1. Introduction

Since the early 1990s, the European Union has expressed its understanding of occupational pensions as a pillar of social protection for European citizens in numerous policy documents.1 The message in those policy documents seems clear: occupational pensions are a key element of the social wellbeing of European citizens.

Historically, the organisation of pension systems has been the purview of national governments. That prerogative is fiercely protected by them, inter alia because of the social aims that pension schemes pursue as well as the value of the welfare state to political parties.2 The EU, however, has busied itself with the regulation of a number of aspects regarding pension systems. Although it has no direct powers regarding pensions, the EU has a number of ways in which it can influence those systems, such as through its powers to regulate the internal market.3 It seems that that very crown jewel of EU achievements – the internal market – can clash with the social objectives pursued by the pension policies of the Member States. This article will focus on occupational pensions, referred to in pension parlance as ‘second-pillar’ pension schemes.4

Because of their design, some occupational pension schemes may run counter to demands and projects linked to a ‘well-functioning’ internal market. Member States with strong occupational pension systems typically feature some form of compulsory or quasi-compulsory membership with pension schemes operating on the principle of solidarity. The latter means that risks are shared between various categories of pension scheme members (e.g. young and old, men and women, but also high and low wage workers). These ‘social occupational pension schemes’ are generally the product of collective bargaining between social partners:5 they define the content of the pension scheme, choose the operator who will manage it and they make it compulsory for all employers and employees in a given sector (with or without the help of the State).

These two social aspects of occupational pension schemes – (1) collective bargaining and (2) compulsory membership – may be at odds with internal market ambitions: they can lead to a distortion of competition, as compulsory membership means granting a monopoly to the pension provider in charge of the occupation scheme; they can interfere with the freedom to provide services of other pensions providers, as the latter may not be able to access the market in the sector of activity covered by the pension scheme.

A question then arises: which one of the two seemingly conflicting objectives – the protection of social occupational schemes or the promotion of the internal market – prevails from the perspective of EU law and policy? Are collectively negotiated social occupational pension schemes protected from the requirements of the internal market or are occupational pensions to be designed in such a way as not to prevent the development of market forces in this sector?

The article aims to provide an answer to this question. Firstly, it explores how primary law deals with these two conflicting objectives. To do so it draws on the Court of Justice of the European Union (CJEU) case law applying competition law and freedom to provide services to pension schemes in particular and to social protection schemes in general. Secondly, the article focuses on secondary law adopted to regulate occupational pension schemes at EU level and attempts to identify their preferences towards social or market oriented occupational pension schemes. More precisely it examines two EU directives – the IORP II Directive6 and the Directive on supplementary pension rights7 – as well as a proposed regulation – the proposal for a Pan-European Personal Pension Product (PEPP)8 – and explains what effects they have or may have on the degree of solidarity of occupational pension schemes. The article concludes with a summary of how the EU manages the tensions between the social components of occupational pensions and the primary and secondary law of the internal market, and the implications arising from it.

2. Social occupational pension schemes and primary internal market law

EU primary law, in principle, requires the freedom to provide services9 and free competition.10 Under certain circumstances, these Treaty on the Functioning of the European Union (TFEU) requirements appear to make organising compulsory collective pension schemes challenging. The compulsory affiliation to an occupational pension scheme and the monopoly enjoyed by the fund operating that scheme can be obstacles to both the freedom to provide services as well as a restriction of competition. In fact, this compulsory membership has been the subject of CJEU case law, first challenged under EU primary law in the Albany11 case, a case on competition law. Despite what may have been thought for years, this case has not definitively solved the issue, as more recent developments in the Court’s case law have shown. We start with recalling the Albany’s lessons before exploring whether they are still valid before examining these lessons in light of subsequent case law.

2.1. Albany’s lessons: competition law

In the well-known Albany case, the CJEU dealt with the issue of whether mandatory occupational pension schemes established by social partners, managed by one single pension provider, is contrary to European competition law. Their relationship to the freedom to provide services was not reviewed in this case. The Albany judgment contained two lessons. First, the collective agreements (social element no. 1) that beget compulsory pension schemes are under certain conditions excluded from the scope of competition law ‘by virtue of their nature and purpose’:12 the collective agreement at issue in Albany, said the CJEU, was the outcome of a social dialogue – i.e. collective bargaining – and served to improve working conditions. Second, the introduction of a compulsory pension scheme (social element no. 2) managed by a single operator is not contrary to competition law if the scheme in question is sufficiently solidary.13 In that case, solidarity mainly meant distributive solidarity, i.e. solidarity between contributors in the form of non-proportionality between contributions and expected benefits.

These two lessons have been confirmed several times in subsequent judgments.14 They reflected a conciliatory approach to the social aspects of occupational pensions with the organisation of an internal market in which competition is not distorted, since these social aspects are immune, on fairly flexible conditions, from the rules that organise such competition. It is important to note, however, that the CJEU in Pavlov explicitly restricted the application of that conciliatory approach to occupational pension arrangements that are the result of collective bargaining between employers and workers.15 Mandatory occupational pension schemes not resulting from collective bargaining enjoy no such exclusion from the competition rules.16 This ‘overly formalistic approach’ was the result of the Court’s focus on the nature of the instrument used to establish the scheme rather than its ‘social objective and its solidarity-based structure’.17

Despite its somewhat limited scope, the Albany case law is a good example of how social objectives in general can be shielded from market law. Nevertheless, the Albany judgment concerned competition law, and not the four internal market freedoms. Since the Albany judgment, the EU has seen a number of developments, both legislative and in case law. The following section deals with the question whether and to what extent the Albany case law is still relevant. More generally, are mandatory social professional pension schemes (still) insulated from internal market law’s interferences?

2.2. The fate of Albany’s lessons after the Viking and Laval judgments: the four freedoms

After Albany, it briefly seemed that all EU Treaty provisions would generally be inapplicable to important aspects of national social policy.18 But the CJEU finally treated the exemption of collective agreements under competition law and the fundamental freedoms as two wholly separate questions.19 The consequences of this case law for occupational pensions are discussed below. Second, the issue as to whether the other lesson of Albany, regarding compulsory membership, can be transposed into the realm of the freedom to provide services, is considered.

2.2.1. Social element no. 1: collective bargaining

The approach in Albany as regards collective bargaining contrasts with that adopted by the CJEU a few years later in the famous Viking20 and Laval21 judgments. Instead of excluding collective bargaining agreements from the scope of the market rules, these judgments explicitly subject them to – in this case – the fundamental market freedoms. The Court held that the activities of the EU include not only an internal market, but also a ‘policy in the social sphere’ with a high level of social protection.22 Nonetheless, the fundamental freedoms ultimately prevailed over social policy, and the Court appeared to have set a boundary for social policy objectives in relation to the functioning of the internal market.

The CJEU therefore chose not to apply a uniform exception for collective agreements valid both in competition law and for the freedom to provide services. This contrasts with the Court’s previous Wouters judgment, where the Court held that justification under competition law implies justification under freedom to provide services.23 According to many commentators, the Viking and Laval judgments reflect a paradigm shift in the general conception of social and economic relations within the EU.24

This conception has been echoed in other judgments after Viking and Laval concerning the creation of occupational pension schemes.25 The first of these judgments is the 2010 case Commission v Germany.26 The CJEU decided that collective bargaining in the public sector to set up a compulsory pension scheme managed by a single operator must respect European public procurement law, which implements the freedom of establishment and the freedom to provide services. Germany argued that the Albany exception should also be applied to European public procurement law, in other words that collective bargaining should escape the latter in the same way that it falls outside the scope of European competition law. The Court rejected this argument. It recalled Viking and Laval and noted that the right to bargain collectively must be exercised in accordance with the requirements stemming from the freedom to provide services and the freedom of establishment.27

One can also find an echo of the Viking and Laval case law in the UNIS decision, adopted in December 2015.28 In this case, the CJEU decided that the granting of an exclusive right by public authorities to manage a supplementary social protection scheme to a single operator (through extension of the binding force of a collective private sector agreement) must comply with the obligation of transparency, which is required by the freedom to provide services.29 In this instance, the exclusive right granted was the right to manage a compulsory supplementary social security scheme.

Although the UNIS case tied the obligation to comply with the principle of transparency to public authorities, it seems undeniable to us that the onus of complying with that principle is primarily on the social partners: for the public authorities to be able to respect it when they extend the application of collective agreements, it is necessary that the agreement they extend made by the social partners complies with the principle.30 Thus the CJEU again echoes – though now implicitly – the Viking and Laval cases where it points out that the extension decision by the public authorities is not exempt from the requirements of transparency resulting from Article 56 of the TFEU.31

The judgments in Commission v Germany and UNIS thus confirmed, as regards occupational pensions, the CJEU’s choice not to transpose the solution adopted in the Albany judgment to define the system of collective bargaining in European competition law to the field of freedom to provide services.

The previous considerations demonstrate that the CJEU has set clear boundaries to the setting up of pension arrangements in place of an all-out exemption from the fundamental freedoms: when engaging in collective bargaining, the social partners must take into account the obligations stemming from inter alia the freedom to provide services. Although this means that the social partners – and the public authorities extending the application of collective agreements – cannot blindly appoint a provider, these obligations do not seem to place social pension schemes at the mercy of unbridled market logic. This will be demonstrated below.

2.2.2. Social element no. 2: compulsory membership

As discussed in the previous sub-section, the CJEU rejected the extension of the Albany exception to the fundamental freedoms. That raises the question whether the other lesson of the Albany judgment, according to which the introduction of a compulsory pension scheme executed by one single fund is not contrary to competition law if the scheme in question is sufficiently solidary, is valid in the field of the freedom to provide services. In other words, the question arises, following the Viking and Laval judgments and the echo they have found in CJEU judgments relating to occupational pensions, whether the compulsory membership to an occupational pension scheme managed by a single operator is compatible with the freedom to provide services.

Compulsory membership: a restriction of the freedom to provide services

There is no doubt, and the CJEU has repeated this on several occasions, that occupational pension providers provide services within the meaning of Article 57 of the TFEU.32 Compulsory membership – although conducive to solidarity – in principle restricts the two components33 of the freedom to provide services, that to provide services and that to receive services.34 On the one hand, occupational pension providers are deprived of the possibility of offering their services to companies that must be affiliated to the pension fund or insurance undertaking. On the other hand, these companies are deprived of the possibility of contracting with a service provider other than the pension fund or the insurance undertaking to which they are necessarily linked.

Nor is there any doubt that this restriction must be justified under Article 56 of the TFEU, regardless of the nature (law or collective agreement) of the measure imposing the membership obligation. Indeed, the Viking and Laval judgments clearly stated that Article 56 of the TFEU has a direct horizontal effect, i.e. it applies not only to measures enacted by States, but also to measures of general application enacted by non-state actors such as, for example, collective agreements adopted by the social partners.35 The question therefore arises as to the justification for the restriction constituted by compulsory membership.

Compulsory membership: a restriction that can be justified using Albany?

As already stated, in the Albany judgment compulsory membership of a pension fund was justified under competition law because of the solidarity features of the pension scheme managed by that fund. Can the solidarity of an occupational pension scheme also constitute an overriding reason of public interest capable of justifying a proportionate restriction on the freedom to provide services?36 To our knowledge, the question was never put to the CJEU. To answer this question, we can only try to extrapolate the position that the Court might adopt from other cases relating to the freedom to provide services.

As a preliminary point, it appears that this question need not be analysed through Article 106(2) of the TFEU, which allows States to derogate from certain provisions of European law for the organisation of services of general economic interest. This article played a central role in the Albany judgment. Several authors are of the opinion that this article could also apply to the freedom to provide services,37 which would make it possible to develop a uniform approach to exceptions to competition law and the freedom to provide services. However, it must be noted that the CJEU has never applied this article in the area of freedom to provide services, except in rare and long-standing exceptions38 or in cases where it did not ultimately play a role in the Court’s decision.39 In any case, we are of the opinion that the application of that exemption would likely ultimately amount to the same result as the public interest exception under the freedom to provide services. We will therefore examine the compatibility of compulsory occupational pension schemes with the freedom to provide services by applying the Court’s standard formula i.e. by determining whether an overriding reason of public interest can likely justify the restriction, and then address the matter of the proportionality of that restriction in the light of the objective pursued.

In several judgments, the CJEU has accepted that the need to safeguard a social security system against a risk of serious harm to its financial equilibrium constitutes an overriding reason of public interest capable of justifying a restriction on the freedom to provide services.40 Yet this risk of damage to the financial equilibrium of the system is directly linked to the existence of solidarity between the individuals covered by the scheme, the maintenance of which requires that membership of the scheme be compulsory. In accepting that maintaining the financial equilibrium of a social security system constitutes an overriding reason of public interest, the Court therefore also accepts that the protection of solidarity existing within that system may justify a restriction on the freedom to provide services.41

However, this case law cannot simply be transposed occupational pension schemes (pillar II), as those cases concerned statutory schemes (pillar I). Nonetheless, these findings could be a precedent for occupational pension schemes with strong solidarity between members and which occupy an important place, in terms of their coverage and the amounts they award, in the pension system of the Member State concerned, so that questioning their operation would have repercussions for the system as a whole.

Assuming that the protection of the solidarity of an occupational pension scheme is accepted by the CJEU as an overriding reason of public interest, the proportionality of the restriction is to be established. In this respect, the Court’s case law concerning compulsory insurance of the social security system seems less easily reconcilable with compulsory occupational pension schemes. In the Kattner judgment on German health insurance funds42 and the Freskot judgment on compulsory natural risk insurance for Greek farmers,43 the Court seemed to attach great importance to the fact that the insurance in question offered only basic cover, thus leaving room for an offer of supplementary insurance by private operators – to the extent that commercial providers could provide such coverage44 – to guide national courts in assessing the proportionality of the restriction.

What, then, would the CJEU decide about the obstacle constituted by compulsory membership of an occupational pension scheme in addition to the basic pension provided by the State? In order to be proportionate, is it necessary for the benefits of compulsory occupational pension schemes to be sufficiently low so that their beneficiaries can also contribute, if they so wish, to individual pensions? One can see the difficulties that this kind of decision would lead to. If this path is followed, it would be necessary, for example, to fix the amount from which the benefits paid by the occupational pension scheme are considered too high to leave sufficient room for individual pensions. Such an assessment would be eminently complicated, as well as eminently political.

The situation of compulsory occupational pension schemes with regard to the freedom to provide services in Europe is therefore uncertain, and is bound to remain so given the absence of a specific CJEU decision on this question. It is also uncertain because, even if the Court accepts solidarity as a justification for restricting the freedom to provide services, it cannot be said with certainty what degree of solidarity must be guaranteed for this justification to be proportionate.45

2.3. Drawing conclusions on the CJEU’s case law

The previous section has raised the question of the place of occupational pensions within European competition law and the freedom to provide services. Two aspects seem to be of particular importance for social occupational pension schemes: collective bargaining and compulsory membership. Initially, after the Albany judgment, it seemed that EU Treaty provisions would not be applicable to important aspects of national social policy. That notion was disposed of by the CJEU after Viking and Laval. The cases in the field of social protection that followed the Albany case law show that the social partners – as well as the public authorities involved in extending the application of collective agreements – are to heed the requirements stemming from the fundamental freedoms, such as the requirement of transparency when selecting a provider.

The possibility to make membership to a pension scheme or fund compulsory is also governed by the Treaty provisions on, for instance, the freedom to provide services. The obstacle created by compulsory membership must be justified. Whether in the field of competition law or the freedom to provide services, we have seen that the existence of solidarity between members of the occupational pension scheme is or would most probably be a central condition for the compatibility of compulsory affiliation with European law. However, uncertainty prevails in this respect, as it is difficult to determine the degree of solidarity required, and whether the standard is the same for competition law and for the freedom to provide services.

Some authors opine that, in order to draw conclusions about the relationship between the case law discussed here and possible future developments making pensions less solidary, it is necessary to closely examine the CJEU’s considerations on what defines the principle of solidarity.46 However, although the Court references certain elements which ‘tend to demonstrate’47 that the schemes at issue in its case claw apply the principle of solidarity – such as the fact that a scheme ‘is financed by contributions the rate of which is not systematically proportionate to the risk insured’ or ‘that the amount of benefits paid is not necessarily proportionate to the insured persons’ earnings’48 – it provides no definition or precise guidance on what solidarity is. One could, so it appears, reasonably surmise that the guiding question is whether the features of the scheme could be offered by commercial providers.

But in the absence of clear guidance relating to solidarity of pension schemes, it appears difficult to gauge what consequences the well-documented and steadily ongoing trends of individualisation and marketisation49 of pension schemes – whereby more and more risk is shifted from the employer and/or pension provider to employees/retirees – have for compulsory membership. Such trends occur even in Member States with strong occupational pension systems featuring compulsory membership. It is difficult to predict how the CJEU would assess these trends in the light of the standard of solidarity it has developed in its previous case law.

Be that as it may, it seems that today the decisions of the CJEU are no cause for great alarm regarding the fate of social occupational pension schemes. But uncertainty might arise tomorrow from the condition of solidarity to justify derogations from competition law and probably from the freedom to provide services, which is the fragile point of the Court’s reasoning.

3. Social occupational pension schemes and secondary internal market law

In view of the discussion above on the CJEU’s interpretation of primary EU law in relation to occupational pension and social security schemes, it is interesting to examine in this section what some important legislative developments that have recently occurred at EU level might have in store for social pension schemes. Indeed, it seems that some of them could lead to or strengthen the trends mentioned above towards a decrease in the degree of solidarity within occupational pensions. If it were so, they would consequently weaken the protection of mandatory occupational schemes against the requirements of market rules enshrined in European law.

The EU has only limited legislative powers in the social policy sphere itself. It has nevertheless been able to influence Member States’ social policies through the adoption of secondary law instruments based on its internal market regulation powers, notably Articles 46 and 114 of the TFEU. It has done so recently in the field of occupational pensions by adopting two Directives, the revamped IORP Directive (IORP II)50 and the Supplementary Pension Rights Directive.51 Both are discussed hereunder. We outline their main objectives and how these are supposed to be achieved. Our aim is to highlight their potential effects on solidarity within occupational pension schemes. In addition to the two Directives, we briefly discuss a new proposed Regulation for the creation of a Pan-European Personal Pension Product, with the same objective of highlighting what its implications could be for social occupational pension schemes if it were to be adopted.52

Before we begin this section, it is useful consider the types of occupational pension schemes in order to facilitate understanding of the discussion that follows. There are two archetypical occupational pension schemes, although in practice most schemes can be positioned anywhere between the two ‘poles’. Defined contribution (DC)53 schemes are schemes in which the contribution that the employer and/or employee contribute every month is fixed. The level of the pension benefits is a direct function of the contributions and becomes known only on the date of retirement. It is the pension scheme member who bears the ‘risk’ of longevity and the resultant lowering of benefits, as well as risks related to, for instance, fluctuating interest levels and stock market yields. Defined benefit (DB) schemes are schemes in which the benefits, instead of the contributions, are defined in advance, for instance as a percentage of the average or final salary. The risk is borne by the employer and/or the pension provider to achieve the promised benefits. Such schemes require additional financial buffers.54

Solidarity is absent in pure DC schemes as the benefits (the pension) are directly dependent on the personal amount of contributions paid by or for the worker: at the end, everyone is supposed to receive what he has contributed to the scheme. On the opposite, solidarity is present in DB schemes. But its intensity is varying. The weaker the correlation between amount of contributions that have been paid by/for the worker and the benefits to be received, the stronger the solidarity is within the scheme.

3.1. The IORP II Directive

3.1.1. Introduction

After a long process that commenced in 1980s, the Institutions for Occupational Retirement Provision (IORP) Directive I55 was passed in 2003, which was recently recast and is now called IORP II.56 The revised IORP II Directive came into force as of 12 January 2017. It explicitly acknowledged the social role that occupational pensions play,57 whereas IORP I only regarded IORPs as ‘financial service providers’,58 an approach that significantly influenced the content of the Directive.59

Despite the revised Directive’s recognition of the social function of IORPs, both versions of the IORP Directive are actually financial service directives, with the ‘limited ambition’ of creating a prudential framework for pension providers that allows mutual recognition of supervisory frameworks for pension providers.60 That mutual recognition facilitates cross-border activity by IORPs between Member States. During the IORP I Directive’s conception, the European Parliament pushed for the inclusion of provisions regulating elements of solidarity, but the Commission and the Council rejected the regulation of such elements.61 The IORP II directive does not provide for such ‘social’ provisions neither.

The IORP II Directive does not in itself hinder the legal ability of Member States to promote or to impose social occupational pension schemes in their territory: it does not directly regulate the pension systems of the Member States but instead focuses on the regulation of the providers (called ‘IORPs’). However, it could have an indirect influence on solidarity within these schemes. This is linked to the requirements contained in the Directive with regards to IORPs engaged in cross border activities.

3.1.2. The Directive’s full-funding requirement: an obstacle to social pension schemes?

The Directive mandates that if IORPs are active on a cross-border basis, i.e. if they operate pension schemes in another Member State than the one in which they are established – they must be fully funded, irrespective of whether they offer DB schemes, DC schemes or a combination of both.62 This means that the pension fund must have sufficient assets in its coffers to cover the accrued benefits. However, for DB schemes or any other scheme featuring guarantees for which no external sponsor bears risks, the IORP II Directive (Article 15) requires that the IORP has additional assets: so-called regulatory own funds to finance such coverage.63

The full-funding requirement that cross-border IORPs are subjected to makes setting up DB schemes more expensive than DC schemes to the parties bearing the risks associated with them – in the case of DB schemes the risk-bearer is the employer and/or the pension provider. They are legally required to make up any funding deficits, whereas such an obligation does not necessarily exist for purely national schemes. Indeed, per definition DC schemes are fully funded, because the amount of the benefits they pay depends directly on the amount of funds available to them. On the contrary DB schemes are often underfunded, and making up funding shortfalls can be an expensive undertaking. The difficulty for setting up cross-border schemes is even greater for DB schemes because of the additional assets requirement to which they are subject.

It is noteworthy that the full funding requirement was to be removed by the IORP II Directive. The Commission, in its Impact Assessment for the IORP II Directive, noted that retaining the requirement ‘will not help attain the objective and would hamper IORPs’ willingness to engage in cross-border activities’.64 The removal of the requirement would have meant that cross-border IORPs would have been treated in the manner as IORPs that do not engage in cross-border activity.65 To the dismay of a number of stakeholders, the full funding requirement was ultimately retained.66 However, as a compromise, the wording of the requirement that cross-border IORPs be fully funded was changed: the new definition jettisoned the words at all times, suggesting some leniency regarding the funding status of cross-border IORPs. But although that textual change appears to make temporary underfunding possible, the full funding requirement is still in place, raising the question whether there is any change in practice.

The retention of this requirement arguably makes the provision of DB schemes by cross-border IORPs more burdensome than DC schemes.

3.1.3. Legislative requirements

DB schemes are also more difficult to organise on a cross-border basis for another reason than funding requirements. As the Commission explained, IORPs organising DB schemes on a cross-border basis are more likely to be subjected to complex national legislation from a Member State different from the one they are established in, owing to the complexity of the schemes and features that do not exist in DC schemes.67 That makes the adoption of DB schemes less plausible than the one of DC schemes.

3.1.4. Conclusion

It can therefore be seen that, if cross-border pension schemes were to proliferate further – which is one of the very objectives of the Directive – they would probably mainly consist of ‘DC oriented’ schemes with fewer or no elements of solidarity rather than ‘DB oriented’ schemes due to the higher cost and complexity of operating such schemes across borders. Yet, the more a scheme is DC oriented, the less likely it is to meet the solidarity criteria developed by the CJEU in competition law, which could be problematic for compulsory social pension schemes. As discussed, those criteria are probably also relevant for the freedom to provide services. This means that achieving the IORP II Directive’s objective of developing IORPs’ cross-border activities would likely lead to the development of pension schemes which according to EU primary law could not be mandatory and managed by a single operator because of their lack of (sufficient) solidarity as defined by the Court. However, the importance of this plausible development should not be overestimated, as so far there have been few cross-border activities of pension funds.68

3.2. The Supplementary Pension Rights Directive

Another important piece of EU law on occupational pensions is the Directive on the acquisition and preservation of supplementary pension rights.69 It aims to facilitate free movement of workers by providing for a minimum harmonisation of certain aspects of the regulation of supplementary pensions. More specifically, it lays down four important requirements regarding acquisition and preservation of supplementary pension rights, in addition to improving obligations of information provided to workers. First, waiting and vesting periods, i.e. periods before which no supplementary pension rights are acquired, shall not exceed three years in total.70 Second, the minimum age for vesting pension rights shall not exceed 21 years.71 Third, the acquired rights of an outgoing worker, i.e. ‘dormant rights’, must be treated in line with the value of the rights of active scheme members or pension benefits currently in payment or at least to be treated ‘fairly’ in relation to them.72 Fourth, where an outgoing worker’s vested rights are below a certain threshold, to be fixed by Member States, supplementary pension schemes can commute them into a capital sum to be paid to the worker, provided that he or she gives his or her consent.73

In its very aim to promote worker mobility, i.e. to smooth out the consequences of this mobility as regards supplementary pension rights, the Directive appears to be in tension with solidarity within a pension scheme. Indeed, strong solidarity can only be organised within a precisely defined and in a largely permanent collective subject to common rules. The greater the numbers of exit, the smaller the real possibilities of organising solidarity. This is precisely one of the reasons why it is easier to establish solidarity-based occupational pension schemes at sectoral or occupational level, because in this way, as long as the worker who leaves a company continues to work in that sector or profession, i.e. as long as he or she remains a member of the collective within which solidarity is established, mobility does not run counter to solidarity.

This tension between the Directive and solidarity is not only in terms of objectives, it can also be traced in the measures it enacted. Indeed, the Directive tends to reinforce the concept of the supplementary pension as a counterpart equivalent to the contributions paid, rather than a right attached to the worker’s situation and at least in a good part independent of his or her contributions, i.e. a right based on solidarity. Firstly, the reduction in waiting and vesting periods for pension rights insists on the right to receive a supplementary pension in return for the work actually done, to the detriment of the idea that this right is based on the worker’s ‘time-tested’ membership of a (solidary) collective. Secondly, the alignment of the valuation of dormant rights with the valuation of active members’ rights also implies the conception that the supplementary pension is basically a counterpart to the contributions paid – so there is no justification for differentiating the treatment between contributions currently paid by active members or in the past by former workers – rather than a counterpart to the professional career – where there is then a reason to make a difference between those whose careers are different. Thirdly, by providing for the refund of contributions paid by workers if they leave their jobs before they have started to acquire additional pension rights, the Directive also underlines the idea of a direct link between contributions and pension rights.

The logic behind the Directive therefore seems less easily reconcilable with solidarity-oriented DB schemes, where the link between contributions paid and benefits is loose, than with DC schemes, where benefits are the direct counterpart of contributions.74 It is interesting to note in this respect that a report carried out in 2007 for the Commission established that at that time 15% of DB schemes required an acquisition period of at least five years and 32% of them a period of more than two years. Moreover, 25% of DB schemes did not offer any revaluation for dormant rights, while most DC schemes made no difference between dormant rights and current rights as regards the allocation of the remuneration of their assets.75 These figures show that the requirements contained in the Supplementary Pension Rights Directive are more in line with the practice of DB schemes than with DC schemes.

As with the IORP II Directive, the Supplementary Pension Rights Directive does not oppose the establishment of solidarity pension schemes. But as with the IORP II Directive, however, it could serve as a catalyst for a broader trend within Member States to shift from DB schemes to more DC oriented schemes. If this trend were to increase thanks, inter alia, to the Directives, one can doubt that solidarity would be sufficient, according to the ECJ, to justify compulsory membership of a scheme managed by a single operator.

3.3. The Pan-European Personal Pension Product (PEPP)

A development that seems to fit in well with the aforementioned76 trend of pension individualisation is the proposed Pan-European Personal Pension Product (PEPP). As early as 2012, the Commission asked the European Insurance and Occupational Pensions Authority (EIOPA) to provide advice on the creation of a single market for personal (i.e. third pillar) pension products with a view to strengthening the market for such products.77 EIOPA’s preparatory work resulted in a proposed Regulation that was published on 29 June 2017. At the time of writing of this contribution, the proposal is still under review by the European Parliament.78 The PEPP is, according to the proposed Regulation, a category of retirement products in the third pillar with individual membership and individual accounts.79 By contrast to occupational plans, personal plans are pension plans to which access does not depend on an employment relationship.80 Consumers acquire them voluntarily and on an individual basis.

The PEPP is described by the proposal as an important element in bolstering Europeans’ retirement savings and increasing the availability of retirement products in the face of the pressure pension systems are facing as a result of demographic change.81 It can be seen as a further development in the privatisation and individualisation of pensions, whereby individuals are increasingly responsible for their own retirement income provision.82 As the responsibility of employers in retirement provision recedes, individuals are left with a greater responsibility of their own in ensuring the adequacy of their retirement provision. In that connection, Stevens identifies a ‘silent pension pillar implosion’ whereby the second pillar (i.e. occupational pensions) is being displaced by the third pillar.

The PEPP’s impact, therefore, could extend beyond the scope of the third pillar. Indeed, it may have potential ‘negative side-effects’ for social occupational pensions. Stevens notes that:

[i]t is well known that there are policy effects between forms of pension or pillars within countries. If a country promotes, either fiscally or socially, one of the pillars [i.e. in that case the third pillar], the other pillars are affected by this.83

If PEPP becomes a success, that success could come at the cost of more established sources of pension provision, such as social occupational pension schemes.84 Stevens finds it ‘surprising’ that both EIOPA as well as the European Commission did not take into account these more than likely consequences of their project.85

However, the precise effects of the proposed Regulation will depend on which types of PEPP will become popular: a PEPP is not one single kind of pension product, but rather a product class within which it appears possible to offer PEPPs with a multitude of features – including, so it seems, DB-type PEPPs. The proposed regulation leaves a lot of room for providers to design their PEPP as they see fit, even – so it seems – as a product very similar to a traditional DB scheme. There seems to be nothing in the Regulation preventing the possibility of establishing an employer-sponsored collective PEPP.

As with the two Directives discussed above, the PEPP as such would have no legal effect on social occupational pension schemes. But as with the two Directives, and perhaps even more so, its side effects could be the weakening of these social schemes correlatively with the increased individualisation of pensions, be they private or occupational.

4. Conclusion

Taking as a departure the statements of the EU – especially the Commission – about the important contribution of occupational pensions to the wellbeing of citizens, this article raised the issue of what effects EU (case) law has on the types of occupational pension schemes. In several Member States, occupational pensions are designed in such a way that they may be in tension with the objectives of the internal market. This is particularly the case for ‘social occupational pension schemes’, i.e. occupational pensions providing for a strong solidarity between their members – which requires that they be managed by a single operator – and set up by social partners in the context of collective bargaining. In such a situation, is EU law a support for social occupational pensions prevailing over internal market objectives or is the opposite true?

To answer this question, this article examined both the case law of the CJEU applying competition law and freedom to provides services, on the one hand, and a number of EU legislative initiatives, on the other hand, in order to examine the relation between social – i.e. collective bargaining and solidarity – and internal market values in the field of occupational pensions.

Our discussion of the CJEU case law in competition law and freedom to provide services began with the famous Albany judgment. In that judgment the ECJ immunised collective agreements from the provisions of competition law. Moreover, it considered the occupational pension system at issue in that case, which featured compulsory affiliation to a pension scheme based on the principle of solidarity and was operated by a single pension provider, was not contrary to competition law thanks in particular to its strong solidarity. This case law is a typical example of how the objectives of the internal market can be adapted to ensure that social values are respected.

However, in the field of occupational pensions as in other fields, attempts have been made before the CJEU to invoke the exceptions of Albany against the freedom to provide services. Those attempts were unsuccessful. In the contentious Viking and Laval cases, the Court destroyed any illusions about the exemption of collective agreements from competition law found in Albany applying also to the fundamental freedoms.

In fact, it is not just the collective agreements that must comply with the fundamental freedoms, but so too must decisions of public authorities extending the application of such agreements and – we argue – the social partners’ selection for a pension provider. What could be the consequences of this? The social partners must take into account the freedom to provide services including, inter alia, the principle of transparency when selecting a provider. The same goes for public authorities extending the application of the collective agreement. This is potentially a restriction on the autonomy of collective bargaining because the social partners can in principle no longer establish specific links with an operator without justifying it through a procedure transparent enough to meet European standards.

Given the Viking and Laval case law, it is clear that occupational pension systems featuring compulsory membership to a scheme managed by one single pension provider must also comply with freedom to provide services; they cannot invoke Albany against the latter. Under certain circumstances, we are of the opinion that the presence of elements of solidarity could be a saving grace invoked as an exception in the public interest. But as any exception, it must be proportional. By themselves, these requirements for justification would not be an affront to systems of solidarity: both compulsory membership as well as an exclusive right for a provider to operate such a scheme can probably be justified under the freedom to provide services. However, they do limit the freedom within which governments and social partners can set up occupational pension systems. Moreover, the requirement of solidarity to justify restrictions to freedom to provide services introduces uncertainty; if legislation or practice in the field of occupational pensions changes, it is necessary to speculate whether solidarity is still sufficient in the eyes of the CJEU.

In that respect, it was interesting to move on to the second part of our analysis and to ask whether European legislation encourages or perhaps threaten solidarity within occupational pension schemes. The EU lawmaker seems to lack the power to regulate the Member States’ pension systems directly. That is a power that still, to a large extent, is the prerogative of the Member States. EU secondary law has, however, influenced national pension systems through the IORP Directives as well as the Supplementary Pension Rights Directive, and will do so soon once the PEPP Regulation passes the scrutiny of the EU lawmaker. These are internal market initiatives, aimed primarily at the freedom of pension providers – or IORPs in EU parlance – to provide cross-border services and the freedom of movement of EU citizens, respectively. Does this, then, mean that the EU’s legislative initiatives in the field of occupational pensions are a threat to pension systems relying on solidarity?

The discussion has shown that the IORP II Directive seems to encourage the setup of DC pension schemes by cross-border IORPs, which do not offer any solidarity, rather than DB schemes, where solidarity is more or less present between the members of the scheme. The same observation can be made with regard to the supplementary pension rights directive: in its logic and requirements, it is more easily reconciled with less solidarity-based but more DC-based schemes. And even if this Directive formally covers only cross-border situations, national legislators will probably adapt all national legislation to the Directive, as it would be technically too difficult to adopt a law applying only to workers who actually or potentially move to another Member State in parallel with existing national law. As for the PEPP, time will tell how this product proliferates on the European market and the possible consequences it may bring for the other pension pillars; it can nevertheless be reasonably argued that its proliferation would probably be to the detriment of occupational pension schemes.

The article thus showed that, in the field of occupational retirement provision, the CJEU’s interpretation of primary law does not today in itself constitute a threat to solidarity, whereas it could constitute a constraint to the autonomy of collective bargaining by subjecting the choice of the provider to the freedom to provide services. However, solidarity could be put under pressure due to secondary law, adopted to promote an internal market for pension providers and to deepen the internal labour market. Thus, contrary to what is usually found with regard to other branches of national social law facing the impact of internal market law, the threat to the social aspect of occupational pensions could come from ‘positive integration’ rather than ‘negative integration’ – to use the famous distinction made by Fritz Scharpf.86 Nonetheless, if this pressure from secondary law were to produce its effects, together with other factors, it would have to be asked in return whether compulsory membership of schemes managed by a single operator is still compatible with primary law; the weakening of solidarity weakens the position of these schemes with regard to competition law and the freedom to provide services.