had been a statutory possibility since 1978. The right did not do so until 2006.
Rémi Lefebvre, Les primaires socialistes: La fin du parti militant, Paris, Raisons d’Agir, 2011.
William Cross and Anika Gauja, “Designing candidate selection methods: explaining diversity in Australian political parties”, Australian Journal of Political Science, 49(1), 2014, 22-39.
Richard Katz and Peter Mair, “Changing models of party organization and party democracy: the emergence of the cartel party”, Party Politics, 1(1), 1995, 5-28.
Gilles Serra, “Why primaries? the party’s tradeoff between policy and valence”, Journal of Theoretical Politics, 23(1), 2011, 21-51.
Bernard Caillaud and Jean Tirole, “Parties as political intermediaries”, Quarterly Journal of Economics, 117(4), 2002, 1453-89.
Christian Le Bart, L’ego-politique: Essai sur l’individualisation du champ politique, Paris, Armand Colin, 2013.
In the “intra-party politics based models” developed by several Anglophone authors, primaries are seen above all as a way of regulating intra-party conflicts. The intra-party variable is decisive for explaining the decision to use primaries, which can help to preserve party unity, to avoid schisms, and to banish dissent. See Hortala-Vallve and Mueller, “Primaries”.
On 1 October 2009, more than 68% of PS members voted for open primaries for the presidential election.
Fondation Terra Nova, “Les primaires”, p. 23.
Air time figures from the CSA.
See Hortala-Vallve and Mueller, “Primaries”.
Gérard Grunberg and Florence Haegel (eds), La France, vers le bipartisme? La présidentialisation du PS et de l’UMP, Paris, Presses de Sciences Po, 2007, p. 42ff.
In 2006 the PS launched their “twenty-euro activist” program in the wake of a sharp rise in UMP membership.
Le Monde, 12 October 2011.
Libération, 18 September 2011.
Libération, 14 May 2013.
Interview with Maël de Calan, director of the think tank La Boîte à Idées, 4 April 2014.
Interview with Jérôme Chartier, LR deputy, 20 March 2014.
These authors argue that significant institutional pressures are at play, which encourage the homogenization of modes of action within specific “zones” of institutional life. Organizations which face similar constraints tend to converge. (The authors call this logic of convergence “isomorphic institutional change”.) The homology or symmetry of positions within the political field between the UMP and the PS makes this hypothesis of isomorphism plausible–an isomorphism which, these authors claim, can be either normative, coercive, or mimetic. See Paul Dimaggio and Walter Powell, “The iron cage revisited: institutional isomorphism and collective rationality in organizational fields”, American Sociological Review, 48(2), 1983, 147-60.
As an ally of Sarkozy said: “The statutes are a living, moving body. And I don’t think o bligatory’ or imperative’ is written next to the word primary’, is it?” (Libération, June 10 2014).
Le Figaro, 6 November 2014.
Florence Haegel, Les droites en fusion: Transformations de l’UMP, Paris, Presses de Sciences Po, 2012, 30.
Éric Treille, “La démocratie partisane à l’épreuve des primaires”, Pouvoirs, 163, 2017, 97-111.
Le Monde, 28 November 2014.
Interview with Benoist Apparu, LR député, 15 March 2016.
Bruno Le Maire received 29.18% of members’ votes.
Interview with Thierry Solère, 26 May 2016.
“Because they do not only involve the members of a political group or, as far as possible, its candidates, rules about primaries should not be part of the group’s statutes, and should be the subject of a distinct and specific document which the statutes of the initiating party refer to in order to guarantee that there is no doubt about its application”, Anne Levade, “Le droit des primaires: règles, contrôle, finances, sanctions”, Pouvoirs, 154, 2015, 99-109.
Unlike the PS, the Haute Autorité is responsible for organizing polling stations, and particularly for deciding the strategically important question of how many there should be.
Interview with Anne Levade, 16 March 2016.
Interview with Benoist Apparu, 15 March 2016.
Bernard Lacroix and Jacques Lagroye, Le président de la République: Usages et genèses d’une institution, Paris, Presses de Sciences Po, 1993.
Les Échos, 17 June 2016.
Libération, 13 April 2016.
Hortala-Vallve and Mueller, “Primaries”. Hollande’s term saw new forms of dissent within the party. Frédéric Sawicki has shown that this disunity is a result of lack of work by the party to produce unity. See Frédéric Sawicki, “L’épreuve du pouvoir est-elle vouée à être fatale au Parti socialiste? Retour sur le quinquennat de François Hollande”, Pouvoirs, 163, 2017, 27-41.
Unlike the American system–where the status of outgoing officeholders is typically integrated into the primary itself–and the UMP, who drew up an article ad hoc. (“No primary will be organized when the president of the Republic comes from the Movement and is a candidate for a second term”.)
Le Figaro, 19 December 2014.
AFP, 28 April 2015.
Le Monde, 24 July 2015.
Libération, 11 April 2016.
Interview with Jean-Christophe Cambadélis, 28 November 2017.
Le Monde, 11 April 2016.
The president dismissed the claims of three Socialist activists who had asked the PS to invoke article 5.3.2 of its statutes.
Daniel Gaxie, “La liberté d’organisation des partis politiques” in Julie Benetti, Anne Levade, and Dominique Rousseau (eds), Le droit interne des partis politiques, Paris, Mare et Martin, 2017, pp. 23-32.
The Belle Alliance Populaire consists of the PS, the Union des Démocrates et des Écologistes, the Parti Écologiste, and the Front Démocrate.
Interview with Rachid Temal, 26 June 2017.
“The PS invented primaries, and everyone else imitated it. It invented primaries for an outgoing president, and everyone is going to imitate it”, Cambadélis wrote on Twitter on 19 June 2016. As he summed things up, Hollande “couldn’t turn up in December and declare himself the candidate”. Libération, 19 June 2016.
Jean-Christophe Cambadélis, Chronique d’une débâcle: 2012-2017, Paris, L’Archipel, 2017.
The first secretary’s language to the press was quite explicit: “We need an elegant fight, not a punch-up”; “a primary of reassurance and discussion”, Libération, 20 June 2016.
He expected there to be 6,000 to 8,000 polling stations, compared to 10,000 in 2011. Speaking on France 2 on Thursday, 14 April 2016, Hollande said that he would announce his decision on whether he had decided to stand “at the end of the year”.
François Kalfon, an ally of Montebourg, declared: “We reject this PR stunt of a Belle Alliance primary. We want a primary to increase mobilization in December, not one designed to rob us in January”. Le Monde, 19-20 June 2016.
This confirms that the Socialist primary was the tool the ruling elites used against the ambitions of the party’s minority factions, in order to pull the wool over their eyes and so to stop them leaving. See once more Hortala-Vallve and Mueller, “Primaries”.
According to the Ipsos-Cevipof-Le Monde survey of 27 September 2016.
Gérard Davet and Fabrice Lhomme, Un président ne devrait jamais dire ça… Les secrets d’un quinquennat, Paris, Stock, 2016.
Bartolone asked that “both the president and the prime minister stand in the primary”. As Frédérique Espagnac, a Socialist senator and Hollande’s former press secretary, told France Info, “The principle of the primary is to bring people together. But today we have four external candidates, and we face the question of whether to cancel it”.
On Hollande’s withdrawal, see also Vincent Martigny, “À gauche, la fin de la synthèse social-démocrate” in Pascal Perrineau (ed.), Le vote disruptif: Les élections présidentielle et législatives de 2017, Paris, Presses de Sciences Po, 2017, pp. 43-58.
As Hamon said to Libération on 6 January 2017, “It feels like everyone has already internalized the left’s defeat in the presidential race”.
Sawicki, “L’épreuve du pouvoir”, 35.
Piergiorgio Corbetta and Rinaldo Vignati, “The primaries of the centre left: only a temporary success?”, Contemporary Italian Politics, 5(1), 2013, 82-96, here 85.
Quoted by Duhamel, Les primaires pour les nuls, 10.
Barnea and Rahat, “reforming candidate selection methods”, 375-94.
Ozge Kemahlioglu, Rebecca Weitz-Shapiro, and Shigeo Hirano, “Why primaries in Latin American elections?”, Journal of Politics, 71(1), 2009, 339-52.
Serra, “Why primaries?”
On the left, Cambadélis has proposed a procedural solution similar to the closed primary held in 2006: “We can’t retreat on primaries. We’re finished with congresses, and with the first secretary being the natural candidate. My suggestion is to introduce a membership fee, which will take us from a primary with three million voters to one with 500,000”. Interview with Jean-Christophe Cambadélis, 28 November 2017.