EU Executive Rule-Making and the Second Directive on Institutions for Occupational Retirement Provision

The legislative system of the European Union is determined by the distinction made between legislation and administrative rules and between delegation and implementation pursuant to the EU Treaty. This article analyses the implementation of this system in practice, on the basis of the revision of the IORP Directive. Not only do both distinctions give rise to conflicts between mostly national and EU legislatures but they also lead to disputes regarding the substance of the EU rules. This article also provides more detailed insight into the role of the European Insurance and Occupational Pensions Authority (EIOPA).

acquired a constitutional dimension, which is reflected in the fact that political scrutiny has become a key characteristic thereof. 14 The substantive criterion is that 'essential elements' of an issue need to be regulated by the legislature itself whereas 'non-essential elements' may be delegated to the executive.
In the Schengen Borders Code case the CJEU made clear that it has a substantial role in defining the distinction between legislative acts and non-legislative acts. Long before the Treaty of Lisbon, the basis for that conclusion had already been developed in the case law of the CJEU in which 'essential' elements were distinguished from 'non-essential' elements. 15 Although the terminology is quite open-textured, the findings as to which elements are essential are '(…) [to be] based on objective factors amenable to judicial review' (Paragraph 67). 16 According to the CJEU, the decision of the Council supplementing the Schengen Borders Code (in the form of what we would now refer to as an implementing act) did indeed relate to essential elements. They should have been laid down in the basic act (i.e. a directive, secondary EU Law) and the decision of the Council therefore had to be annulled. The essential elements that had to be decided at the legislative level included the enforcement powers granted to border guards to take coercive measures. According to the CJEU, not only did the Council decision thus affect fundamental rights, but it also involved political choices: a balance had to be found between conflicting interests (protection of the migrant vs. the interest involved with checks at the EU's external borders). 17 Political choices as a criterion for defining essential elements is, however, problematic: if only since political decision-making can never be fully excluded in the context of executive rule-making. 18 In any case, the discretion of the legislature has been effectively restricted by the CJEU. 19 This is obviously manifested first of all by the outcome of the case. Moreover, the CJEU did not review whether or not the Council, as the executive institution, had remained within the margins set by the EU legislature (the delegation provision from the Schengen Borders Code) but rather whether the EU legislature had respected the general, constitutional distinction between essential and non-essential elements. In other words, the CJEU focused on the limits on the legislature rather than on the limits on the delegate.
The dividing line between legislative and non-legislative acts is also relevant from an inter-institutional perspective. In a recent case the CJEU had to rule on the Commission's decision to withdraw a legislative proposal for a Regulation on macro-financial assistance to third countries. 20 The Commission opposed the Council's and the European Parliament's amendment of the proposal to remove an implementing power from the proposal according to which the Commission would have the power to decide on the actual conferral of assistance. The European Parliament and the Council had lodged an appeal against the withdrawal of the proposal by the Commission. The Commission disagreed with the decision of the European Parliament and the Council in that the European Parliament and the Council were of the opinion that such decisions should be made via the ordinary legislative procedure (and thus took the position that such decisions were part of the essential elements of the Regulation). The CJEU, however, found that the Commission's withdrawal was justified. One of the main objectives of the proposal is to accelerate the decision-making on providing assistance and thus its effectiveness. In light of this, the implementing power was in fact an essential element of the proposal, according to the CJEU (Paragraph 91). Furthermore, the withdrawal was justified Utrecht Law Review | Volume 12 | Issue 1 (January) 2016 based on the Commission's position regarding the trialogues. In practice, these informal consultations between the three legislative institutions are of vital importance in order to reach consensus. The CJEU used the Commission's position during these trialogues as a justification for the subsequent withdrawal of the proposal. It did not answer the material question of whether this case involved essential or non-essential elements. The case should be seen as an answer to the question regarding the scope of application of the powers of legislative institutions to determine what elements of the regulation are essential.

The distinction between delegation and implementation
The Treaty of Lisbon has introduced another distinction that relates to the distinction between delegation and implementation. 21 This distinction, although it may appear as quite technical, has indeed become extremely politicised. A lack of consensus on the choice between delegation and implementation frequently leads to considerable delays in the legislative process. This has been the case in major legislative dossiers such as the Multiannual Financial Framework, the reform of the Common Agricultural Policy and the series of measures relating to the Single Market Act I (April 2011) and II (October 2012). 22 The diverging and even opposing interests of EU institutions and the lack of clear substantive criteria and definitions in the Treaties are conducive thereto. Delegation is based for the most part on ex post control by the Council and the European Parliament. 23 Depending on the legal basis, these institutions may have the power to object to a specific delegated act (as a result of which it cannot enter into force) or they may even have the power to revoke the delegation altogether. Moreover, a legislative act may provide that both of these powers are connected to a specific delegation. The power to revoke the delegation is a radical one as it essentially means that the Council and the European Parliament can unilaterally amend a legislative act. However, as far as the authors are aware, this power has not yet been actually exercised. Both the Council and the European Parliament have, however, raised several objections to specific delegated acts. 24 Thus, delegated acts have certainly become included in processes of democratic oversight. 25 Not surprisingly, the European Parliament strongly prefers Article 290 TFEU because of these ex post control mechanisms 26 which it lacks under Article 291 TFEU. 27 By contrast, the Member States prefer implementation over delegation because of the ex ante control mechanism which applies to implementing acts via committees of national experts (comitology). The Member States value this mechanism of ex ante control mechanism highly and have tried to apply it to the adoption of delegated acts as well. Although the Commission already promised in 2009 to 'systematically consult experts from the national authorities of all Utrecht Law Review | Volume 12 | Issue 1 (January) 2016 the Member States' 28 -and this commitment has indeed been incorporated into various legislative acts 29 -Member States view this as an insufficient guarantee. These diverging subjective institutional views on delegation and implementation now make the legislative choice for either Article 290 TFEU or 291 TFEU necessarily easy to understand. Moreover, control by the CJEU is limited as the Court grants the legislature a wide margin of discretion. In its decision in the Biocides case, the CJEU concluded that judicial control is limited to whether the choice of the legislature falls within the conditions laid down in Articles 290 and 291 TFEU. 30 In a later case, the CJEU decided that whether the Commission has been granted a discretionary power or not (or, if so, the scope thereof) is not a relevant factor to determine whether Article 290 or 291 should apply. 31 Despite the legislative discretion that determines the choice between Article 290 and 291 TFEU, Schütze has argued that delegation and implementation are fundamentally different in nature: in his view Article 290 TFEU concerns the delegation of legislative power (as it involves the power to amend primary legislation) whereas Article 291 TFEU concerns the delegation of executive powers (as it involves the implementation of primary legislation). 32 The respective control mechanisms reflect this different status: control in the context of Article 290 TFEU focuses on democratic guarantees, whereas control by Member States is the key element in the context of Article 291 TFEU. The latter element can be explained by viewing implementing acts as a manifestation of executive federalism: Member States feel the need to be able to control the (supranational) Commission.
Such an idea of a fundamental difference between delegation and implementation is difficult to reconcile with the EU's legislative practice. Delegation is limited to acts of general application but this does not imply that implementation may not include measures of general application as well. 33 Moreover, whereas implementation is supposed to only give effect to existing provisions of primary legislation (and, thus, to refrain from adding new elements thereto), the reality is that it often inevitably involves value judgements and political choices. 34 Thus, implementing acts and delegated acts are not easy to distinguish on the basis of their content. 35 With the Interinstitutional Agreement on Better Law Making (IIA), 36 the EU institutions have tried to remove some of the controversy by bringing the procedure to adopt delegated acts closer to that of the committee procedures under Article 291 TFEU in that: -it is now an obligation for the Commission to consult Member State experts; -the Commission is obliged to react to the consultation of Member State experts and will state how it will take the experts' views into consideration; and -Member State experts have the right to give a new opinion in case of a change in the material content of a draft delegated act. Consequently, Member States' opposition to delegation may be expected to diminish and, thus, the choice between delegation and implementation may be expected to be depoliticized to a certain extent. But the Commission had wished for more -and, for that matter, the European Parliament as well -and tried to introduce a more substantive delineation of delegation and implementation. To that end, it identified a set of objective criteria that should guide the choice between the two forms of secondary legislation. In its proposal for a new IIA 37 the Commission identified a number of criteria that in fact directly flow from the Treaty system, such as the rule that measures designed to lay down additional substantive rules and criteria to be met can only be adopted by way of delegated acts as they supplement the basic act. 38 Other criteria would indeed have clarified and further elaborated the existing legal framework, such as the rule that 'the annual and multiannual work programmes implementing financial instruments should be adopted by means of implementing acts'. However, all of the Commission's criteria have been removed from the final version of the text of the IIA, including those that in fact add little to the existing legal framework. Even a vague commitment to negotiate substantive criteria between the three institutions following the entry into force of the IIA has not made it to the final text. 39 The legitimacy of EU secondary legislation is not only relevant from an interinstitutional perspective. The involvement of stakeholders and public consultation are increasingly seen as valuable elements of the decision-making on delegated acts in particular. 40 The IIA, however, brings little new in this regard, focused as it is on Member State expert participation. It only provides that decision-making on delegated acts 'may include' consultations with stakeholders. 41 Experiences with obligations to that effect that already exist with regard to the powers of financial agencies of the EU have thus not found their way to the IIA.
More in general, the IIA completely ignores the position of EU agencies in EU law making. The EU financial agencies that were set up in 2011 have been given quasi-regulatory powers and are authorized to propose draft delegated and implementing acts. The status of such draft acts is strengthened by legislative constraints on the discretion of the Commission. The Commission may not simply ignore such draft acts. Thus, the existence of such quasi-regulatory powers raises questions as to their exact status and scope, their legitimacy and the implications for the EU institutions. These questions have remained unaddressed by the IIA. This is all the more surprising since the basic Treaties have failed to recognize the status of EU agencies as well. 42 Thus, no constitutional framework exists to regulate the position of agencies as recipients of (quasi-)delegated or (quasi-)implementing powers. The gap between the constitutional provisions and the actual institutional practice is only widening. 43 We will analyse the effects of this gap on the position of the EU agency in the field of pensions: the European supervisory agency EIOPA (European Insurance and Occupational Pensions Authority

Resistance to executive rule-making
Article 30 of the Commission's proposal to revise the IORP II Directive contains a legal basis for delegation for the Commission. Under this provision, the Commission may adopt a delegated act specifying, inter alia, the remuneration policy, the risk assessment for pensions and the pension benefit statement. Moloney 45 concludes that delegation provisions are usually drawn up in very general terms and that it is therefore difficult to see how they relate to the requirements of Article 290 TFEU. Moloney's view is inter alia supported by the European Economic and Social Committee (EESC). In connection with the MiFid II proposal 46 the EESC objected to the 'extreme and disproportionate use of delegated acts', as provided by Article 94 of the proposal. 47 According to the EESC, delegated acts should relate to specific and well-defined affairs and should be exercised within a certain period of time.
In the proposal for the MiFIR Regulation 48 the EESC 49 notes that there is a compatibility problem between Article 40 of the draft Regulation (the legal basis for delegation) and Article 290 TFEU. 50 The number, content and provisions of the delegated acts are not consistent with the provisions of the Treaty and place too many fundamental aspects of the Regulation outside the scope of the normal legislative process, according to the EESC. The EESC therefore recommends examining how Article 40 of the proposal can be aligned with the requirements of Article 290 TFEU.
The EESC was also critical of the proposed legal bases for delegation in the IORP II proposal. 51 The EESC urged the Commission to be particularly cautious when drawing up the delegated act mentioned in Article 54 of the proposal (regarding requirements for the pension benefit statement), 'given the potential costs of such a solution'. The Committee therefore suggested that when listing information requirements for scheme members, the Commission should take into account the nature of the schemes which they belong to. 52 By doing so, the EESC essentially said that the Commission should be careful not to affect the essential elements of the rules. As to making information available to members and beneficiaries of pension funds, the EESC expressed concerns about the suitability of a statement of standardised information, and proposed that more experience should be gathered before the delegated act could be carried out. 53

EIOPA and executive pension regulation
As stated in Section 2, the EU financial agencies have powers to draw up draft rules in preparation of delegated acts and implementing acts. 54 EIOPA 55 has such powers too. They are somewhat controversial, mostly due to the fact that EIOPA deals with rather sensitive issues and seems to act, sometimes, on its own initiative.
On its own initiative, EIOPA is currently working (in connection with the implementation of the current IORP Directive) on tightening the requirements for the solvency of pension institutions. 56 These solvency Utrecht Law Review | Volume 12 | Issue 1 (January) 2016 requirements, set out in the 'Holistic Balance Sheet (HBS)', entail a mechanism that should facilitate a clearer comparison between European pension funds. 57 In the course of its consultations EIOPA reiterated the Commission's statement that more technical information on the solvency of pension funds was required. 58 EIOPA acted on this by starting these consultations on its own initiative, and emphasized that there was no connection with the Commission's proposal for IORP II. PensionsEurope, a representative body of national pension funds, has been critical of the EIOPA's plans for the HBS, and has even referred to the plan as being 'conceptually wrong.' It claims that the HBS is not an adequate regulatory instrument and moreover entails an 'unacceptable burden' for IORPs. 59 Aside from the HBS, EIOPA issues several reports and guidelines, including ones on the costs of IORPs 60 and individual value transfers between Member States. 61 This last report 62 identifies eight bottlenecks for cross-border individual value transfer for which EIOPA offers 'good practices'. These should be observed by the Member States.
These EIOPA activities do not automatically lead to acts with legally binding effect. This could however be the end result, either through the mechanism of Article 290/291 TFEU, or otherwise (see the next section). This could give EIOPA a greater role than is perhaps assumed. This will be explained in what follows.

Further capital requirements for IORPs?
As said, under the current measures of the European Commission and EIOPA, and based on the current draft text of the IORP II Directive, no new capital requirements seem to be imposed on IORPs. 63 Nevertheless, as stated at the end of the preceding section, it cannot be ruled out that such requirements will eventually come into the picture. Three developments can be mentioned.
First of all, Article 30 of the draft IORP II Directive empowers the European Commission to adopt implementing acts 'in cases not foreseen by this Directive'. 64 In practice, the differences between the national capital requirements are a complicating factor when it comes to the cross-border provision of pension schemes. Precisely these differences could be a reason for applying Article 30 and imposing capital requirements. This is affirmed by the wording of the recitals of the IORP proposal that state that it is principally the sponsoring undertaking rather than the pension institution that either covers the longevity risk (for example) or guarantees certain benefits or certain investment performance. According to the this recital, however, in some cases it is the pension institution itself which provides such cover or guarantees and the sponsor's obligations are generally exhausted by paying the necessary contributions. In these circumstances, the recital continues, the products offered are similar to those of life-assurance companies and the institutions concerned should hold at least the same additional own funds as life-assurance companies. 65 Accordingly, although the scope of application of the legal basis for delegation is fairly broad, no essential elements of the regulation are in dispute in this sense. Partly in light of the recitals, the separation between legislation and executive rule-making would thus not be jeopardized.
In this light it is -in se -at least doubtful whether removing the delegation provision of Article 30 IORP II Directive would result in a situation in which no capital requirements can be imposed on IORPs.
Utrecht Law Review | Volume 12 | Issue 1 (January) 2016 EIOPA also has at its disposal the powers set out in Article 8 of Regulation 1094/2010 (the EIOPA Regulation). 75 They are fairly broad, especially in relation to insurers and include a) advisory and coordinating powers, b) the development of drafting standards and c) independent supervisory powers. Article 1(2) of this Regulation moreover provides that EIOPA will act within the powers conferred by this Regulation and within the scope of secondary EU legislation (including the Solvency II Regulation and the IORP Directive), but also 'including all directives, regulations, and decisions based on those acts, and of any further legally binding Union act which confers tasks on the Authority'.
The EIOPA Regulation moreover provides for a mechanism to resolve differences of opinion between national supervisory authorities. 76 Where a national supervisory authority takes issue with the procedural or substantive terms of a measure taken (or not taken) by another national supervisory authority, EIOPA must assist the authorities at their request in reaching agreement within the timeframe determined by EIOPA. 77 If the dispute continues, EIOPA must be able to settle the matter. 78 These dispute resolution powers are broad and can therefore also extend to the interpretation of delegated acts. Via these powers too, the capital requirement can therefore fall within the scope of EIOPA's powers.

Conclusion and closing remarks
The distinction between legislative and non-legislative acts is not only relevant from a constitutional perspective, but defines concrete legislative dossiers as well. In EU pensions law a key issue is whether capital requirements for pensions institutions may be adopted by executive regulation. As we have contended, various legal bases and options exist that may allow for this. At least, they would not a priori prohibit such rules. Yet, given the political controversy it would be likely that setting such requirements would amount to regulating essential elements of the area. Indeed, the making of political decisions would be involved which is, according to CJEU, reserved for the legislature, albeit that the CJEU would look at the actual substantive political interests at stake, rather than at the existence of political controversy.
The case of pensions law illustrates, moreover, that the demarcation between legislative and executive power is not only a matter of institutional powers, but affects the vertical power balance between the Member States and the European Union. This is a strong argument in favour of CJEU control over the issue at a level that is higher than the one we find in some domestic legal systems.
The distinction between delegated and implementing acts plays a role as well in the national opposition against delegated acts. Some governments have issued a critique on the proposed Article 30 of the draft IORP Directive, which provides for the power to adopt delegated acts. 79 Arguably, the possibility that the Commission may invoke this provision to impose capital requirements on pension institutions is especially worrying for Member States in light of the lack of ex ante control via committee procedures (comitology) that does exist under implementing acts. Here, the new Interinstitutional Agreement on Better Law-Making may make a difference on the point of the consultation of Member State experts. In any case, the key element here is Member State control. The constitutional discourse on the nature of the difference between delegated and implementing acts seems to be playing no role whatsoever in the decision-making process on the Directive.
Utrecht Law Review | Volume 12 | Issue 1 (January) 2016 The political controversy over Article 30 of the Directive is relevant from another perspective as well. The legal basis for the adoption of delegated acts has become the focus of attention with the opposition to capital requirements. As we have argued, however, such capital requirements may, however, acquire an EU dimension in other ways than Article 30 alone. This seems to remain ignored. Thus, we may expect that the political salience of Article 290 TFEU will remain to be higher than Article 290 TFEU despite the horizontal rather than vertical divide between them.
Finally, the role of EU agencies: this article has argued that EIOPA could have an important influence on the solvency requirements that may be imposed on pension institutions. This is not only due to the extensive quasi-regulatory powers that EIOPA possesses on the basis of numerous legislative acts, but is also due to its other powers (such as dispute resolution powers). In fact, this scenario is quite likely in view of EIOPA's actual activities in recent years. The lacuna in regulating EU agencies' position in rule-making in a more general way is thus manifested as an ever increasing problem. 80 The role of the EIOPA is now being developed in an ad hoc way. The new Inter-Institutional Agreement on Better Law-Making fails to address the issue, however.
In light of the controversy over executive rule-making powers, it would make sense if capital requirements for IORPs were directly and effectively regulated by the legislature itself. This would strengthen their legitimacy and be in line with the case law of the CJEU. As an essential element of the IORPs, capital requirements can also be regarded as an essential element of the regulation itself and as such they belong to the legislative domain.