Navigation – Plan du site

AccueilNuméros31 | 2nd semestreVariaIn search of the bloc bourgeois

Varia

In search of the bloc bourgeois

À la recherche du bloc bourgeois
En búsqueda del bloque burgués
Bruno Amable

Résumés

Cet article propose une analyse du bloc social qui pourrait soutenir la stratégie politique de changement radical du modèle socio-économique français. Ce bloc social, le bloc bourgeois, rassemblerait les groupes aisés et éduqués des blocs sociaux traditionnels de gauche et de droite qui ont structuré la compétition politique pendant la Cinquième République. L’analyse se fonde sur une étude empirique des attentes de l’électorat en matière de politiques publiques à partir de l’enquête électorale française de 2012. Un modèle de classes latentes conduit à une partition de l’électorat français en dix-huit clusters. Cette analyse apporte un éclairage sur la composition des bases sociales respectives de la gauche et de la droite ainsi que sur la possibilité d’existence du bloc bourgeois. L’article contribue également à expliquer les raisons de l’échec de François Hollande ainsi que les difficultés de l’actuelle présidence d’Emmanuel Macron.

Haut de page

Texte intégral

Introduction

  • 1 For a definition of neoliberalism, see Pierre Dardot and Christian Laval’s book (2009) and Bruno Am (...)

1Contrary to a widespread cliché, the French socio-economic model is anything but “impossible to reform” and one could on the contrary argue that it has undergone significant institutional change (Amable, 2017) since the 1980s. What is true however is that some of these “structural reforms”, from the plan Juppé of 1995 for social protection to the 2016 labour market flexibility enhancing loi travail, have met considerable social resistance and later led incumbent governments to electoral defeat. Nevertheless, these political difficulties have not deterred mainstream political parties from keeping on with their intentions to transform the French socio-economic model. Yet whereas there was in the 1970s competition between the left and the right concerning the direction that these transformations should take – towards socialism or, more modestly, social democracy for the former, towards a more liberalised economy for the latter – the matter seems to be settled since the 1980s. Neoliberal1 reforms have been implemented by left as well as right governments; they first concerned financial deregulation and privatisations, but they have later affected the employment relationship and social protection.

2If the objective is relatively clear, the political means to achieve it are more evasive. Implementing a radical economic transformation project demands a social base that successive governments have failed to find not only on the left but also on the right. Analysing the problems faced by the left government at the beginning of the first François Mitterrand presidency, Alain Lipietz (1984) identified the crisis of Fordism as a crisis of hegemony, with the Fordist period defined as the outcome of a partial and temporary alliance between a fraction of industrial capital and wage-earners. The crisis of Fordism expressed the dissolution of the dominant social alliance. Finding a new model implied thus to build a new one.

3The aim of this paper is to analyse the possibility of a social base for the neoliberal economic, social and political transformation project in France. Following Bruno Amable and Stefano Palombarini (2009), the concept of social bloc is used to assess the stability of a political strategy for institutional change. Stability is defined as the existence of a dominant social bloc, i.e. a set of social groups whose demands are taken into account in the definition of public policy and institutional design. The identification of this potential dominant social bloc is performed with a statistical analysis of individual policy demands and expectations, in order to determine endogenously the social groups that could be united by a political strategy geared towards neoliberal reforms. Using data from the 2012 French electoral study (FES2012), a latent class analysis defines social groups based on the proximity of the policy demands of the individuals that compose them. The aim is to see how these groups could possibly be aggregated in a potentially dominant social bloc.

4This paper follows and extends the contribution of Bruno Amable, Elvire Guillaud and Stefano Palombarini (2012) in the identification of social groups and social blocs starting from the individuals’ expression of policy preferences using a different (and better) data source. The concept of bloc bourgeois is given an empirical content that goes beyond a description in terms of social groups defined ex-ante (e.g. the middle class or professionals, etc.). The paper proceeds as follows. The first section presents the theory of social blocs. Then, the gradual breakup of the traditional left and right social alliances and the difficulties of the search for an alternative socio-political alliance are presented. The following section turns to the empirical analysis of the policy preferences and presents the latent class model’s results. The possibilities of aggregating a new social bloc, the bloc bourgeois, are then discussed in another section. A brief conclusion follows.

1. Social blocs

  • 2 I avoid to use the word “class”. There is a plethora of definitions of social classes in the litera (...)
  • 3 This does not imply that all demands of all agents in a social group are perfectly identical.
  • 4 One could distinguish between social groups, defined by a proximity of position in the socio-econom (...)

5The main concept upon which the analysis that follows is built is that of a social bloc. In the French theory of regulation, the concept is directly taken from Antonio Gramsci and represents “a stable system of relations of domination, alliances and concessions between social groups (dominant and subordinate). A social bloc is hegemonic when its interests correspond to those of the whole nation.” (Lipietz, 1994, p. 340). In Lipietz’ view, the hegemonic bloc must be comprehensive and only a small proportion of interests are discounted. Amable and Palombarini (2005; 2009) have a different position. They define social groups2 by a community of interests as they are perceived by the group members that may, under condition of organisation, lead to the expression of demands for specific policy options or institutional change in various areas. A social group is a set of agents united by common perceived interests, a perception mediated by the system of representations of these agents, in short ideology. They therefore express common policy preferences3. A social bloc gathers social groups that express different and sometimes antagonistic demands4. It is unified by a political strategy that aims to satisfy the most important demands of the relevant social groups. The role of political leadership is to select among social demands those that will be satisfied and those that will be left out. This choice is made under different economic and internal compatibility constraints. Yet the main criterion that political actors consider is the ability of the different social groups to supply political support, which includes but is not limited to vote. Support does not imply a perfect satisfaction of all the demands of all the social groups that form the bloc. Not all groups are equal inside a bloc; the chosen political strategy and the priority given to certain demands through the implementation of specific policies or “reforms” reflect the hierarchy of groups in terms of political support.

6A social bloc is dominant when the strategy that unifies its constituent groups is politically successful. The contestation stemming from social groups whose demands have been disregarded (the dominated groups) is limited to a minority position of the political representation space. With respect to Lipietz’ Gramscian definition of a social bloc, two differences are worth pointing out. Firstly, it is not necessary that only a small number of social groups be excluded from the dominant social bloc; in fact, looking at France in the late 2010s, early 2020s, the bloc bourgeois, although dominant in the sense exposed above, is rather limited in size (Amable & Palombarini, 2018). One could not say that this bloc is hegemonic in the Gramscian sense. Secondly, several social blocs may exist and compete for dominance. Amable, Guillaud and Palombarini (2012) have analysed the political competition in France as opposing two different strategies aiming for the support of two differentiated social blocs.

2. The breakup of the left and right social blocs in France

  • 5 For a typology of models of capitalism, see The Diversity of Modern Capitalism by Amable (2003).

7One could distinguish two different social blocs corresponding to two opposed political strategies, a liberal and post-Gaullist alliance on the right and a socialist-communist (Parti socialiste-Parti communiste français, PS-PCF) alliance on the left (Amable, Guillaud & Palombarini, 2012; Amable, 2017). To give a broad picture, the left social bloc included the majority of the public sector employees and of the working classes. The core of the right bloc were the medium and superior categories of the private sector, the self-employed and professionals, and farmers. The policy expectations of these blocs were strongly antagonistic. The left bloc expected an extension of social protection and industrial democracy, an economic policy promoting real wage increases and greater state control on the economy. The expectations of the right bloc were a mix of social conservatism and public intervention in the economy, at least until the crisis of the 1970s, and a more or less progressive transition to a neoliberal model of capitalism from then on5.

  • 6 Chirac, who had once declared that he was looking for a French style Labour party (travaillisme à l (...)
  • 7 How radical Sarkozy would have been is uncertain considering that the 2008 crisis compelled him to (...)

8Both the left and the right social blocs started to gradually fall apart in the 1980s. On the right, the 1980s saw the growing divergence between a radical neoliberal core, favouring drastic market liberalisation reforms, and a more moderate fraction that wanted to preserve the essential elements of the French social model. As analysed in André Gauron’s work (1983; 1988), tensions were present during most of the Fifth Republic. The first attempt to implement neoliberal reforms was made under Valéry Giscard d’Estaing’s presidency, with prime minister Raymond Barre succeeding Jacques Chirac in 1976. Unlike Chirac, whose economic policy orientations were at that time still influenced by post-war “Keynesianism”, Barre was convinced that the only way out of the 1970s’ economic crisis implied a structural transformation of the economic model, with large firms adapting to international competition, which implied to relinquish the Gaullist economic development model and introduce a radical change of macroeconomic policy away from “Keynesianism”. However, the most drastic attempts at implementing a neoliberal structural reform program took place in 1986, when a right coalition tried to emulate the Thatcherite and Reaganite liberalisation programme6. The attempt split the right social base (Amable, Guillaud & Palombarini, 2012), with a part of the right bloc rejecting the radicality of the programme while another part, more skilled and affluent than the former, welcomed it. After the electoral failure of 1988, the right, when in power, was always cautious in its attempts to liberalise the French socio-economic model, conscious that the right social bloc was itself divided on the issue of the radicality of the neoliberal transformation (Amable, 2017). Even Nicolas Sarkozy, who presented himself as a radical reformer in the presidential election of 2007, promised that the labour market reform he intended to implement would take the form of “flexicurity” and not be an all-out flexibilisation7.

9On the left, after Mitterrand’s electoral victory in 1981, the U-turn in economic policy in 1982-1983 (Lordon, 1998) established an enduring contradiction between on the one hand the supply-side economic policy that PS-led governments wanted to implement, geared towards European monetary unification (EMU) and the achievement and deepening of the single market, and on the other hand the policy expectations of the left bloc.

  • 8 Roughly one third of PS in 1979 (Fulla, 2016).
  • 9 The Programme commun de gouvernement elaborated with the communist party and PS’ satellite Radicaux (...)

10Already at the end of the 1970s, a fraction8 of PS were hostile to the full implementation of the left economic manifesto9 and were inclined to adopt a more orthodox macroeconomic policy along with, later, implementing “structural reforms” liberalising the French socio-economic model. This would have consequences for both the political alliance and the social base of the left; both would have to change for PS to stay in power. In the decades that followed Mitterrand’s victory in 1981, this change proved too difficult to achieve. The failure was epitomised by former minister and president of the European Commission Jacques Delors renouncing to be PS’ presidential candidate in 1995 because he considered that the pro-European integration and pro-structural reforms economic policy he wanted to implement required a political coalition that was impossible to form.

  • 10 Lamy P. (2005), « Le modèle français vu d’Europe. Entretien », Le Débat, no 134, p. 31-41.
  • 11 Le Figaro, 24 mars 2005. Original citation : « l’Europe a été notre programme d’ajustement structur (...)

11The European issue appeared increasingly linked to that of “structural reforms”. European integration took a specific turn in the 1980s with the single market and EMU. The former promoted competition at the level of the union and more or less directly affected the institutions that had been the core of the post-war socio-economic model, in particular the public sector. The controversial social protection reform of 1995 was justified by prime minister Alain Juppé as a consequence of the requirements of the Maastricht treaty for public deficits. The former Medef vice-president Denis Kessler considered “Europe” as “a machine to reform France against its will”. Pascal Lamy, former chief of staff of Jacques Delors at the European Commission, considered that “the reordering and the ‘marketisation’ of the French economy […] were made by Europe, thanks to Europe and because of Europe10”. Economist Jean Pisani-Ferry stated bluntly that “Europe was our structural adjustment program. […] France has let European integration play the role that the IMF or the World Bank play for poorly governed countries11”.

  • 12 For a discussion of Europe as a new political divide, see Les Français contre l’Europe ? by Nicolas (...)

12Consequently, “Europe” emerged progressively as a salient political divide cutting across the left and the right blocs as it became clear during the referendums on the Maastricht (1992) and the constitutional (2005) treaties12. For a part of the left bloc, European integration appeared to lead to the imposition of a series of reforms that were the opposite of the evolutions that they wanted for the socio-economic model: restrictive monetary and budget policies, privatisation of public sector activities, increasing labour market flexibility, welfare state retrenchment, etc. The situation was somewhat different for the right bloc, where the opposition to European integration stemmed partly from the will to keep control of the macroeconomic policy but was also related to the consequences of increased competition for some protected activities, as well as the will to defend national sovereignty in other areas of public policy, including immigration (Amable, 2017).

13The pro-European integration social groups were split across the traditional left and right blocs. The most skilled and better-off parts of both blocs expressed similar demands regarding the continuation of the integration process. The less affluent or skilled groups of both the left and right blocs expressed less confidence in that process. The rising importance of the European integration issue in the 1990s and 2000s brought the pro-EU groups closer together, as the analyses of the 1992 and 2005 referendums on respectively the Maastricht treaty and the constitutional treaty showed (Lehingue, 2007). The possibility to structure political competition around the issue of European integration appeared as a way to escape from a political deadlock once summed by Jean-Claude Juncker as the impossibility to implement allegedly necessary (structural) reforms without losing the following election.

14Confronted with such a problem, the left first tried to find a new compromise and extend their social base on the right rather than change it drastically. In the late 1990s, prime minister Lionel Jospin claimed that PS’ “sociological base […] has been renewed and extended. This is why [PS] must find the best trade-off between social classes. Those who are rather satisfied with the current state of society and do not want to incur the ‘cost’ of increasing equality [and those] for which the notion of equality and its concrete deepening are fundamental.” (Jospin, 2000, p. 50). His spectacular failure at the 2002 presidential election proved the difficulty of finding the best trade-off.

  • 13 The Terra Nova report made more or less explicit reference to some political science analyses focus (...)
  • 14 When in 2013 a journalist pointed out to Hollande that his decision to let Mittal Steel mothball th (...)
  • 15 This bloc is characterized by the absence of working classes, which were present in the traditional (...)
  • 16 The traditional left/right cleavage is expressed in terms of economic policy preferences, redistrib (...)

15The need to find alternative social support became gradually more pressing. In 2011, shortly before the presidential election, the so-called progressive think tank Terra Nova urged “the left” (2011), in fact PS, to jettison the working classes because their policy demands would be impossible to satisfy: their economic demands, because globalisation and the evolution of modern capitalism had allegedly made the welfare state unsustainable and the neoliberal reforms inevitable; their “cultural” demands, because these were authoritarian and clashed with the “values of the left” as defined by Terra Nova13. Terra Nova (2011) recommended that the mainstream left look actively for a new, rather fuzzy, social base, “la France de demain”, away from the traditional constituency of the left. If François Hollande did not heed this recommendation during his left-oriented 2012 presidential campaign, his presidency was certainly marked by the search for a new social base14 and the successive governments of his mandate were formed excluding not only parties to the left of PS but eventually also the left wing of PS. Under his presidency, structural economic policy spectacularly took a neoliberal direction, culminating with a labour market reform, the loi travail, which promoted labour market flexibility far more than any previous reform implemented by right governments. Regarding the social base for such a project, the construction of a new social bloc excluding the working classes ambitioning to be dominant, the bloc bourgeois15 (Amable & Palombarini, 2014; 2017), gathering the high-skilled and affluent pro-EU social groups demanded that the political actors looking for it should actively go beyond the traditional left/right cleavage16 and promote a European divide instead.

  • 17 Giscard d’Estaing is one of the very few politicians to have claimed the label: “The most scientifi (...)

16The search for a social base likely to support the transition of the French socio-economic model towards a neoliberal market economy in connexion with the deepening of European integration was not limited to the left. Already in the 1970s, the liberal fraction of the right had expressed similar ambitions. President Giscard d’Estaing in the 1970s was the first leading French politician of the post-war period to have an explicitly neoliberal agenda17. He called for the emergence of a new social alliance, the “central group”, that would, according to him, gather two-thirds of the French population united by “common cultural attitudes” and “similar lifestyles” (Giscard d’Estaing, 1984, p. 246). François Bayrou, a centre-right politician several times minister in conservative governments, tried unsuccessfully to follow Giscard’s footsteps, promoting European integration, structural reforms, and fiscal conservatism. His electoral failures illustrated the difficulties for a politician identified with the support of a traditional social bloc to transcend the left/right cleavage and reach out to social groups coming from the opposite bloc.

3. Identifying socio-political groups

  • 18 For all questions, there was a possibility of no answer.

17The possibility of a new dominant social bloc can be investigated with the structural analysis of the political demands by the different socio-political groups. In order to identify these groups, a bottom-up approach is adopted. Rather than consider already defined social groups, on the basis of occupation, for instance, the analysis will start from the policy preferences expressed by individuals. The latter will be allocated into endogenously defined groups according to the similarity of their policy preferences. These will be identified by the answers18 given to the electoral study of 2012 (Sauger, 2012), a post-election survey comprising 2014 face-to-face interviews, representative of the French population registered on the electoral roll.

  • 19 The 2012 survey included some questions on economic policies and institutional reforms which made p (...)

18The 2012 electoral survey (FES2012) is chosen for two reasons. First, as explained in the previous section, the date corresponds to a turning point for the stability of the traditional social blocs. Apparently, the traditional opposition between the right and left blocs was the same as it ever was, but in fact the breakup of the two blocs was already underway. François Hollande was designated as the presidential candidate for PS following a primary that was open beyond the left party members, with the aim of gathering a broader, and in fact different, social base. More significantly, as mentioned before, Hollande never envisaged a left governing coalition like those that had governed in 1981 or 1997. Once elected, he oriented his political strategy away from the traditional support of the left, eventually facing an internal left opposition among PS deputies. In terms of social base, Hollande was the first left president to govern without looking for the support of the traditional left bloc. The second reason is that, in comparison to other available surveys19, FES2012 is remarkable by the quality of the questionnaire regarding policy preferences, which includes precise questions on sensitive policy options that can be used for our empirical analysis.

19The issue underlying the empirical analysis is that of the neoliberal transformation of the French socio-economic model. The choice of items in FES2012 is therefore guided by the question of the possibility to aggregate a social bloc that could support such a transformation. Ideally, one would have liked to have questions affecting all five institutional areas used to define a model of capitalism in Amable’s book (2003) – the labour market, product markets, the financial system, social protection and the education and training system – in addition to the broad economic policy orientation. However, one has to take into account the constraints of the existing questionnaire.

Four different areas of demands regarding policy options or institutional change are distinguished. The first one is the broad economic domain, with the general orientation of economic policy (wage policy, public expenditure, etc.) and the structural reforms issue. The most sensitive structural reforms concern the employment relationship and social protection.

  • 20 For instance Manuel Valls, Hollande’s prime minister.

20Regarding labour market reform, one central issue of the 2000s was the replacement of all open-ended and fixed-term employment contracts by a single employment contract. The so-called contrat de travail unique (CTU) would have workers’ rights in terms of benefits and entitlements as well as firms’ firing costs increasing with tenure. Formally, the new contract would be open-ended, but firms would be relieved of some of their obligations in case of employment termination: no obligation to propose a redeployment plan for fired workers in case of collective dismissal, or to provide an economic motive for the termination of the contract and therefore no possibility for a judge to check whether such a motive is valid. The single employment contract was one of the reforms promised by conservative Nicolas Sarkozy during the 2007 presidential campaign but was never implemented (Amable, 2014). The proposition popped up again before the 2012 presidential election, and the centre-right candidate to the presidential election, François Bayrou, included it in his programme. Some prominent PS politicians expressed interest in this measure before and after the election20. A question in FES2012 asked respondents for their opinion on the contrat unique. Another question related to labour market regulation and social protection asked whether the respondent thought that the unemployed could find a job if they really wanted to. Although this is not a question directly implying a policy decision, the responses could be informative on how receptive the respondent would be to labour market flexibility-enhancing policies or cuts in unemployment benefits.

  • 21 For this question and others concerned, the initial range of response from 0 to 10 was re-expressed (...)
  • 22 These issues are still at the centre of the 2020 pension reform.

21Regarding social protection, an important reform of the Sarkozy presidency concerned pensions. It included many changes, the most significant being a two-year extension of the working period necessary for obtaining a full pension. The reform led to a significant union-led contestation in autumn 2010 and one electoral promise of PS candidate François Hollande was to partially reverse it. Social protection financing was also a reform discussed in the years preceding the 2012 election. A project put forward by Sarkozy was to partially substitute a financing of social protection by value-added tax (VAT) to the system based on employers’ and employees’ social contributions. The drop in social contributions would lower labour costs and allegedly boost employment. However, the transition would have distributional as well as institutional consequences: VAT is a regressive taxation and the contribution-based funding legitimates the involvement of trade unions in the management of unemployment insurance. Two questions were considered, one on the fairness of Sarkozy’s pension reform21, the other on the desirability of a rise of the value-added tax, with the implication that the extra receipts would finance the social protection system22.

22Two questions dealt with the size of the public sector. One asked whether the respondents wanted to increase or decrease the number of civil servants. A second question asked whether the respondent had a positive or negative opinion on privatisations. A traditional question on how favourable the respondent is to the income inequality-reducing action of the government was also included, as well as a question on whether the priority of economic policy should be to improve firms’ competitiveness or employees’ condition.

23The second policy area considered is European integration. As mentioned before, this issue is central in the breakup of the traditional left and right blocs and the unification of a tentative bloc bourgeois. One question dealt directly with the desirability of the deepening of European integration: respondents were asked whether they would like to see more power given to the French state or to “Europe”.

  • 23 See the paper by Sheri Berman and Maria Snegovaya (2019).

24The third area concerns the broadly “cultural” issues: the attitude towards immigration as well as the tolerance towards alternative lifestyles. Regarding immigration, respondents could give their opinion on whether to stop or continue it. They were also asked whether they thought that homosexual couples should have the right to adopt children. These questions are included in the analysis firstly because a large political science literature has focused on the emergence and existence of a “second dimension” of political divide based on “cultural” issues, and secondly because the social-democratic parties that implemented neoliberal reforms have often put forward such “postmaterialist” issues in order to make up for their shift to the centre on economic issues23. Identifying social groups for which “cultural” issues matter may be important for the definition of the bloc bourgeois.

Finally, in France as in other countries, the environment issue has seen its importance increase as a political divide. A sensitive question has concerned nuclear energy, because of the size of the nuclear programme launched after the oil shocks and the controversies regarding the safety issue. A question on the desired future of nuclear energy (stop it or go on with it) is considered.

4. The identification method

4.1. The latent class model

25The objective is to identify the groups that could form a new social bloc. These will be defined by the proximity among their members regarding policy preferences. In order to achieve this, it is helpful to start from the expression of preferences at the individual level and see how one can endogenously determine social groups. An adequate method for doing this is latent class modelling. The objective of the latent class analysis is to find a categorical structure for the individuals of FES2012. The model considered in the empirical analysis has the following form:

26γit is the answer of individual i to one of the T survey questions included as indicators in the model. x is a latent variable designating the cluster (group) to which the individual i may belong (1 ≤ x ≤ K). f is the density corresponding to a particular set of γi and P(x) is the probability to belong to a certain cluster. One supposes a multinomial distribution for the γit and a model of multinomial or ordered logistic regression according to the variable. The values of the latent variable are also assumed to come from a multinomial distribution and a multinomial regression model is used.

  • 24 The test provides information on the improvement that an additional cluster provides to the model. (...)

27There is no definitive criterion for choosing the number of clusters. Information criteria (AIC, BIC…) are often used. The simulations performed by Karen Nylund, Tihomir Asparouhov and Bengt Muthén (2007) lead to the conclusion that the indicator giving the best results is a log-likelihood difference-based test (bootstrap likelihood-ratio test), using samples obtained by bootstraps to estimate the value of the statistics24. This test is used in determining the number of clusters in the model presented in the next subsection.

4.2. Model estimation

  • 25 The models were estimated with Latent Gold 5.0.

28A series of models were estimated25, increasing each time the number of clusters until the bootstrap likelihood ratio test indicated the lack of significance of adding another cluster to the model. After a series of estimations, a 19-cluster model was eventually rejected by the test. The chosen model comprises therefore eighteen clusters for the sample of 2014 respondents to the French electoral survey.

29The eighteen clusters are defined by the policy expectations of the individuals. One may also want to look at the variables influencing (probabilistic) cluster membership. Although it would have been possible to include predictors directly in the latent class analysis, it would have been cumbersome to do so while testing the significance of the different variables. Also, these variables would have influenced the outcome of the classification, which was not desirable. Therefore, the so-called three-step approach was chosen instead. After having estimated the model of interest (first step), individuals were assigned to latent clusters using the posterior cluster membership probabilities (second step) and then the association between the assigned cluster memberships and external variables could be investigated. The corresponding results will be presented after the characterisation of the clusters.

5. The 18-cluster model

5.1. Policy preferences

30A succinct description of the groups is given in table 1, with the relative weight of each one in the sample, their main policy demands and some general characteristics. Cluster 1 gathers individuals opposed to neoliberal reforms (single labour contract, pension reform, income inequalities, decrease in the number of civil servants...) and broadly in favour of European integration. Cluster 2 members are somewhat similar to cluster 1’s in terms of policy expectations but differ with respect to European integration. A majority of this cluster would want more power given to the French state rather than to the European Union [EU]. Cluster 3 individuals are divided on the issue of the single employment contract and rather hostile to giving more power to the European Union. They are not particularly favourable to immigration or homosexuals’ rights; they tend to agree with the idea that the unemployed could find a job if they really wanted to. They are not in favour of neoliberal policies such as privatisations, pension reform or reducing the number of civil servants; they are pro-redistribution and favourable to an economic policy oriented towards wage-earners. The people in cluster 4 are divided on the issue of the single labour contract as well as other neoliberal reforms (income inequality policies, number of civil servants). They are rather hostile to immigration and homosexuals’ rights and do not want more power given to the European Union. Cluster 5 are not in favour of giving more power to the EU and are on the whole hostile to neoliberal reforms. Cluster 6’s policy preferences lead them to support European integration. They support some neoliberal reforms (pension reform, privatisations) without necessarily being strongly in favour of all of them, such as the single labour contract. They support inequality-reducing state intervention, but they consider that economic policy should be geared towards improving firms’ competitiveness.

Table 1. Cluster description

Group

Weight

Policy preferences

Characteristics

1

10.3%

Pro-European integration; hostile to neoliberal reforms

Young; high income; high education level; public sector

2

9.1%

Not pro-European integration, hostile to some neoliberal reforms

Middle-aged; low income and education level; clerks, workers, artisans, women

3

8.8%

Not pro-European integration; hostile to some neoliberal reforms

Middle-aged; low or middle income; technical education

4

7.9%

Anti-European integration; favourable to some neoliberal reforms; anti-immigration and homosexuals’ rights

Low income and education level; old

5

6.7%

Anti-European integration; hostile to some neoliberal reforms

Young, high education level; public sector

6

6.5%

Pro-European integration; ambiguous on neoliberal reforms; pro-competitiveness and VAT increase

High income, wealth and education level; upward social mobility

7

6.4%

Undecided on European integration, undecided on neoliberal reforms except pro-privatisations

High income; higher technical education; upward social mobility, women

8

6.1%

Anti-European integration; pro-CTU; pro-redistribution, against nuclear energy, against immigration

Women, young, middle income

9

5.4%

Divided on European integration; hostile to immigration and homosexuals’ rights; favourable to some neoliberal reforms

Men, old; high income; self-employed

10

5.4%

Pro-European integration; pro-neoliberal reforms; anti-immigration

Men, old; high income & education level

11

4.7%

Not pro-European integration; pro-redistribution; in favour of some neoliberal reforms; not favourable to immigration

Old

12

4.7%

Not pro-European integration; pro-redistribution; not favourable to immigration

Young; middle income

13

4.3%

Express no opinions on many issues

Women; low income and education level

14

4.3%

Against neoliberal reforms; pro-immigration

Young; low income

15

3.9%

Divided on European integration; favourable to neoliberal reforms; against immigration

Old; high income

16

2.1%

Pro-European integration; favourable to neoliberal reforms

Young, high or middle income

17

2.1%

Divided on European integration; favourable to some neoliberal reforms; against immigration

Old, middle income

18

1.5%

Pro-European integration; favourable to neoliberal reforms; against immigration

High income

Source: Bruno Amable, 2018

31Cluster 7 are split on the matter of giving more power to the EU or the French state and the decrease in the number of civil servants. They are rather opposed to the single labour contract but not to the pension reform. They are mildly opposed to redistribution (relatively to the population average), in favour of nuclear energy, privatisations and homosexuals’ rights, but not strongly supportive of immigration. Also, they approve of an economic policy favouring competitiveness over wages. Cluster 8 are against more power given to the EU, favourable to the single labour contract but opposed to the pension reform, and in favour of an economic policy improving wages. They are no supporters of redistribution or immigration and are divided on homosexuals’ rights. Cluster 9 are split on the European integration issue, rather against the single labour contract but in favour of the pension reform. They want an economic policy for competitiveness and are no supporters of homosexuals’ rights or immigration. Cluster 10 are pro-European integration, broadly in favour of neoliberal reforms, and against immigration and homosexuals’ rights. Cluster 11 are not in favour of more power given to the EU and express some views in favour of neoliberal reforms (the unemployed could find a job, fewer civil servants…).

32The policy demands of cluster 12 are not in favour of European integration. They expect the redistributing action of the state and are otherwise split or undecided on most neoliberal reforms. They do not support immigration or homosexuals’ rights. The most significant characteristic of cluster 13 is that they express no definite policy preferences (“no answer” to most questions). The policy expectations of cluster 14 do not favour European integration or most neoliberal reforms (single labour contract, pension reforms, privatisations) but the group support an economic policy oriented towards wages as well as immigration and homosexuals’ rights. Cluster 15 are divided on European integration and in favour of most neoliberal reforms (CTU, pension reforms, privatisation, decrease in the number of civil servants) and nuclear energy; they are against redistribution, immigration and homosexuals’ rights. Cluster 16 are pro-European integration and express otherwise standard conservative policy demands: in favour of the single labour contract and the pension reform, and against redistribution.

33Cluster 17 are divided on European integration and have conservative policy demands: favourable to the single labour contract, privatisations and pension reform; against income redistribution and immigration. Cluster 18 are favourable to European integration, a pro-competitiveness economic policy and neoliberal reforms (single labour contract and pension reform); they are hostile to income redistribution and immigration. Clusters 17 and 18 are rather similar in their policy demands but cluster 17 are not so keen on a policy oriented towards competitiveness and would rather welcome a decrease in VAT whereas cluster 18 are favourable to an increase.

5.2. Individual characteristics and cluster composition

  • 26 The income variable selected is the level of total household income divided by the square root of t (...)

34One can precise the identity of the different social groups by looking at variables that have not been taken into consideration for the definition of the cluster grouping but define some common social characteristics of the clusters nonetheless. The following remarks are based on the difference between the proportion of a cluster in the sample and the relative weight of certain categories. Cluster 1 membership is associated with a lower-than-average proportion of seniors: cluster 1 represents 10% of the individuals of the sample but gathers only 5% of those over 65. Individuals (probabilistically) belonging to this cluster are better-off (over 16% of the cluster’s membership in the top two deciles of the distribution of household income26) and have a higher education level (28% of the membership with a university degree higher than bachelor) than the average population. The proportion of civil servants is high (20% of the civil servants of the whole sample are in this group). Apart from the age structure (more middle-aged individuals) and the higher than average proportion of women (61% of the cluster), the main differences with cluster 1 are the income and education levels: cluster 2 have lower degrees (the cluster represents 9% of the individuals but regroups less than 5% of the individuals with a university degree) and income levels (the cluster gathers less than 3% of the individuals in the top two deciles of the distribution of household income) than cluster 1. The only distinctive characteristics of cluster 3 is a higher-than-average proportion of individuals with technical education and a slightly higher proportion of workers (12% of the workers in the sample for a cluster representing under 9% of the sample).

35Cluster 4 (8% of the sample) tend to have low income and education levels (15% of the individuals with a primary education level), with a majority of middle-aged individuals. Workers and craftsmen are overrepresented. Cluster 5 is composed of young (49% of the cluster under 40) highly educated (44% with tertiary education) individuals working in the civil service (19% of the cluster). Cluster 6 (6% of the sample) is mostly composed of affluent (it regroups 14% of the top income decile) and highly educated (13% of the sample individuals with a university degree higher than bachelor) individuals. Cluster 7 has a high proportion of women (58% of the cluster), civil servants (21% of the cluster) and well-off highly educated (43% of the cluster have a tertiary education) individuals.

36Cluster 8 regroups a higher-than-average proportion of young (42% of the cluster under 40), low-income (74% of the cluster below median household income) women (64% of the cluster) with a low level of education (73% of the cluster have at most a secondary education). Cluster 9 regroups proportionately more senior (49% of the cluster over 55) men (60% of the cluster) with a higher income (61% of the cluster above median household income) than the average population. Cluster 10 are very similar to cluster 9 in this respect but possess a higher-than-average education level (42% of the cluster have a tertiary education). Cluster 11 is characterised by a high proportion of seniors (55% of the cluster over 55) who tend to give no answer to many questions. Cluster 12 regroups young (55% of the cluster under 45) and not necessarily highly educated individuals (59% of the cluster have at most a secondary education).

37Cluster 13 has a high proportion (62%) of women, young (27% under 25) and old (35% over 65) as well as religious (19% of the cluster) individuals. Most individuals of this cluster have a low level of education (70% of the cluster with at most secondary education) and income. Cluster 14 have a rather low income (40% of the cluster in the bottom three deciles of household income distribution) in spite of not having a low level of education (29% of the cluster have a tertiary education). Civil servants are particularly represented (20%) in this cluster.

38Cluster 15 gathers older (67% of the cluster over 55), religious (22% of the cluster) and relatively well-off individuals (27% of the cluster with wealth over 300,000 euros). Cluster 16 are more difficult to define but they are more religious than the average (18% of the cluster). There are relatively few younger individuals in cluster 17 (only 11% of the cluster under 40). The cluster has a high (31%) proportion of self-employed. Cluster 18 regroups well-off (55% of the cluster in the top three deciles of household income distribution) and educated (39% of the cluster have a tertiary education) individuals.

39In order to come to a general characterisation of the cluster grouping, the third step of the three-step model mentioned above is performed. Different variables likely to influence cluster membership are considered. Individual characteristics such as sex and age were included in the variables possibly explaining group membership. In addition, income, education, occupation and religious participation (the respondent attends a religious service at least twice a month) were considered. Various indicators for the education level were considered for inclusion in the model. Elimination of non-significant variables led to keep indicators for technical education and higher education. Occupation variables proved to be jointly insignificant with the exception of the indicator for civil servants. The results are presented in table 2 and broadly confirmed the characterisation obtained by considering the statistics of cluster composition that were exposed above.

Table 2. Predictors of class membership

Classes

Woman

Under 40

Over 65

Household income per person

Technical education

Higher education

Civil servant

Religious

1

-0.04

0.01

-0.41

2.19

-0.58

0.69

0.37

-0.90

0.41

0.12

2.77

2.34

4.28

5.93

3.62

2.59

2

0.29

-0.23

-0.45

-3.42

0.22

-0.31

-0.01

0.10

2.77

1.98

2.58

1.97

2.07

2.20

0.03

0.50

3

0.01

-0.12

0.05

-1.88

0.29

-0.10

0.20

-0.45

0.01

0.86

0.40

1.09

2.71

0.72

1.25

1.49

4

-0.14

-0.41

-0.22

-5.76

0.08

-0.74

-0.06

-0.03

1.42

3.25

1.79

2.59

0.82

3.07

0.35

0.16

5

0.01

0.38

-0.55

-0.98

0.05

0.35

0.39

-0.78

0.08

2.31

1.96

0.51

0.33

2.96

2.78

1.21

6

-0.07

0.11

0.02

3.88

0.04

0.69

-0.21

-0.11

0.61

0.93

0.13

4.16

0.32

5.45

1.33

0.54

7

0.26

0.14

-0.11

3.17

0.20

0.28

0.35

-0.23

2.22

1.03

0.68

2.81

1.65

2.15

2.46

0.79

8

0.42

0.35

0.12

-4.26

0.10

-0.32

0.26

0.10

2.94

2.44

0.73

2.51

0.87

1.92

1.54

0.54

9

-0.26

0.01

0.32

3.88

0.22

0.21

0.03

-0.15

2.26

0.06

2.42

4.07

1.66

1.55

0.19

0.62

10

-0.25

0.16

0.29

3.84

-0.13

0.32

-0.41

0.11

2.21

1.12

2.15

3.84

0.96

2.47

1.77

0.65

11

-0.03

0.13

0.39

-3.04

0.02

-0.38

-0.34

0.16

0.24

0.82

2.54

1.01

0.18

1.77

1.28

0.91

12

-0.02

0.39

0.00

4.16

0.19

-0.47

-0.45

0.22

0.14

2.74

0.02

2.65

1.37

2.48

1.17

1.04

13

0.21

0.40

0.35

-1.10

-0.18

-0.33

-0.18

0.38

1.84

2.68

2.37

0.69

1.59

1.87

0.83

2.68

14

-0.01

0.19

-0.23

-4.52

-0.15

0.17

0.34

-0.04

0.07

1.42

1.24

2.57

1.27

1.19

2.17

0.20

15

-0.10

-1.08

0.22

4.37

-0.33

0.29

-0.77

0.56

0.77

2.38

1.51

4.74

2.07

1.74

1.96

3.43

16

-0.20

0.15

0.20

-3.43

-0.01

0.09

0.08

0.44

1.10

0.67

0.89

0.83

0.07

0.45

0.32

2.02

17

0.04

-0.77

0.12

-1.97

0.01

-0.96

0.13

0.21

0.28

2.01

0.64

1.13

0.08

1.61

0.53

0.91

18

-0.12

0.16

-0.11

4.87

-0.05

0.51

0.28

0.41

0.61

0.76

0.36

4.92

0.24

2.50

1.25

1.67

Note: Z-value under the coefficient.

Source: Bruno Amable, 2018

5.3. Political self-placement

40The political hue of each group may be assessed with the help of the self-placement of their members on a left-right scale (0 to 10). For the sake of simplification, this was re-expressed in five categories: far left (0 to 2), left (3 and 4), centre (5), right (6 and 7) and far-right (8 to 10). Table 3 presents the results of a regression of individuals’ political position on cluster membership. These may be summed up as follows. Clusters 1, 5, and 8 have broad left leanings. Clusters 2 and 3 are somewhat more centre-left than cluster 1. Clusters 4 and 11 appear split but the weight of right and far-right is larger (40% of the group) than left and far-left (29%) in cluster 4. Cluster 6 are centre-left. Clusters 7 and 12 are centrists. Clusters 9, 10, 15, and 18 are strongly right. Unsurprisingly, cluster 13 express no clear opinion. In spite of having a strong component of the far left, the opinions of cluster 16 appear widespread. Cluster 14 are far left, cluster 17 far right.

Table 3. The influence of self-placement on cluster membership

Cluster

Far left

Left

Centre

Right

Far right

1

1.24

0.74

-0.81

-0.97

-3.42

7.80

6.06

-3.69

-3.19

-13.49

2

0.63

0.32

0.22

-0.06

0.13

3.53

2.36

1.98

-0.41

0.51

3

0.86

0.17

0.30

-0.26

-0.02

4.97

1.16

2.74

-1.64

-0.07

4

0.46

-0.32

0.13

-0.41

1.16

2.47

-1.73

1.22

-2.42

5.70

5

0.63

0.79

-0.15

-0.23

-0.75

3.32

5.74

-0.95

-1.34

-1.32

6

-0.07

0.78

0.26

-0.09

-1.83

-0.25

5.81

2.29

-0.58

-0.74

7

-0.17

0.05

0.53

0.17

0.32

-0.59

0.28

4.92

1.34

1.34

8

0.51

0.30

0.16

-0.07

0.44

2.53

1.99

1.19

-0.46

1.84

9

-3.15

-0.60

-0.21

0.96

0.89

-16.52

-2.19

-1.30

8.40

4.07

10

-1.40

-0.13

-0.01

0.59

0.96

-1.23

-0.70

-0.08

5.19

4.35

11

-0.02

0.35

0.01

-0.02

0.63

-0.06

2.25

0.09

-0.12

2.68

12

0.32

0.26

0.31

0.23

0.15

1.37

1.62

2.39

1.69

0.50

13

0.29

0.09

0.23

-0.12

-0.18

1.32

0.52

1.90

-0.74

-0.52

14

1.79

-0.01

-0.70

-0.86

-2.32

9.35

-0.07

-2.52

-2.45

-1.31

15

-1.72

-1.95

-0.11

0.58

1.28

-0.87

-1.27

-0.67

4.53

5.84

16

0.74

0.12

0.00

0.12

0.15

3.09

0.54

0.01

0.61

0.41

17

-0.44

-0.71

-0.39

0.06

1.69

-0.83

-1.72

-1.47

0.28

6.90

18

-0.49

-0.25

0.23

0.36

0.71

-0.77

-0.77

1.11

1.71

2.44

Note: Z-value under the coefficient.

Source: Bruno Amable, 2018

6. Aggregating a social bloc

6.1. The possible social blocs

41From these results and considering the policy preferences, the left bloc can be said to include clusters 1, 2, 3, 5, 8 and 14. The right bloc comprises clusters 4, 9, 10, 15, 17 and 18. Other clusters can be considered either as centre-left or centre-right (6, 7 and 12), or neither specifically left nor right (11, 13 and 16). These clusters are at the periphery of the left and right blocs and could join one or the other depending on the political strategies followed by conservative or social-democratic parties or durably stay outside of either bloc. Table 4 shows the partition of clusters according to the social blocs to which they could be aggregated, their appreciation of neoliberal reforms and their position with respect to European integration.

Table 4. The clusters and the divides on European integration and neoliberal reforms

Neoliberal reforms

Blocs

Pro

Unclear/split

Anti

Left

Pro-EU

1

Unclear

14

Anti-EU

8

2, 3, 5

Centre-left or centre-right

Pro-EU

6

Unclear

7

Anti-EU

12

Right

Pro-EU

10, 15

Unclear

9, 17

Anti-EU

18

4

Unclear

Pro-EU

16

Unclear

11

13

Anti-EU

Source: Bruno Amable, 2018

42As mentioned previously, the European integration issue splits both the left and right blocs. Regressing the answers to two questions regarding respectively whether the EU is a risk for social protection and national identity on group membership, one obtains the results documented in table 4. On the left, cluster 1 are pro-EU partly because they do not think that it threatens either the social model or the national identity. Cluster 14 do not think the EU is a threat to the national identity. For the other clusters of the left bloc, on the other hand, the EU represents a threat to both social protection and national identity (2 and 3) or simply to social protection (5). Such risks are also perceived by some clusters of the right bloc: loss of national identity (groups 9 and 17).

Table 5. Risks associated with European integration (logit estimation)

EU is a risk for

Cluster

Social protection

National identity

1

-0.13

-0.89

-1.52

-7.43

2

0.33

0.27

2.75

2.74

3

0.42

0.29

3.16

2.79

4

0.39

0.74

3.43

5.93

5

0.34

-0.18

2.57

-1.71

6

-0.17

-0.70

-1.72

-5.26

7

0.04

-0.04

0.36

-0.38

8

0.17

0.34

1.41

2.89

9

0.01

0.26

0.04

2.29

10

-0.60

-0.04

-5.29

-0.39

11

-0.05

0.11

-0.43

0.93

12

0.42

0.67

2.67

4.21

13

-0.09

-0.19

-0.80

-1.69

14

0.18

-0.35

1.37

-2.87

15

-0.57

-0.23

-4.58

-1.89

16

-0.57

-0.09

-3.36

-0.53

17

-0.03

0.34

-0.16

2.04

18

-0.11

-0.31

-0.55

-1.60

Note: Z-value under the coefficient.

Source: Bruno Amable, 2018

6.2. The possibility of a bloc bourgeois

43The possibility of aggregating a social bloc around the key policy options of the continuation of European integration and the implementation of neoliberal structural reforms can be assessed. Figure 1 summarises the respective positions of the 18 classes on the policy issues considered as indicators for the latent class analysis in a matrix. One can then assess the profile of each cluster according to policy preferences. Positions that reflect a policy preference compatible with a political strategy aiming for a bloc bourgeois are represented with a plain (yellow) cell. Policy preferences antagonistic with such a strategy are represented with a hatched cell. Blank cells represent neutral or undecided positions.

Figure 1. Policy preferences

Figure 1. Policy preferences

Note: plain (yellow) = for, hatched = against, blank = split or undecided.

Source: Bruno Amable, 2018

Cluster 6 is the key social group for the formation of a bloc bourgeois because this group are in favour of European integration as well as significant neoliberal reforms (pensions, privatisations, competitiveness). What other groups could possibly be aggregated to cluster 6 in order to form a bloc bourgeois?

44The aggregation of a social bloc hinges on the possibility to find a political strategy based on the satisfaction of policy demands sufficiently important for the groups potentially forming the bloc. The demands of these groups must be sufficiently compatible to guarantee the stability and political support of the bloc. Starting from cluster 6 as the core group, clusters 1, 15 and 16 share the same pro-European integration leaning. However, cluster 1 is strongly opposed to neoliberal reforms or policy orientations, making their integration in the prospective social bloc next to impossible. Clusters 16 and especially 15 on the other hand express policy demands compatible with such an integration. Moving to neoliberal policy orientations, cluster 7, split on the issue of further European integration, express demands close to those of cluster 6. More generally, taking clusters broadly in favour of neoliberal economic policy and structural reforms and favourable or not opposed to European integration, on could obtain a new social bloc gathering groups 6, 7, 9, 10, 15, 16 and 17, which would represent roughly 26% of the sample. Some groups are sometimes more favourable to neoliberal economic policy than cluster 6. The extension of the social bloc beyond the core social group would therefore probably require the elaboration of a political strategy oriented towards labour market and social protection reforms.

  • 27 One may note that this bloc is not particularly “culturally progressive” if one is to judge by the (...)

45A significant characteristic of this bloc is that contrary to the traditional left and right blocs, it would exclude the popular classes and have core groups with a high level of education and income. Considering the sociological composition (age, income, education), cluster 7 are not too dissimilar to cluster 6 and reinforce the bourgeois character of the bloc. Clusters 9, 10, 15, and 16 regroup individuals with high income or education level too. These groups would then correspond to the bloc bourgeois27 identified in Bruno Amable and Stefano Palombarini’s chapter of Hideko Magara’s publication (2014).

Conclusion

46The presidency of François Hollande represented a key moment for the transformation of the French socio-economic model and the associated socio-political (dis)equilibrium. The gradual breakup of the social blocs that had structured political competition during most of the Fifth Republic was completed, and this made possible the victory of an outsider with a “movement” formed barely one year before the 2017 elections whose social base was composed of the better-off segments of the traditional left and right social blocs. In the first round of the 2017 presidential election, Emmanuel Macron obtained his best results with the skilled workforce and professionals (37%) and his weakest score with workers (15%). He scored 35% of the voters with a university degree but only 17% of those under the secondary education level, 14% with voters whose monthly household income is under 1250 euros, but 32% for those above 3000 euros (Ipsos & Sopra-Steria, 2017). By promoting the key issues for the formation of the bloc bourgeois, Emmanuel Macron was able to exploit an opportunity created by the emergence of the European integration issue as a key political divide (Amable & Palombarini, 2018).

  • 28 A key element in Macron’s victory in 2017 was the particular political institutions of France and t (...)

47The analysis performed in this paper has shown that the core of the bloc bourgeois is rather narrow and therefore extending the bloc in order to integrate other social groups is a political necessity. The analysis suggests that this extension would not be possible, at least not to a significant extent, with “nonbourgeois” groups of the former left bloc. An extension on the right appears more feasible, based on the satisfaction of demands for the implementation of neoliberal reforms. Yet even such an extension would not guarantee a majoritarian social alliance28. This seems nevertheless the option chosen by Macron, whose economic and social policy options resemble those of the traditional right parties.

  • 29 A study of Macron’s party LREM (Terra Nova, 2018) reveals the existence of several fractions. The s (...)

48The stability of the bloc bourgeois depends crucially on the acceptability of neoliberal reforms by the groups of that social alliance. As shown by the empirical analysis, there exist potential conflicts within the extended bloc bourgeois on the area subject to “reforms” and on the extent of the transformations. Unification of the bloc bourgeois on these issues would make it increasingly resemble the more affluent part of the traditional right bloc. Another possible issue is immigration. A “liberal” attitude in this respect is important for the core bloc bourgeois but possibly antagonises some groups of the former right bloc that would be necessary for the extension of the bloc. On the other hand, too repressive an attitude in this respect would be in contradiction with the expectations of the bourgeois part of the former left bloc29. Nevertheless one should not overestimate the “progressive” or “culturally liberal” values of the bloc bourgeois. Our analysis has shown that the classes of the bloc bourgeois were not particularly culturally liberal in comparison to other classes. Gilles Finchelstein (2018) showed that a large fraction of Macron’s constituency (46%) were in favour of having as a head of state a “strong man” that would not have to worry about elections or the parliament. Of all supporters of other parties, only those of Front national were more in favour of that option (55%).

49Finally, an analysis such as that performed in this contribution is highly dependent on the quality of the information contained in the data sources used. The good quality of the questionnaire of the 2012 electoral survey made the analysis meaningful because this questionnaire included questions concerning economic policies and reforms. In this respect, the 2017 survey is more than disappointing. An extension of the analysis performed here would require precise questions regarding the policy expectations and demands for or against “reforms” that individuals may have, particularly with respect to the five institutional areas defining a socio-economic model (Amable, 2003): the employment relationship, the product markets, the financial system, the social protection and the education sector. A questionnaire including all these elements could be of great help for the analysis of the sustainability of socio-economic models.

Haut de page

Bibliographie

Amable B. (2003), The Diversity of Modern Capitalism, Oxford, Oxford University Press.

Amable B. (2011), « Morals and politics in the ideology of neoliberalism », Socio-Economic Review, vol. 9, no 1, p. 3-30.

Amable B. (2014), « Who wants the contrat unique? Social support for labour market flexibility in France », Industrial Relations: a journal of economy and society, vol. 53, no 4, p. 636-662.

Amable B. (2017), Structural Crisis and Institutional Change in Modern Capitalism: French capitalism in transition, Oxford, Oxford University Press.

Amable B. & S. Palombarini (2005), L’économie politique n’est pas une science morale, Paris, Raisons d’agir, coll. « Cours et travaux ».

Amable B. & S. Palombarini (2009), « A neorealist approach to institutional change and the diversity of capitalism », Socio-Economic Review, vol. 7, no 1, p. 123-143.

Amable B. & S. Palombarini (2014), « The bloc bourgeois in France and Italy », in Magara H. (ed.), Economic Crises and Policy Regimes: the dynamics of policy innovation and paradigmatic change, Cheltenham, Edward Elgar, p. 177-216.

Amable B. & S. Palombarini (2018), L’illusion du bloc bourgeois. Alliances sociales et avenir du modèle français, nouvelle éd. actualisée et augmentée, Paris, Raisons d’agir, coll. « Raisons d’agir ».

Amable B., Guillaud E. & S. Palombarini (2012), « Changing French capitalism: political and systemic crises in France », Journal of European Public Policy, vol. 19, no 8, p. 1168-1187.

Amar C. (2014), Jusqu’ici tout va mal, Paris, Grasset.

Armand L. & V. Giscard d’Estaing (1969), Quel avenir pour l’Europe ? Dialogue entre Louis Armand et Valéry Giscard d’Estaing, Paris, Publicis conseil.

Atkinson W. (2015), Class, Cambridge, Polity Press, coll. « Key concepts ».

Berman S. & M. Snegovaya (2019), « Populism and the decline of social democracy », Journal of Democracy, vol. 30, no 3, p. 5-19.

Bornschier S. & H. Kriesi (2013), « The populist right, the working class, and the changing face of class politics », in Rydgren J. (ed.), Class Politics and the Radical Right, London, Routledge, coll. « Routledge studies in extremism and democracy », p. 10-31.

Bourdieu P. & L. Boltanski (2008) [1976], La production de l’idéologie dominante, Paris, Demopolis, Raisons d’agir.

Dardot P. & C. Laval (2009), La nouvelle raison du monde. Essai sur la société néolibérale, Paris, La Découverte.

Finchelstein G. (2018), Portrait-robot des sympathisants de La République en marche, Paris, Fondation Jean Jaurès.

Ford R. & W. Jennings (2020), « The changing cleavage politics of Western Europe », Annual Review of Political Science, vol. 23, p. 293-314.

Fulla M. (2016), Les socialistes français et l’économie (1944-1981). Une histoire économique du politique, Paris, Presses de Sciences Po, coll. « Académique ».

Gauron A. (1983), Histoire économique et sociale de la Cinquième République, t. 1 Le temps des modernistes, Paris, La Découverte, Maspero, coll. « Économie critique ».

Gauron A. (1988), Histoire économique et sociale de la Cinquième République, t. 2 Années de rêves, années de crises (1970-1981), Paris, La Découverte, coll. « Cahiers libres ».

Giscard d’Estaing V. (1984), Deux Français sur trois, Paris, Flammarion.

Hout M., Clem B. & J. Manza (1993), « The persistence of classes in post-industrial societies », International Sociology, vol. 8, no 2, p. 259-277.

Inglehart R. (1990), Culture Shift in Advanced Industrial Society, Princeton (NJ), Princeton University Press.

Inglehart R. & S.C. Flanagan (1987), « Value change in industrial societies », American Political Science Review, vol. 81, no 4, p. 1289-1319.

Ipsos & Sopra-Steria (2017), 2017 élection présidentielle. 1er tour : sociologie des électorats et profils des abstentionistes, Ipsos. URL: https://www.ipsos.com/fr-fr/1er-tour-presidentielle-2017-sociologie-de-
lelectorat
[accessed on 1 August 2021]

Jospin L. (2000), Le socialisme moderne, Paris, Fondation Jean Jaurès, coll. « Notes de la Fondation Jean Jaurès ».

Kitschelt H. (1993), « Class structures and social democratic party strategies », British Journal of Political Science, vol. 23, no 3, p. 299-337.

Kriesi H. (1998), « The transformation of cleavage politics: the 1997 Stein Rokkan lecture », European Journal of Political Research, vol. 33, p. 165-185.

Kriesi H. (2010), « Restructuration of partisan politics and the emergence of a new cleavage based on values », West European Politics, vol. 33, no 3, p. 673-685.

Lamy P. (2005), « Le modèle français vu d’Europe. Entretien », Le Débat, no 134, p. 31-41.

Lehingue P. (2007), « Le Non français au traité constitutionnel européen (mai 2005). Sur deux lectures “polaires” du scrutin », Actes de la recherche en sciences sociales, no 166-167, p. 123-139.

Lipietz A. (1984), L’audace ou l’enlisement. Sur les politiques économiques de la gauche, Paris, La Découverte, coll. « Cahiers libres ».

Lipietz A. (1994), « Post-Fordism and democracy », in Amin A. (ed.), Post-Fordism: a reader, Oxford, Blackwell, coll. « Studies in urban and social change », p. 338-357.

Lordon F. (1998), « The logic and limits of désinflation compétitive », Oxford Review of Economic Policy, vol. 14, no 1, p. 96-113.

Nylund K.L., Asparouhov T. & B.O. Muthén (2007), « Deciding on the number of classes in latent class analysis and growth mixture modeling: a Monte Carlo simulation study », Structural Equation Modeling: a multidisciplinary journal, vol. 14, no 4, p. 535-569.

Oesch D. (2006), Redrawing the class map: stratification and institutions in Britain, Germany, Sweden and Switzerland, Basingstoke, Palgrave Macmillan.

Oesch D. (2008), « The changing shape of class voting », European Societies, vol. 10, no 3, p. 329-355.

Piketty T. (2018), Brahmin Left vs Merchant Right: rising inequality and the changing structure of political conflict, World inequality database working paper, no 2018/7.

Piketty T. (2019), Capital et idéologie, Paris, Le Seuil, coll. « Les Livres du nouveau monde ».

Sauger N., Brouard S. & E. Grossman (2007), Les Français contre l’Europe ? Les sens du referendum du 29 mai 2005, Paris, Presses de Sciences Po, coll. « Sciences Po. Fait politique ».

Sauger N. (2012), French Electoral Study 2012, TNS-Sofres. URL : https://cdsp.sciences-po.fr/en/ressources-en-ligne/ressource/fr.cdsp.ddi.FES2
012/
[accessed on 1 August 2021]

Terra Nova (2011), Gauche : quelle majorité électorale pour 2012 ? Projet 2012 – Contribution no 1, Paris, Terra Nova.

Terra Nova (2018), La République En Marche : anatomie d’un mouvement, Paris, Terra Nova.

Haut de page

Annexe

Table A1. The 18-cluster latent class model

Number of individuals

2014

Number of parameters

1097

Degrees of freedom

917

Classification errors

0,1657

Reduction of errors

0,8153

Entropy R-squared

0,8367

Standard R-squared

0,7441

Source: Bruno Amable, 2018

Table A2. Likelihood-ratio test. Test 18 vs. 19 classes

Bootstrap

Log-likelihood Statistics

-2LL Diff

p-value

s.e.

CV

Log-likelihood (LL)

-34193,3673

40,9271

0,2040

0,0180

88,3383

Source: Bruno Amable, 2018

Conclusion : reject the 19-cluster model.

Haut de page

Notes

1 For a definition of neoliberalism, see Pierre Dardot and Christian Laval’s book (2009) and Bruno Amable’s paper (2011). In what follows, the term will refer to the attempts to change the institutions of the French socio-economic model to make them more similar to those of a neoliberal model of capitalism as defined in Amable’s work (2003): flexible labour markets, privatisation of public utilities, education and social protection, liberalisation of product and financial markets, etc.

2 I avoid to use the word “class”. There is a plethora of definitions of social classes in the literature, be it in sociology, political science, economics or any other social science, and they do not necessarily correspond to the same object nor are they necessarily compatible with one another (see for instance Will Atkinson’s work, published in 2015). The use of the word “class” has been limited as much as possible in order to avoid confusion among readers who may link this word to their preferred meaning. As Pierre Bourdieu once said, even taking a position on the issue of social classes depends on one’s position in the class structure.

3 This does not imply that all demands of all agents in a social group are perfectly identical.

4 One could distinguish between social groups, defined by a proximity of position in the socio-economic structure, from socio-political groups, to which corresponds a proximity in the policy demand space. For the sake of simplification, this distinction is ignored.

5 For a typology of models of capitalism, see The Diversity of Modern Capitalism by Amable (2003).

6 Chirac, who had once declared that he was looking for a French style Labour party (travaillisme à la française), promised that when the right would come back to power, it would implement a “rather brutal breakup” with the existing economic model.

7 How radical Sarkozy would have been is uncertain considering that the 2008 crisis compelled him to change his initial plans, leading him to declare that the age of laisser faire was over.

8 Roughly one third of PS in 1979 (Fulla, 2016).

9 The Programme commun de gouvernement elaborated with the communist party and PS’ satellite Radicaux de gauche.

10 Lamy P. (2005), « Le modèle français vu d’Europe. Entretien », Le Débat, no 134, p. 31-41.

11 Le Figaro, 24 mars 2005. Original citation : « l’Europe a été notre programme d’ajustement structurel […] la France a fait jouer à l’intégration européenne le rôle que les pays mal gouvernés confient au FMI ou à la Banque mondiale. »

12 For a discussion of Europe as a new political divide, see Les Français contre l’Europe ? by Nicolas Sauger, Sylvain Brouard and Emiliano Grossman (2007).

13 The Terra Nova report made more or less explicit reference to some political science analyses focusing on the cultural divides (Inglehart, 1990) or based on a two-dimensional view of the political space: economic and cultural (Kitschelt, 1993).

14 When in 2013 a journalist pointed out to Hollande that his decision to let Mittal Steel mothball the Florange blast furnaces would entail significant political costs, Hollande simply replied: “among blue-collar workers, yes, but it does not matter” (Amar, 2014, p. 65).

15 This bloc is characterized by the absence of working classes, which were present in the traditional left and right blocs.

16 The traditional left/right cleavage is expressed in terms of economic policy preferences, redistribution and public spending vs. low taxes, or regulation of economic activities vs. market liberalisation for instance. There is a huge political science literature on the left-right cleavage and the emergence of “new” cleavages. For an overview, see Robert Ford and Will Jennings’ paper (2020).

17 Giscard d’Estaing is one of the very few politicians to have claimed the label: “The most scientific form of modern economic thought is the liberal thought [… it is necessary] to give it a modern name: neo-liberalism” (Armand & Giscard d’Estaing, 1969, cited in Bourdieu & Boltanski, 2008, p. 33).

18 For all questions, there was a possibility of no answer.

19 The 2012 survey included some questions on economic policies and institutional reforms which made possible the exercise performed in this paper. The 2017 electoral survey on the other hand did not include questions on policy preferences regarding important reform areas such as the labour market or the social protection system. For the most part, it favoured questions on respondents’ opinions on various topics (robotisation, climate change, inequality…) over questions on the approval or disapproval of specific policy measures. For instance, a question asks respondents who they think were the victims of the economic crisis. This is a question that may reflect the vision that respondents have of the economic crisis, but not what their policy demands are.

20 For instance Manuel Valls, Hollande’s prime minister.

21 For this question and others concerned, the initial range of response from 0 to 10 was re-expressed as going from 0 to 5 or 6 depending on the profile of responses.

22 These issues are still at the centre of the 2020 pension reform.

23 See the paper by Sheri Berman and Maria Snegovaya (2019).

24 The test provides information on the improvement that an additional cluster provides to the model. If the difference in model fit (measured with the squared likelihood L²) is not significant, the more parsimonious model is chosen. If it is significant, the less restricted model is chosen instead. The test is based on the reduction in L² that can be expressed in terms of twice the increase in log-likelihood LL associated with the increase in number of clusters from n to n+1. The associated p-value is estimated by means of a parametric bootstrap, whereby the model is estimated for a certain number (here 500) of replication samples (the so-called Monte Carlo simulation) that are generated from the probability distribution defined by the maximum likelihood estimates of the restricted (n-cluster model). The estimated bootstrap p-value is defined as the proportion of bootstrap samples with a larger -2LL-difference value than the original sample.

25 The models were estimated with Latent Gold 5.0.

26 The income variable selected is the level of total household income divided by the square root of the number of household size.

27 One may note that this bloc is not particularly “culturally progressive” if one is to judge by the attitude regarding the right to adopt children by gay couples.

28 A key element in Macron’s victory in 2017 was the particular political institutions of France and their presidential character. The outcome of the legislative elections is strongly dependent on that of the presidential election. This, and a particularly low participation rate, made it possible for La République en marche (LREM) to obtain an absolute majority in parliament while representing only 15% of the electorate in the first round.

29 A study of Macron’s party LREM (Terra Nova, 2018) reveals the existence of several fractions. The so-called “liberal progressive” holding economic neoliberal and culturally progressive values represent slightly less than a third of the party members, being outnumbered by the (liberal- and moderate-) “conservatives” that represent together 42% of LREM membership.

Haut de page

Table des illustrations

URL http://journals.openedition.org/regulation/docannexe/image/20350/img-1.png
Fichier image/png, 6,7k
Titre Figure 1. Policy preferences
Légende Note: plain (yellow) = for, hatched = against, blank = split or undecided.
Crédits Source: Bruno Amable, 2018
URL http://journals.openedition.org/regulation/docannexe/image/20350/img-2.png
Fichier image/png, 56k
Haut de page

Pour citer cet article

Référence électronique

Bruno Amable, « In search of the bloc bourgeois »Revue de la régulation [En ligne], 31 | 2nd semestre | Autumn 2021, mis en ligne le 17 janvier 2022, consulté le 19 avril 2024. URL : http://journals.openedition.org/regulation/20350 ; DOI : https://doi.org/10.4000/regulation.20350

Haut de page

Auteur

Bruno Amable

Professor of political economy, department of History, Economy and Society, University of Geneva, Unimail, boulevard du Pont d’Arve 40, 1205 Geneva; bruno.amable@unige.ch

Articles du même auteur

Haut de page

Droits d’auteur

CC-BY-NC-ND-4.0

Le texte seul est utilisable sous licence CC BY-NC-ND 4.0. Les autres éléments (illustrations, fichiers annexes importés) sont « Tous droits réservés », sauf mention contraire.

Haut de page
Search OpenEdition Search

You will be redirected to OpenEdition Search