Varieties of capitalism and pension schemes : determinants of change in the case of France

(1)

Varieties of capitalism and pension schemes:

determinants of change in the case of France.

Thesis submitted in partial fulfilment of the requirements for the degree of MSc in

Political Economy of the University of Amsterdam’s Graduate School of Social

Sciences

June 2019

Scott L. Santer

Id: 12296139

(2)

2

Abstract

The central focus of this paper is the examination and explanation of why the French political economy set-up partial capitalisation of its Pay-As-You-Go general pension regime in 1999, and why the first occupational and personal pre-funded pension schemes were set-up in 2003. A theoretical framework is elaborated which shows why, on the one hand, these reforms are paradoxical in relation to French comparative institutional advantages, while on the other hand legitimise a State-centred approach to undertake analysis. The method employed to operationalise this approach is a systematic tracking of the process leading to the reforms through the analysis of parliamentary debates as well as law propositions, commission reports and inter-chamber notes. In-so-doing I show that the 1999 capitalisation stemmed from a Left-wing initiative which was in itself a reaction to an attempt by the Right-wing to entirely privatise welfare. Furthermore, I show that the first occupational and personal pre-funded schemes were set-up mainly to enlarge the pool of domestic capital in France in order to minimise the significance of foreign capital on the Parisian CAC-40 on the one hand, and in order to decentralise industrial relations on the other.

(3)

3

1. Introduction

In developed political economies, the origins of systems for retirement income provision date back more than a century. Arguably, governments of the time began to establish mandatory social insurance against social risk in order to conciliate workers and capitalists and provide social stability (Jessop 1990). Over time, social policies have become more and more closely linked, though always implicitly, to market regulation (Schelkle 2012). This research paper investigates these implicit links by analysing the causalities behind a puzzling set of reforms implemented between 1999 and 2003 in France. These reforms are relevant for research in that they constitute a bifurcation from the social insurance tradition in France.

Today, all social insurance systems have in common a basic scheme for poverty alleviation through taxes. This is true of whole social insurance models as well as their pension component. In this realm of retirement income, States show different levels of implication and generosity, in many cases going further than minimal poverty prevention by collecting additional social contributions from citizens. Nations which only provide the bare minimum for poverty prevention are liberal, mostly Anglo-Saxon, political economies which leave large developmental potential for supplementary private schemes to be set-up. A more nuanced public-private mix characterises continental European political economies, where a number of State-led social insurance models go further by guaranteeing, through taxes and redistributive policies known as Pay-As-You-Go (PAYG) systems, universal income security to all citizens. Going further for pension provision, some redistributive welfare regimes define benefits based on previous earnings, tending to reinforce status differences (Palier 2010). The Liberal model of minimum public welfare induces increased self-reliance for retirement income. This model has progressively made privately managed pre-funded pension schemes the predominant arrangement for pension provision. In the recent past, this has led some authors to go as far as labelling the concerned capitalist system (the United States in this case) ‘pension fund capitalism’ or ‘socialism’ because of the theoretical increased power they give to contributors; i.e. workers (e.g. Drucker 1976; Clark & Hebb 2004). Save for Switzerland and the Netherlands, the pre-funded scheme is characteristic of Anglo-Saxon political economies – in particular the United States and United Kingdom – where assets under management of independent pension funds have higher aggregate value than the GDP of the respective countries.1 This significant size of pension funds in Anglo-Saxon political economies is symptomatic of a strong degree of ‘financialisation’ whereby financial motives, markets, actors and organisations are important components of the political economy (Epstein 2006). In Continental Europe, France for example, the predominant model is traditionally one of constant redistribution (the PAYG model) where workers’ contributions finance the pensioners’ benefits. The PAYG model is resilient to change because of how difficult it is to rewrite the intergenerational

(5)

5

contract that they imply (Pierson 2001). Furthermore, an important strand of academic study shows how, theoretically, each model induces complementary, self-reinforcing, effects on other institutional clusters such as financial systems, industrial relations and corporate governance regimes (e.g. Vitols 2005). In turn, these complementarities induce comparative institutional advantages which characterise political economic Varieties of Capitalism (VoC) (Hall & Soskice 2001).

The ability to go further than examining the resilience of the PAYG model has become necessary in light of the visible policy change away from the model that a number of political economies have put in place. Indeed, contemporary Continental European economies are more and more characterised by a dynamic of change from PAYG to privatised, pre-funded pension systems (Natali & Stamati 2013). The question of how this general liberalising trend is taking place is discussed extensively (e.g. Streeck & Thelen 2005). Going one step further, academic debates concerning the consequences and country-specific modalities of liberal changes in traditional PAYG systems are also numerous (e.g. Dixon 2008). Yet these academic accounts often struggle to, or simply avoid explaining why pension system liberalisation is taking place or how it is legitimised. This is mainly due to the difficulty researchers face when untangling exactly how a given institutional cluster such as social insurance and its pension component interacts with other institutional clusters, and where causality for change between each is to be found (Höpner 2005). Therefore, many publications simply assume that liberalisation is taking place, and that the main subject of research should be the country-specific modalities of this type of change (e.g. Streeck & Thelen 2005; Ebbinghaus 2011; Palier 2010). I place myself upstream from this debate, deeming it necessary to identify precisely what triggers the first steps of change towards privatisation and capitalisation of pension systems in particular.

In order to do so, this paper uses a State-centred approach to analyse the first reforms constituting wholesale breaks from the traditional French PAYG pension system. After years of law proposals from the two traditionally dominant Right-wing parties Union pour la Démocratie Française (UDF) and Rassemblement pour la République (RPR) being refused, not implemented or abrogated, it is paradoxical that the first steps towards capitalisation and privatisation were taken by a majoritarian Left-wing government, followed by a cohabitation government with no clear majority, and finally were consolidated by a Right-wing government and majority. A variety of publications discuss different components of the general trend towards this type of reform (e.g. Dixon 2008; Ebbinghaus 2015). In the specific case of France, many publications discuss and can account for the reasons why the attempts prior to the successful implementations failed (e.g. Naczyk 2013; 2016). However, it remains puzzling why, reforms or additional legislation for capitalisation and/or privatisation succeeded in 1999 and 2003.

For this paper to be able contribute to understanding causal factors behind pension system capitalisation and privatisation, the research question ‘Why and how do reforms and/or legislation changes that appear to contradict a political economy’s social insurance tradition and comparative

(6)

6

institutional advantages succeed?’ will be central. In order to answer it, this paper is divided into six sections.

After this introduction section, a second section focuses on the presentation of the case studied and will provide necessary contextualisation to the research. Third, a theoretical framework will clarify how the traditional French pension system contributes to the country’s comparative institutional advantages and will show that the pension reforms hereby analysed have implications far beyond the social insurance realm. This literature, which gives the tools to understand how pension models affect other institutions and thus the macro-economic configuration of a given political economy, does not provide tools for understanding why pension models would change. An additional theoretical contribution is necessary to be able to do so. Therefore, literature on institutional change will be reviewed in order to precisely conceptualise the change taking place. With this done, it will be shown that literature on financialisation provides the best suited approach for this research: a State-centred analysis. Fourth, a methodology section will present how I undertook a State-centred analysis. A fifth section will present the results of my analysis. Finally, a sixth section will conclude.

In summary, I show that by analysing the reforms from the perspective of the State as a unitary actor, capitalisation occurred in order to consolidate the PAYG system, and privatisation occurred in order to make the pension system more sustainable by adding supplementary incentives for private savings. Results are a lot more informative when the process leading to reform is traced with partisan units of analysis. In this case, it is clear that traditional Right-wing parties are long-standing proponents of capitalised and privatised welfare, while the opposite is true for the Left. My analysis shows that the Left changed its position concerning the issue primarily because of foreign ownership on the Parisian CAC-40. This was the main element which enabled consensus to be reached, and planted the first seeds towards pre-funded pension schemes in the French political economy.

2. Case study: systemic changes in France’s general pension

system

The solid foundations for the contemporary French pension system were laid by the temporary government of the World War II aftermath (Palier 2010). The main characteristic of the system is its emphasis on solidarity and status maintenance while relying on PAYG financing. The system today is somewhat fragmented, with five components (i.e. regimes) in place (Sarfati & Ghellab 2012). The PAYG general regime (régime général) which applies to all private-sector wage-earners; the pre-funded public servants scheme (Préfon ; Retraite et Prévoyance de la fonction publique); special schemes (régime spéciaux) which apply to employees of State-owned enterprises or public utilities;2

2_{Apart from the teachers’ arrangement, all special regimes are organised on a PAYG basis. They concern}

employees of the Société Nationale des Chemins de Fers Français (SNCF), of the Régie Autonome des

Transports Parisiens (RATP), and of companies Energie de France (EDF) and Gaz de France (GDF) (Palier &

(7)

7

two mandatory PAYG pension schemes for all wage-earners except those covered by the special schemes;3 and since 2003, pre-funded pension schemes which can consist of individual or collective agreements, open to all wage-earners.4 Additionally, since 1994, the self-employed receive tax incentives for funded saving plans (Palier 2008). Research is required in order to understand a structural change of the general regime through partial capitalisation in 1999, and the creation of additional pre-funded schemes in 2003. The remainder of this section precisely presents these case studies.

In 1999, a Left-wing parliament majority, under a coalition government, passed a law which put in place a wealth fund – the Fonds de Réserve pour les Retraites (FRR) – in order to guarantee the long-term sustainability of the general regime. This partial capitalisation constitutes the first case study of this research. Although the reform is only of capitalisation, and not privatisation, it appears as a significant change to the traditional functioning of the general regime; by making it rely on capital returns in order to remain balanced. It is also worthy of study because of how it appears contradictory to the Left’s previous stances on capitalisation. Indeed, when previous occurrences of reform for capitalisation and privatisation were proposed by Right-wing parties and actors, the Left had always been able to contest, amend, or repeal them. For example, in 1987 a law making pension provision individual responsibility5 is passed by a RPR-UDF coalition, only to be repealed in 1989 by a Left-wing government defending the redistributive PAYG system of the general regime and supported to do so by unions of both workers and employers (Naczyk 2016: 210). A decade later, a similar attempt for privatisation and capitalisation is formally proposed and accepted by an RPR-UDF coalition,6 only to be repealed on similar grounds by the following government and parliament in 1998. These successive attempts and repeals show that pension reform is a highly contentious policy realm in France, where the Right has been pushing for capitalisation, privatisation and even marketisation for quite some time already, and where the Left has defended the status-quo by repealing attempts to do so. Therefore, all significant succesful reforms of the general regime before 1999 were centred around issues of legal retirement age, contribution periods, or benefit indexing (Amable et al. 2012).7 Clearly, then, the Left acted in apparent opposition to its previous stances on capitalisation by setting up the FRR in 1999. This is the first case studied to provide answers to the question this research addresses.

3_{Namely Association pour le Régime de Retraite Complémentaire des Salariés (ARRCO) for blue-collar}

workers and Association Générale des Institutions de Retraite des Cadres (AGIRC) for managerial staff.

4_{Namely, the collective agreement-induced Plan d’Epargne pour la Retraite Collective (PERCO) and the}

individual 3rd pillar-equivalent Plan d’Epargne Retraite Populaire (PERP).

5_{Namely, the Plan d’Epargne pour la Retraite (PER) put forward by the Chirac-Balladur government under}

President François Mitterand.

6_{Namely, the Loi Thomas.}

7_{In 1982, the legal retirement age is reduced from 65 to 60, with a minimum contribution period of 37.5 years to}

obtain full pension. In 1993, the minimum contribution period is extended to 40 years and pension benefits are indexed on prices instead of wages (Amable et al. 2012).

(8)

8

As the general overview of the French pension system indicates, it is in 2003 that supplementary pre-funded schemes are set up. This is the first successful legislation change towards capitalised and privatised retirement income schemes open to all citizens in France. This is the second case studied in order to provide answers to the question this research addresses. As hinted in the introduction, these reforms are major breaks from the French social insurance tradition, which itself has been part-and-parcel of the country’s Variety of Capitalism (VoC), hence to its comparative institutional advantages.

3. Theoretical framework

The following section reviews how to theoretically understand institutions and organisations. This enables an empirical study of the French VoC and its comparative institutional advantages, with emphasis given to where and how pension provision is to be situated. In order to properly conceptualise the reforms as occurrences of institutional change, it then reviews literature centred on such a topic. With the French political economy characterised and the reforms conceptualised, it finally reviews literature explicitly centred on institutional change towards financialisation and in-so-doing puts forward the appropriate approach to follow for analysis.

3.1. Institutional economics

Initial questioning of the real-world foundations of classical economic theory was made through works published at the end of the 19th and beginning of the 20th century mainly by Thorstein Veblen, John Commons (1932), and Ronald Coase (1937). Since then, the concepts of institutions and organisations in economic theory have become of paramount importance in explaining cross-national specificities in aggregate production and consumption regimes. In order to characterise the French political economy and its institutional configuration, I shall make use of the VoC framework established by Hall & Soskice (2001), which is itself an empirical acknowledgement of an array of theoretical contributions made by the Institutional Economics school of thought (Williamson 2007 [1985]; North 1990).

The initial insight provided by Commons (1932) is the fact that economics, as a field of research, are not a natural science; the economist’s task is not one of uncovering natural, quasi-divine, laws of economic functioning in the way the dominant classical theory posited at the time. To Commons (1932), any economic order is maintained by subjective arrangements found to reconcile conflicting interests. Around the same time, Coase (1937) shows that classical economic theory is based on the ill-founded assumption that price-movements suffice in organising production by market actors. If this was the case, production could be carried out without any additional productive organisation (i.e. firms; companies) at all. Coase (1937) argues that the reason they do exist lies in the fact that markets are not frictionless; there is a cost to any market transaction. The solution found to minimise these costs is vertical integration of production and the creation of organised firms. Going further in conceptualising and accounting for external costs to market transactions, Williamson (2007 [1985])

(9)

9

theorises the place of contractual costs. Together with transaction costs presented by Coase (1937), the theory of incomplete contracts made by Williamson (2007 [1985]) provides a first step towards understanding different modes of organisation that companies take (which Williamson (2007 [1985]) refers to as hierarchies). He does so by showing that different contractual specificities, such as geographical location of production relative to retail, physical specificities of the produced good, market size, reputation of producing firm, amongst others, influence ex ante and ex post contractual costs. In turn, these affect the strategic organisational choices of firms.

With the role and modalities of economic organisations theorised, the final theorisation of institutional determinants of organisational behaviour is made by North (1990). North (1990: 4) emphasises the need to consider institutions as both formal and informal constraints on organisational modalities. By focusing only on formal constraints, institutional change and divergent path-dependencies can hardly be understood. With this in mind and by explicitly differentiating institutions from organisations, North (1990: 73) sums up all previous institutional economics literature: institutions are the rules of the game, organisations are the players; “organizations [are] purposive entities designed by their creators to maximize wealth, income, or other objectives defined by the opportunities afforded by the institutional structure of the society. In the course of pursuing those objectives, organizations incrementally alter the institutional structure […] Organizations […] will be created as a function not simply of institutional constraints but also of other constraints (e.g., technology, income, and preferences). The interaction of these constraints shapes the potential wealth-maximizing opportunities of entrepreneurs (economic or political).”

As customs, traditions, and codes of conduct are the cultural constraints which apply to actors of a given organisation, they also forge behavioural patterns and decision-making choices within the formal institutional structure. Thus, they are determining features of the techniques used by any organisation pursuing its objectives. In turn, the techniques used to achieve objectives gradually affect the institutional structure, hence the rules of the game. Therefore, informal constraints have a determining influence on processes of incremental institutional change, and cultural homogeneity of a given society make them sources of political economic path-dependencies (North 1990; 44-5).

However, the works of North (1990) remain theoretical and lack explicit empirical basis. It is the VoC framework pioneered by Hall & Soskice (2001) which provides the tools to examine how institutional constraints, both formal and informal, vary from one political economy to another, how this variance leads to different hierarchies and modes of interaction on economic markets, and how political economies end up being characterised by variegated macro-economic systems of both consumption and production.

The cultural aspect of informal constraints as well as the way organisations gradually affect institutional structures discussed in North (1990) are still hard to incorporate in any framework of

(10)

10

economic analysis, making the institutional change facet of institutional economics a contentious issue needing additional theoretical scaffolding to be examined. The VoC approach completes literature on economic coordination, which had until North (1990) focused on internal coordination within organisations or hierarchies. The framework developed in Hall & Soskice (2001) focuses on explaining how companies coordinate activities between each other and therefore provides complete information serving to differentiate varieties not only of intra-hierarchy coordination, but also varieties of inter-hierarchy and inter-agent coordination.

3.2. Pension systems and Varieties of Capitalism

To differentiate aggregate production and consumption patterns, Hall & Soskice (2001: 7) focus on five spheres in which companies must develop relationships in order to resolve coordination problems. These are a) industrial relations, b) vocational training and education, c) corporate governance, e) inter-firm relations and f) coordination vis-à-vis employees. The following section briefly presents the monography’s conclusions before examining the case of France.

By analysing these five spheres within which coordination problems must be solved, Hall & Soskice (2001) ideal-typically differentiate Liberal Market Economies (LMEs) from Coordinated Market Economies (CMEs).8 This characterisation is primarily made by systematically comparing German (CME) companies behaviour from Anglo-Saxon (LME) ones and shows how production strategies of companies differ according to the institutional structure under which they operate. In general, the difference lies in the fact CMEs provide non-market institutions which reduce the uncertainty surrounding other actors’ behaviour.

The following review will show the VoC approach is a powerful explanatory tool, yet it struggles to account for institutional change in a direction opposite to the previously followed path of development. Hall & Soskice (2001)’s logic is able to argue that a firm’s behaviour tends to reinforce the pre-established institutional structure. Institutional ‘spheres’ are categories of institutions which constantly interact with each other through the organisations which perform under their rules. Debate as to the causal relations between institutional spheres themselves has led to no consensus (Crouch et al. 2005). Nevertheless, characterising the French political economy and the place of pension provision therein following Hall & Soskice (2001)’s logic is useful exactly for this purpose: to show how puzzling the case studies here analysed in fact are.

In summary, an LME configuration makes company behaviour rely on transparent market-coordination, whereas a CME provides additional non-market scaffolding for coordination between economic actors. There is no normative positioning as to which coordination type is better or worse than the other, albeit Hall & Soskice (2001: 45) end up arguing that measurable economic

8_{This nomenclature is somewhat misleading: coordination occurs and is necessary within any institutional}

configuration. The CME-LME differentiation is not a distinction between coordination and non-coordination, it is a distinction between coordinating institutions.

(11)

11

performances of a political economy are function of institutional coherence. Coherence refers to ideal LME and CME sets of institutional spheres. This is so because, according to their framework, coherence fosters complementarity; whereby the presence of one institution increases the returns from another. Pension systems are significant institutions regarding complementarities and coherence. Below, I will go through the way pre-funded, privatised and marketised – i.e. financialised – pension systems are characteristic of LME complementarities and coherence, and how PAYG pension systems are characteristic of CME complementarities and coherence. To do so, both cases will be confronted to the five spheres of necessary coordination.

3.2.1. Industrial relations

In the sphere of industrial relations, companies face coordination issues in wage-bargaining and working conditions, in relations with workers organisations such as unions, and with other employers. Pension provision systems are intrinsically linked to this sphere in that they are part-and-parcel of wage negotiations. Each political economy has different historical backgrounds leading to social insurance and pension systems. The place given by institutions for social dialogue between the State, employees and employers, and the relative force and organisation of workers and capitalists at moments of social insurance system constitution – often following the Second World War (Crouch 2000) – plays a determining role in the type of development that the system follows (Macheda 2012; McCarthy et al. 2016). With the social insurance and pension systems in place, complementarities are fostered thanks to the adaptative nature of human agency (Höpner 2005), and path-dependencies guide the theoretical logic of change, as presented above.

Following the VoC ideal-typology concerning industrial relations, the CME model provides non-market coordinating institutions, inexistent in LMEs, for relations between employers, unions, employees and other employers. This enables encompassing wage-bargaining institutions as opposed to more individualistic ones in LME, and provides pension and unemployment benefits more generous than in the LME ideal-type (Estevez-Abe et al. 2001). The empirical basis for this claim is the differentiation between German industrial relations and Anglo-Saxon ones. In Germany, the encompassing wage-bargaining institutions are tightly linked to pension systems in that both are negotiated at a firm- or more frequently sectoral-level (Estevez-Abe et al. 2001). This is caused by non-market institutions set up by the State. These non-market institutions create the main difference between LMEs and CMEs by facilitating credible commitments for long-term strategies between economic actors.

First, the State facilitates inter-employer coordination through the set-up of sectoral employer associations which enable strategic consultation between employers active in the given sector (Culpepper 2001). Then, platforms for mandatory negotiation with unions exist in order for their interests to be taken into account (Culpepper 2001). Additionally, executive bodies of German hierarchies are tied to supervisory boards comprising major shareholders, employee unions, as well as

(12)

12

major suppliers and customers (Hall & Soskice 2001). The industrial relations output of this institutional configuration usually takes the form of employers having knowledge of other firms’ strategies and long-term employment commitments as unionised workers’ interests are mandatorily included in wage and employment condition policies. Finally, the pension component of wage-bargaining is mostly made through sectoral PAYG arrangements, as security of employment, thus contributions, are insured. In this case, firms compete not on a price-basis, rather on a quality basis. In LMEs, these non-market institutions which facilitate or impose coordination do not exist. The only authority exercised by the State is that of guaranteeing ‘fair’ competition on markets by making firms disclose extensive information on their economic performances (Hall & Soskice 2001). Therefore, industrial relations are unable to foster credible commitments for long-term strategies between economic actors. In this case, market indicators guaranteed by the State are the primary source of information defining company strategies. The theories challenging classical economics presented above, in particular the works of Oliver Williamson (2007 [1985]) gives sufficient theory to understand company behaviour. Indeed, without additional non-market coordinating institutions, companies rely only on themselves to solve problems that cannot be addressed by markets alone. The industrial relations output of this configuration is an absence of credible long-term commitments and knowledge of competing firms’ economic strategies, and flexible labour markets entailing individual responsibilities (i.e. pre-funded arrangements) concerning retirement-income provision.

3.2.2. Vocational training and education

The main point made concerning vocational training and education within a given institutional structure concerns the direct influence social insurance models have on them. Therefore, welfare State modalities can be understood as complements to national production systems (Estevez-Abe et al. 2001).

CMEs are generally characterised by generous unemployment and pension policies leading to counter-cyclical budgetary policies (Amable & Azizi 2014). This gives confidence to citizens that it makes sense for them to invest in specific skill-set acquisition through personal training and education.9 Indeed, if they know that in case of job-loss they will maintain revenue until re-employment, it is not a big risk for them to invest in vocational training. To the contrary, if little unemployment protection is guaranteed, workers will rather invest in general transferable skills in order to maximise employment possibilities. Protection against job loss is thus not necessarily an external interference with efficient labour markets, as posited by neo-classical economic theory.

In their chapter, Estevez-Abe et al. (2001: 156-8) also emphasise how in turn, the incentives provided for knowledge acquisition impact levels of economic inequality. In CMEs, industrial relations incite long-term strategies of company-growth as well as employment. In this case, employers have

(13)

13

incentives to set-up vocational training tracks for their long-term employees to remain efficient throughout their career. In LMEs, companies have no insurance that an employee they invest in for training purposes will not be ‘poached’ by a freeriding competing company. Together with the flexible hiring and firing pattern of LME industrial relations, this leads to employers being less inclined to provide firm-specific skills to employees. Acquiring new skills and education in this context is primarily defined by the job performances of the worker once her general skill-set provides her with a job. In this general skill system, completing obligatory school does not lead to the possibility of entering a vocational training track. Therefore, academically weak students face lower returns from their educational investments, and their opportunities to invest remains low after completing general knowledge schooling. Therefore, academically weak students in general skills systems are more likely to be trapped in low-income unskilled jobs.

Together with industrial relations modalities, incentives for vocational training and education give the context necessary to understand the way respective pension schemes affect corporate governance regimes of a given political economy.

3.2.3. Corporate governance

The institutional sphere of corporate governance is the one most directly affected by modalities of pension provision because of the impact pension schemes have on a given country’s financial system. However, it is important to note that pension schemes are not the only variables affecting structures of financial systems; elements such as the timing of industrialisation also influence financial systems (Lazonick & O’Sullivan 1997).10

Redistributive solidaristic PAYG pension schemes such as the

régime général in France, or systems with non-capitalised occupational schemes such as those

predominant in Germany, are institutions which constantly redistribute contributions from active workers and/or employers to retired beneficiaries.11 In-so-doing workers and pensioners benefit from more stable and liquid disposable income than in capitalised and privatised systems which are based on stock-options or other financial products. Therefore, under a redistributive PAYG pension scheme citizens have a propensity to save mainly through bank accounts (Vitols 2001). This leads to a bank-based financial system in which banks have a larger role to play in financing projects. In turn, this incites companies and banks to develop long-term relationships. These relationships are to be incorporated in the general dynamic of CME industrial relations presented above, as they consolidate networks of trust enabling credible commitments for long-term strategies. In summary, industrial relations and social insurance policies which buffer income inequality through redistributive

10_{Also, all institutional spheres interact; industrial relations also have a visible impact on corporate governance}

and vice versa.

11_{The term solidaristic refers to the redistributive nature of the traditional French pension system. The name}

stems from the fact that in a PAYG system financed by employer/employee contributions and taxes, the interests of mid- and low-income households concerning the viability of the Welfare State and the rights of employees are aligned. In a non-solidaristic pension system such a capitalised and privatised one, retirement income flows from capital returns, which tends to align the interests of mid-income households with the interests of high-income households.

(14)

14

arrangements support bank-based financial systems, which in turn support bank-oriented systems of corporate governance.

As above-presented, the flexible labour markets of LMEs which entail individual responsibility for pension provision lead to individualised pre-funded retirement schemes. In countries where retirement income is provided by individual contributions to pension funds – i.e. in pre-funded pension schemes – there is a higher demand for financial products such as marketable securities in order for contributions to remain valued and ideally for them to grow in value. In such systems, citizens tend to hold less wealth in the form of bank deposits, meaning banks are not the primary source of financing. Political economies which have set up pre-funded pension schemes for the retirement income of their citizens therefore have financial systems that heavily rely on impersonal markets for financing needs and asset mobilisation.

In addition to the inequality fostered by the differential incentives provided for skill acquisition, such pension schemes tend to align interests of mid- and high-income households for shareholder value maximisation at the expense of employee rights (Jackson & Vitols 2001: 176). This type of push has been accentuated over time, as an increasing amount of assets under management of pension funds across the world are delegated to third party finance professionals driven by standardized portfolio management techniques (McCarthy et al. 2016). Moreover, pre-funded pension schemes affect levels of ownership concentration in traded companies by making ownership less concentrated, more widespread and less organised because of the lack of coordinating institutions (Roe 2003). This configuration is characteristic of LMEs, where in this situation of widespread ownership governed by norms of shareholder value maximisation, the capacity for firms to grow internally independently from market pressure is weak; as no protection from investor ‘exit’ in case of bad quarterly results is a

priori guaranteed.12 Therefore, corporate governance in LMEs can be conceptualised as being in itself marketized; governing bodies of LME companies are constantly exposed to takeover risks if the company produces bad short-term measurable performances.

To explain sustained levels of high employment in LMEs, flexible labour markets and wage-bargaining institutions are to be considered complementary to the financial system and its pressures on companies; workers must be easily layed off in order to maintain short term profitability. In summary, industrial relations and social insurance policies which accentuate economic inequality and make citizens rely on personal responsibility for pension provision lead to market-based financial systems which in turn accentuate the above-presented characteristics of LMEs, where, for example, products compete especially on the basis of cost advantages.

(15)

15

3.2.4. Inter-firm relations and coordination vis-à-vis employees

At this point, sufficient information has been presented in order to understand the general dynamics of inter-firm relations and coordination vis-à-vis employees characterising CMEs and LMEs. Inter-firm relations are strategically coordinated in CMEs, whereas they are more competitive on LMEs. As for relations vis-à-vis employees, they are by far more inclusive in CMEs than in LMEs, where firms rely on easily laid-off labour in order to maintain short-term profitability.

With this ideal-typology defined, Hall & Soskice (2001: 45) argue that institutional coherence along the five institutional spheres fosters complementarity between them, leading to enhanced measurable macro-economic performances. For example, LME structures are characterised by a set of coherent institutional spheres relying on few non-market coordinating institutions: flexible labour markets, competitive and transparent markets, pre-funded pension schemes, and market-based financial systems, to name a few. CME structures also present coherent institutional spheres relying on non-market coordinating institutions: encompassing wage-bargaining institutions, platforms for coordination between actors, vocational training facilities and bank-based financial systems. However, the claim that institutional coherence is the engine behind successful macro-economic performances has since been widely debated and proven erroneous (e.g. Kenworthy 2006). Quite to the contrary, some case studies show that incoherence can have beneficial effects on performances (e.g. Campbell & Pederson 2007). Anyhow, when it is recognised that different actors have different definitions of complementarity and consequent performance, incoherence can be shown beneficial for a specific set of organised interests while harmful for another (Höpner 2005).13 In the field of institutional change, both concepts can prove useful for analysis, yet require additional conceptualisation of the form of change taking place and a clearly defined angle of approach in order to be operationalised (Deeg 2007: 626-8).

With the ideal-typical LME-CME spectrum in mind, the following section goes through the above-presented institutional spheres in the precise case of the French VoC in order to grasp the particularities of the cases this research studies.

3.3. The French Variety of Capitalism

To say the least, the main specificity of the French political economy lies in the importance the State has in directly intervening for inter-agent coordination. As we have seen, in the German VoC coordination is made mainly through sector-level arrangements and the main function of the State is to guarantee the viability of the non-market coordinating institutions. In the case of France, the State plays a much more active coordinating role. In spite of its coordinating function, the French State’s role is significant beyond coordination, so much so that the political economy can be separated from

13_{For example, if the LME characteristic of high transparency concerning financial operations and performances}

is made mandatory in a CME, it can be considered beneficial from the standpoint of workers’ unions (for wage-bargaining and general negotiations with employers) while detrimental from the standpoint of employers.

(16)

16

the CME category and be labelled “State-Enhanced Economy” with a State which intervenes “where it sees fit” (Schmidt 2009: 521-2). This section reviews the meaning of this nomenclature relatively to the LME-CME spectrum.

Although not the first appearance of social insurance in the French political economy’s history, the instauration of the régime général in 1945 constitutes the institutional template upon which all later social insurance developments have built (Palier 2005: 63-106). It was set-up with solid support from Left wing political forces and centralised workers’ unions. Consensus was easily achieved, as opposition from capitalists and the Right-wing parties was marginalised because of its visible or suspected complicity with the Vichy regime (Palier 2005: 78-9). However, the system should not be understood as a case of class victory, it is built as a social contract of Liberation from German occupation and warfare towards citizens (Palier 2005: 100). Furthermore, although it is strongly supported by political forces and unions, it is not them who elaborate it; high profile bureaucrats of the State were the system’s engineers (Palier 2005: 101). The dominant corporatist paradigm among bureaucrats led to a system centred on the salaried employee; where social security was a function of professional activity, and where pension benefits were based on previous wage (Palier 2005: 113-5; Palier & Naczyk 2011: 91-2).

Since the first legislation concerning the régime général, industrial relations in France have not been characterized by efficient coordination. The main cause for this is how the French State has not encouraged the development of associations that it cannot control, in particular employer associations (Culpepper 2001). Workers’ unions also face disincentives to organise because of how the State is allowed to extend any agreement reached by just one trade union and one employer association to all firms active within the sector or the country (Palier & Naczyk 2011). Therefore, today French companies face difficulties when trying to set up credible commitments for long-term strategies between each other, as information between them does not circulate efficiently. In addition, the 1980’s saw the beginning of a trend of union power erosion because of decentralising efforts from the State and its withdrawal from direct implication for centralised wage negotiations (Schmidt 2009: 524-5). The first major reform in this direction were the Auroux laws which enabled companies to bypass centralised unions and promote firm-level, more local negotiations (Hancké 2001: 325-7). However, the lack of knowledge or experience of workplace negotiations created a strong discrepancy between formal law and actual practice, making the unions almost collapse under the weight of the new situation (Hancké 2001). This decentralising trend has been ongoing ever since (Palier & Naczyk 2011; Schmidt 2009). It has progressively led to industrial relations being characterised by antagonistic sets of interests and preferences between employers and employees, with no specific legal scaffolding to frame negotiations. It is when conflict intensifies that the French State intervenes in order to insure a minimum of coordination (Amable 2003: 137). In summary, the industrial relations

(17)

17

of the French political economy are characterised by minimum incentives for employer coordination and a trend of decentralisation of unions and their bargaining powers since the 1980’s.

The lack of non-market institutions for employer coordination in France has led to only a small quantity of complementary formation tracks proposed by the private sector. The lack of mandatory consultation between employers makes the private sector rely heavily on the State for vocational training and education (Culpepper 2001; Hancké 2001). Indeed, private companies in this situation have little confidence in their competitors’ good faith concerning skilled worker ‘poaching’.

Nevertheless, this trend does not make the French political economy fit the LME criteria of coherence; the relatively generous régime général as well as unemployment protection have remained in place to this day. Therefore, incentives do exist for individuals to follow education tracks other than general skill formations. These are provided by the State or by public-private partnerships. Furthermore, this social insurance system and its PAYG pension component contribute to lower levels of inequality and egalitarian wealth distribution. As the presentation made above shows, these elements together provide the institutional framework for a bank-based financial system to emerge and incites banks and companies to develop long-term relationships of trust. In France, this contributed to the development of patterns of cross-shareholding between networks of banks and companies in order to be insulated from excessive market pressure (Morin 1998). This case of strategic coordination amongst banks and corporations had developed with explicit support and incitation from the State, before its retreat from direct coordination spontaneously reinforced it (Bancel 1999; Morin & Rigamonti 2002). The cross-shareholding network immobilised large blocs of shares within French corporations, making it possible for them to go beyond short-term profitability and adopt a more CME-inclined dynamic of company development.

The theoretical characterisation of the French VoC makes the 1999 and 2003 pension reforms seem contrary to the economy’s theoretically path-dependent institutional spheres. Although saving patterns are not easily transformed (Cesaratto 2005; Vitols 2005), the measures undermine the bank-based financial system, which potentially affects corporate governance regimes. Furthermore, they provide signals to younger generations that the social insurance system should not be relied upon as confidently as before, which could in the long-term affect vocational training and education patterns. Altogether, this could lead to changes in inter-firm relations and methods of coordination vis-à-vis employees. In order to understand the reforms, the following sub-sections examine different conceptualisations of institutional change and academic contributions contributing towards understanding them completely. Specifically, it examines the possibility that literature on the specific type of institutional change of increased financialisation provides appropriate analytical tools. Reviewing this literature strand enables to legitimise a State-centred approach later methodologically systematised and applied to the pension reforms.

(18)

18

3.4. Institutional change

The central focus of the Varieties of Capitalism literature is the historical persistence of differences in the institutional arrangements within political economies and how these incentives and constraints shape the behaviour of companies. Its framework therefore enables the understanding of the fact that within a given political economy, production is organised in specific manners in consequence of particular institutional structures. The approach is firm-centred and shows how institutions define investment incentives and constraints for different agents and assets. Applied to the pension reforms hereby analysed, such a framework is able to explain why a number of attempts to capitalise and privatise pension provision in France failed in the years before the 1999 and 2003 reforms – employers and employees had vested interests in the continuation of the pre-existing PAYG system (Naczyk 2013). However, the VoC framework is ill-suited to fully account for the reasons why political economic institutional structures can break from the resilient characteristics of their framework. Therefore, it needs to be complemented by additional theoretical scaffolding in order for the 1999 and 2003 reforms to be understood.

The punctuated equilibrium model of institutional change is one that has received considerable and ongoing attention to explain institutional change. The way change is conceptualised in the punctuated equilibrium model is by examining the relative forces of agency versus structure (Katznelson 2003). In general, institutional change occurs in this view as a consequence of institutional breakdown providing a moment of critical juncture demanding innovative replacement. The term breakdown does not refer only to the functioning of the institution, it can also concern the paradigm guiding the institution’s governance or a change of the actors in charge of it, for example (c.f. Princen 2013). In spite of attempts to conceptually refine this model, the malleability of its conceptualisation of change makes it hard to systematically apply. Where to start searching for any case of breakdown or increased agency? And what causes the increase of agency in the first place? This theorisation of change will prove useful for the presentation of results rather than for defining an approach for analysis.

The punctuated equilibrium model is challenged by Thelen & Streeck (2005: 9), who are uncomfortable with its malleability and the fact it is “employed, almost by default, by most political scientists, sociologists, and economists working on institutional change.” Without denying that in some cases its framework is in all ways valid – the collapse of the Soviet Union for example – Streeck & Thelen (2005) and Thelen (2009) show that it does not capture the most common way institutional change occurs.

Thelen (2009)’s study shows that significant change rather takes place gradually and through an accumulation of small adjustments, in the absence of obvious ruptures. In sum, Thelen (2009) shows that pressures emanating from globalisation and the relative decline in manufacturing and other traditional economic activities in Germany have not been synchronised to institutional adaptation on the same grounds as the ‘core’ industries. Rather, a process of institutional drifting is conceptualised

(19)

19

as gradually taking place; actors who were involved in the original (i.e. ‘core’) design of coordinating institutions caused gaps to appear in rules applying to them. Consequentially, a gradual process of institutional change is underlined. The reforms analysed throughout this paper are not cases of non-occurrence as is the adaptation of institutions to new forms of labour in Germany; they are not cases of institutional drifting.

A more exhaustive listing of concepts facilitating analyses of institutional change is given by Streeck & Thelen (2005:1-39) who put forward five broad modes of transformative institutional change which they call displacement, layering, drift, conversion, and exhaustion. Layering and exhaustion best characterise the cases studied in this research. The 2003 reforms for occupational pre-funded pension schemes and third pillar equivalents consist in institutional layering because they are novel legislation rather than reforms of the pre-existing system. The 1999 reforms, on the other hand, consist in a form of institutional exhaustion; the PAYG system is viable in the long-run under conditions of full employment and increasing wages. Neither have been guaranteed since the 1970’s and the dissolution of the Bretton Woods institutions (Streeck 2013; Blyth & Matthijs 2017). The concept of exhaustion is similar to that of critical juncture moment or breakdown, only better defined for the cases of this research.

As for the punctuated equilibrium conceptualisation of change, conceptualising the form of change is useful to understand the nature of change ex ante and are hard to directly apply for understanding the causalities behind changes and reform (Streeck & Thelen 2005: 16). In order to identify the most fruitful angle of analysis to uncover the reasons why institutional layering took place, and why capitalisation was the reform chosen in the case of exhaustion of the PAYG system, literature focusing on the pension facet of financialisation proves a State-centred approach is necessary.

3.5. Financialisation

As presented in the introduction, financialisation refers to the process whereby financial motives, markets, actors and organisations are important components of the political economy (Epstein 2006). To be more precise than this general definition, it is possible to disentangle the process of financialisation on three scales; micro, meso and macro-economic (Van der Zwan 2014; 2017: 554-6). From a micro-economic perspective, financialisation occurs through an increase in consumption of financial products. At the meso level, financialisation is linked to the maximisation of shareholder-value as a principle of corporate governance. Finally, on a macro-economic scale, financialisation is associated to new patterns of capital accumulation; when the profitability of non-financial companies is increasingly derived from financial sources. As this definition makes clear, personal or occupational pre-funded pension schemes are symptoms of financialisation; they increase the demand and consumption of financial products and tend to push for shareholder value maximisation. The effects on the macro-level cannot be directly linked to pension systems, although the macro-level undeniably

(20)

20

is affected by processes occurring at the two lower levels (Lysandrou 2011; Bonizzi & Churchill 2017).

Highly financialised economies are typically LMEs such as the United Kingdom or United States (Engelen et al. 2010). This has led to some studies concluding that through pension capitalisation and privatisation, institutional convergence towards Anglo-Saxon templates is taking place (Dixon 2008). However, after further examination of cross-national comparative specificities – especially various management structures of funds once implemented – this claim can be strongly nuanced (e.g. Dixon & Sorsa 2009; Macheda 2012; McCarthy et al. 2016). Along these lines, two main strands of literature can be identified. The first focuses on the influence of the financial industry and characterises its lobbying power as the major driving force of pension reform towards financialisation (e.g. Kemmerling & Neugart 2008; Leimgruber 2012). This literature is complemented by a second strand which characterises the State as an arena of confrontation between actors of industrial relations who push or pull for pension capitalisation and marketisation; it shows that financial lobbies need support from dominant actors in industrial relations in order for their lobbying effort to be successful (e.g. Macheda 2012; Naczyk 2013; 2016; McCarthy 2014; McCarthy et al. 2016). The further development of newly-created funds then follows a specific development path, which States have recently strongly influenced as independent actors because of their considerations for fiscal consolidation (Natali 2018). In addition to the fact that States are necessarily part-and-parcel of pension governance and system change, they sometimes do so primarily for social stability (Jessop 1990) or economic objectives such as market-expansion (Schelkle 2012). This points to a third, emerging strand of literature which my research contributes to (Koreh 2017; Trampusch 2017; Tuytens 2019).

From a financial sector-centred perspective, Kemmerling & Neugart (2008) put forward the increasing influence of financial lobbies as an explanation for pension capitalisation, privatisation and in some cases marketisation. An increase in influence of a typology of preferences which in essence has a vested interest in the development of a market that would benefit its industry is an obvious explanation for the spread of pension financialisation. Kemmerling & Neugart (2008) thus explain the predominance of the pre-funded model of pension provision in the UK stemming from the fact that financial lobbies became very influential in the UK following the Thatcher government. On a trans-national scale and with more refinement given to the financial sector, Leimgruber (2012) shows that the insurance industry has been central, notably through its Europe-wide think-tank commonly called the ‘Geneva Association’, for the diffusion of the ‘three pillar doctrine’ that gives a predominant place to private pension schemes and is already solidly entrenched in, for example, the Swiss and Dutch pension systems.

The second strand of literature goes one step further and incorporates political struggles over the control of funds’ management into the analysis of their diffusion. This approach is very similar to the literature focusing on the general topic of institutional change presented above and has led to precision

(21)

21

being given to how and when lobbying power and influence comes to have a conclusive incidence on policy making. In the case of France, Naczyk (2013) shows that the privatisation attempt led in the late 1980’s failed because of the benefits employers and employees enjoyed through their common administration of the supplementary schemes. It was when these schemes came under heightened pressure because of a necessary (and agreed upon by social partners) increase of contributions in 1996 that the national council of French managers pushed for funded pensions to be part of the political agenda once again (Naczyk 2013).

Along similar lines, McCarthy (2014) shows that in the USA, where pension provision has always been organised on a pre-funded basis, it was the political configurations of social partnership which marginalised workers’ interests and gave heightened importance to employers on trustee boards through legal impositions. The employers made use of the pension funds in an impatient manner to finance operations such as the delocalisation of production – obviously contrary to the workers’ interests. To the contrary, Macheda (2012) shows that labour unions were directly implicated in the financialisation of Icelandic pension funds and the process was undertaken largely to their advantage. These publications show that social dialogue is an essential component of the outcome of financial lobbying concerning pension financialisation. However, it has proven to be inadequate for explaining the puzzling 1999 and 2003 reforms. Indeed, Naczyk (2013: 462), although able to account for the repeal of privatisation laws in the 1980’s in France, is unable through his analysis to explain the emergence of PERCO and PERP, which are the first collective and personal pre-funded retirement saving plans implemented in France in 2003.

In his study on the spread of occupational pension schemes in the United Kingdom, the Netherlands and Italy, Natali (2018) shows that all three political economies had similar pressures exerted on them by the financial industry’s lobbies and that this pressure ended up leading to different outcomes. Concerning the State, Natali (2018) shows that cost-containment measures were linked to modalities of pension provision in the United Kingdom and the Netherlands, where State intervention increased towards this end and was buffered by the varying relevance of social partners – higher in the Netherlands, where unions were able to keep the schemes tied to defined benefits, whereas the UK pension system shifted towards defined contribution arrangements.14

A recent strand of literature takes into account the State as an independent actor. These studies show that social insurance and pension reforms cannot necessarily be understood as reflexions of the concerned actors’ preferences – the State’s interests can come to override them when it comes to pension reforms (Trampusch 2017; Tuytens 2019) or social insurance in general (Koreh 2019). The

14_{The individualization of risk is here also taken as being part of financialization. Arguably, Natali (2018)}

conflates neoliberal morals and ideas described in Dardot & Laval (2009) & Amable (2010) with the process and definition of financialization described by Epstein (2006) and van der Zwan (2014) – the “financialisation of the everyday” is not a sine qua non condition of the general process of financialisation.

(22)

22

element all have in common is the emphasis on the fact that the too much attention has been given to the benefit side of social insurance, and not on its provision.

In a broad analysis of social insurance policies, Koreh (2017) argues that social insurance policies serve the independent interests of States as revenue raising institutions and power consolidators as much as they serve direct beneficiaries for social stability. In other words, Koreh (2017: 115) shows that social insurance policies should be considered as much a consequence as a cause of revenue-related legal frameworks. Using a fiscal-sociology approach, he shows that in certain cases, the fiscal dimension must be taken into account to understand social insurance policy outcomes. For example, the historical development of social insurance policies in Germany is closely linked to federalism and fiscal conflicts created by the country’s fragmented tax structure. Later in time, it became an instrument for shifting costs and budget constraints financed through taxes, to specific social insurance schemes financed by social contributions. Therefore, in some cases, social insurance systems are fiscal mechanisms. However, Koreh (2017: 119) sometimes conflates social insurance expansion with Keynesian automatic adjustments, which somewhat weakens his argumentation. Focusing on pension system reform, Trampusch (2017) intensively tracks the decision-making process which led to pre-funded pension arrangements to emerge in New-Zealand. The work provides insight into the fact that the introduction of pre-funded pension schemes in the country served broad fiscal, economic and financial functions rather than mere welfare functions. For instance, the author shows that the financial lobby’s emergence in the country was more a consequence of pension privatisation than the opposite.

Along similar lines, Tuytens (2019) argues that states should not be analysed only as ‘transmission belts’ of democratic pressure, but as actors who have their own interest in the outcome of welfare financialisation. According to Tuytens (2019), interventions from the State are guided by the motivations of balancing both social-political and commercial viability of private pension provision. This sums up all the literature presented in this section; the State as an actor in itself influences the outcome of both financial lobbying and social partner negotiations. However, as an independent actor, it is neither concentrating solely on the well-being of citizens and social stability, nor is it solely concentrating on its budgetary balance sheets or other economic preoccupations; both elements are at stake in pension reforms and the State can be shown to surpass financial lobbying efforts and social partner negotiations in order to guarantee the sustainability not only of the former as Marxist literature emphasises (Jessop 1990), it can also be guided by considerations concerning its own fiscal sustainability (Tuytens 2019).

However, Tuytens (2019) only provides a general argumentation with no specific case study and assumes that privatisation of pension provision is the long-term interest of any State without going further in examining why this is the case. Nevertheless, the three State-centred approaches reviewed

(23)

23

here do provide legitimacy for applying it to the 1999 and 2003 reforms in France. They show that through this angle it is possible to identify when and why financial lobbying and social partner negotiations can be inconclusive or insufficient to explain pension reforms towards financialisation. In the following section, I give precision as to how this approach is made methodologically operational by first presenting what my conceptualisation of the State is, and why parliamentary debates and the policy documents they build upon provide my primary source of documents analysed. Then, I show how I identified which parliament sessions and policy documents were most worthy of analysis, and how I analysed them.

4. Method

The State is by no means a constant unitary actor. In democratic capitalisms, it is composed of at least two major parties which themselves should be considered as organisations animated by policy paradigms which aim to mobilise different social blocs and represent them for policy-making (Amable et al. 2012). Therefore, the policy choices of a given State at a given point in time should represent the sets of interests most mobilised and represented in the policy-making process. However, as has been presented above, sometimes this logic is not appropriate to explain policy choices. For example, why would Bismarck set-up social insurance programs for workers when the dominant political forces of the time were undoubtably not representing workers’ interests? It has been theorised that this is because the State can override the dominant social bloc’s (or class in Marxist terms) short-term interests in order to guarantee long-term stability and the sustainability of capitalist modes of accumulation and production (Jessop 1990). In more recent cases, it has been shown that consensus for reform can be obtained because of State considerations concerning its sustainability as an organisation in and of itself (Tuytens 2019), or because of considerations concerning broader fiscal, economic and financial policies for the political economy as a whole (Trampusch 2017).

In any case, parties must be taken into account for analysing the State in contemporary democratic capitalisms; for reformative action to take place it is either because the power of the party defending the status-quo eroded, or because its positions changed or evolved over time. To understand these positions and their evolution, the theoretical framework elaborated above makes it possible to understand the potential implicit objectives of social insurance and pension reform. In particular, the concepts of complementarity and coherence can prove helpful to differentiate positions of proponents and opponents of a given form of change (Deeg 2007: 627). However, in order to make full sense of party positions it is necessary to precisely contextualise the moment of analysis. In order for a State-centred approach to be complete, it is therefore necessary to have knowledge of the fiscal and budgetary situation of the State, of the policy targets predominant at the time of analysis, and of the potentially implicit objectives or consequences of a political stance.

Varieties of capitalism and pension schemes : determinants of change in the case of France