SOFT LAW, SELF-REGULATION AND CO-REGULATION IN EUROPEAN LAW: Where Do They Meet? Linda Senden Readers are reminded that this work is protected by copyright. While they are free to use the ideas expressed in it, they may not copy, distribute or publish the work or part of it, in any form, printed, electronic or otherwise, except for reasonable quoting, clearly indicating the source. Readers are permitted to make copies, electronically or printed, for personal and classroom use. Abstract Over the past decade, the EU has been developing a new regulatory policy, which increasingly puts emphasis on the use of alternative instruments or on instruments that are complementary to traditional command-and-control legislation. This aim of diversification of the Union’s regulatory instruments is fundamentally inspired by the concern to enhance the effectiveness, legitimacy and transparency of EU action. These alternative instruments - including inter alia recommendations and voluntary agreements - are often labelled with the general terms of ‘soft law’, ‘self-regulation’ and/or ‘co-regulation’. This article is aimed at providing a general insight into the meaning which these concepts currently have within the context of the EU; it discusses the legal framework for the use thereof and touches upon some possible effects in terms of legitimacy. Furthermore, it addresses the interconnectedness of the phenomenon of soft law on the one hand and of self-regulation and co-regulation on the other. In this respect, particular attention is given to the course taken within the framework of the White Paper on European Governance 2001, the 2002 Commission Action plan ‘Simplifying and improving the regulatory environment, the Interinstitutional Agreement on better law-making of 2003 and the Treaty establishing a Constitution for Europe adopted by the Member States in 2004. Contents 1. Introduction 2. The two pillars of the Union’s legislative policy 2.1 The development of a ‘new legislative culture’ 2.2 The first pillar: ‘Do less in order to do better’ 2.3 The second pillar: Diversification of modes of governance 2.4 The underlying aim of enhancing the Union’s legitimacy 3. The legal framework for European self-regulation and co-regulation 3.1 Conceptualising European self-regulation and co-regulation 3.2 Some important manifestations of European self-regulation and co-regulation 3.3 The development of a general legal framework: The Interinstitutional Agreement on better law-making 3.4 Effects in terms of legitimacy 4. Interconnectedness with Community soft law 4.1 The concept, classification and functions of Community soft law 4.2 The case of recommendations and their link with self-regulation 5. Concluding remarks 1. Introduction In 1998, the Conclusions of the ECOFIN Council Meeting of 1 December 1997 concerning taxation policy were published, to which was attached the ‘Resolution of the Council and the Representatives of the Governments of the Member States, meeting within the Council of 1 December 1997 on a code of conduct for business taxation’. In the later adopted Communication from the Commission ‘Tax Policy in the European Union - Priorities for the years ahead’, it was stated more generally that ‘the use of non-legislative approaches or “soft legislation” may be an additional means of making progress in the tax field’. Reference is made in this respect to instruments such as communications, recommendations, guidelines and notices. More recently, on 1 October 2004, the European Advertising Standards Alliance (EASA) presented its new code of conduct to a group of Commission officials, consumer groups and trade associations. This code of conduct is said to provide ‘basic principles in ethical standards for advertising’ and it states that self-regulation ‘can provide appropriate redress for consumers, a level playing field for advertisers, and a significant step towards completing the Single Market’. The above examples raise the basic question as to how the use of such soft law and self-regulation instruments fit into the broader European legislative and regulatory framework. To begin with, it can be observed that, by the end of the last decade, the European Commission started a more fundamental debate on better EU governance, which was very much inspired by the institutional crisis witnessed by the Commission in 1998. Better regulation is also a key part of the objective set at the Lisbon European Council Summit in 2000, of making the EU the world′s most competitive knowledge-based economy by 2010. In July 2001, the debate culminated in the adoption of a White Paper on European Governance. In essence, this debate centres on the extent to which the traditional - supranational and top-down - Community command-and-control method is still the right way to proceed, and what new forms of European governance - intergovernmental and non-governmental - should be explored and promoted with a view to ensuring good governance in the EU. More in particular, the White Paper takes as one of its starting points that ‘[t]he Union must renew the Community method by following a less top-down approach and complementing its policy tools more effectively with non-legislative instruments’ and that ‘legislation is often only part of a broader solution combining formal rules with other non-binding tools such as recommendations, guidelines, or even self-regulation within a commonly agreed framework’. The Lisbon European Council called on the institutions and the Member States to establish a coordinated strategy for the simplification and improvement of regulation, which was subsequently put forward by the Commission in June 2002 in its Action Plan on ‘Simplifying and improving the regulatory environment’. This Action plan deals with a broad range of modes of governance, including in particular the use of soft law (recommendations), co-regulation, voluntary sectoral agreements, benchmarking, peer pressure, networks and the open method of co-ordination. Next, the European Parliament, the Council and the Commission reached an agreement on the conclusion of the Interinstitutional Agreement on better law-making, which was adopted on 16 December 2003. This agreement also addresses the use of alternative methods of regulation. Of particular interest for our purpose is that it sets out, for the first time, the general framework and conditions for the use of co-regulation and self-regulation mechanisms within the EU context. Finally, the issue of simplification of the Union’s legal instruments was also put on the agenda of the European Convention, convened on the basis of the Declaration adopted by the European Council of Laeken. Concrete proposals have been made for a redefinition of these instruments in the Treaty establishing a Constitution for Europe, on which the EU leaders reached agreement at the European Summit of 17 June 2004 and which was signed in Rome on 29 October 2004. All these developments fit in with the more gradual development of a new legislative policy, set in motion already in the second half of the 1980s. Taking these recent developments as the point of departure, the main aim of this article is to provide a general insight into the meaning of self-regulation, co-regulation and soft-law means and mechanisms within the context of the EU, and to see within what legal framework use is made thereof. It will also touch upon the effects of this framework in terms of legitimacy. This entails that the reasons why, the ways in which and the conditions under which use is made of self-regulation, co-regulation and soft law are investigated (sections 3 and 4). The article also seeks to establish where these alternative modes of European governance meet, or, what connection can be established between the use of soft law, self-regulation and co-regulation in the context of the EU (section 5). The article will start, however, by considering more closely the main foundations of the legislative policy which the EU can be said to now conduct and the alternative modes of governance that fit into this policy (section 2). For it is against this background that the increasing use of self-regulation, co-regulation and soft law is to be regarded. An observation that needs to be made at this point is that the focus will be on how self-regulation, co-regulation and soft law are used as instruments or modes of governance at the European level, not on how these are used at the national levels as a means of implementing European law into the national legal order. For instance, the question whether the use of national environmental agreements constitutes an appropriate means of implementing an EC directive will not be dealt with here. 2. The two pillars of the Union’s legislative policy 2.1 The development of a ‘new legislative culture’ In an article published in 1986, Bruha and Kindermann concluded that EC legislation was still ‘Terra incognita der Gesetzgebungslehre’. According to these authors, a specific Community theory of legislation or legislative policy had not even started to develop. However, the mid-1980s can be seen as a turning point with regard to the way of thinking on European legislation. At that moment in time, the stagnation of the internal market, the national deregulatory tendencies and the criticism of both the quantity and the quality of the body of European legislation constituted a catalyst for the EC to reconsider its legislative task, taking as its starting point the White Paper for the Internal Market of 1985 and the Single European Act of 1986. Further reflections on the existing body of European legislation and new legislation to be adopted and the burden it imposes on national authorities and companies have led to deregulatory and self-regulatory tendencies also at the EC level. Particularly since the early 1990s it has become clear that European legislative policy rests on two main pillars, which in a sense can be seen as each other’s logical counterparts. The first pillar represents the aim to make less use of the instrument of legislation and to reduce the existing body of European legislation. Improvement of the quality of European legislation is also a point of concern here. The second pillar represents the aim to make more use of other modes of governance or regulation, which are of a less compelling or non-governmental nature. In short, such a policy thus aims, on the one hand, at less and better legislation and, on the other, at more diversified European governance mechanisms. This has brought the European Commission to speaking of a ‘new legislative culture’. This culture is very much inspired by the notions of flexibility and differentiation. The concept of flexibility was introduced into the EC Treaty and the EU Treaty in the form of provisions enabling closer cooperation between only a number of Member States. This means that not all the Member States have to agree in order to move forward in respect of a certain matter. More interesting for our purposes, however, is that there is an ambition to arrive at flexibility and differentiation not only from a substantive point of view, but also from an institutional or instrumental one, i.e. when it comes to the modes or instruments by which European integration is to be given shape. More generally, the Treaty establishing a Constitution for Europe now states in its Article I-8 that the motto of the Union is ‘United in diversity’. When we look more closely at the purpose of European legislation, it also becomes very clear that uniformity is not always aimed at and that the aim is rather to establish harmonisation or alignment of national law and policy as far as necessary (in particular with a view to realising the internal market). Increasingly, it now appears that such binding harmonisation legislation may not even be considered necessary, but only - non-binding - coordination of national policies. The pillars of this new legislative policy are also firmly rooted in the principles of subsidiarity and proportionality, laid down in Article 5 EC. The Edinburgh European Council Conclusions, adopted in 1992, demonstrate this, as will be further explained in subsection 2.3. 2.2 The first pillar: ‘Do less in order to do better’ When we look more closely at the way in which the first foundation of the Union’s legislative policy has developed, it becomes clear that simplification and deregulation are the key words for putting the device ‘do less in order to do better’ into practice. This was underlined already in the Molitor Report, according to which simplification must be taken to mean ‘. . . that it is essential to ensure that regulation imposes the least constraint on competitiveness and employment whilst maximizing the benefits of direct government intervention’. Deregulation is taken to mean ‘. . . in some instances, an unavoidable extension of simplification will be the reduction or removal of government regulations, where such regulations are no longer necessary, or where their objectives can be achieved more effectively through alternative mechanisms’. There are thus in fact two sides to this first pillar: a drafting aspect and a policy aspect. As regards the drafting aspect, numerous proposals have been made and initiatives taken by the Community institutions in order to improve the quality of European legislation. In 1997, Declaration no. 39 concerning the quality of the drafting of Community legislation was attached to the Treaty of Amsterdam, pursuant to which the Interinstitutional Agreement on common guidelines for the quality of drafting of Community legislation was adopted on 22 December 1998. In legal writing, it has been considered that these guidelines are still rather limited, compared in particular with national guidelines. More recently, however, it has also been concluded that the existing European guidelines are essentially the same as the national ones and that the EU shares more or less the drafting philosophy of the Member States. The question no longer is considered to be whether the EC drafting rules as such are sufficient, but rather whether they are sufficiently applied. Without going into detail on this, it is clear that the efforts to improve the quality of European legislation continue to be made today. The aforementioned 2002 Commission Action plan ‘Simplifying and improving the regulatory environment’ and the Interinstitutional Agreement on better law-making, published on 31 December 2003, already testify of this, as well as the 2003 Commission Communication ‘Updating and simplifying the Community acquis’. The Commission considers the Interinstitutional Agreement to be the most ambitious effort undertaken up to now for realising better regulation, by uniting the three decision-making institutions in a joint global strategy to ensure better European lawmaking, while respecting the responsibilities of each institution. As will be seen below, this Agreement is also highly important when it comes to the conditions under which self-regulation and co-regulation are considered to be appropriate alternatives to European legislation. As regards the policy aspect, the striving for deregulation definitively acquired its place at the European level, because it explicitly aims at limiting legislative activity to what is necessary. This aim is given shape in various ways, as regards both existing legislation and new legislation. As regards the existing body of European legislation, an operation of ‘up-dating the stock of existing legislation’ and of ‘reducing the volume of the Community acquis’ has been set in motion, entailing not only the consolidation and codification of European legislation but also the removal of obsolete legislation. These initiatives contribute to reducing the complexity of the European body of legislation and to enhancing its accessibility. Consolidation is a semi-official editorial compilation of the various legal texts concerning a particular issue, which takes place outside the formal decision-making procedures. It has no legal consequences and leaves the legal force of the various texts intact. Codification occurs when a formal legal act is adopted, such as a regulation or a directive, on the basis of the prescribed procedures, by which all earlier texts are repealed and replaced by one new text that, in principle at least, does not alter the original contents. A major achievement in this respect is the realisation of the Community Customs Code. Already in 1994, an interinstitutional agreement was adopted to speed up the process of codification. This process involves also repealing those parts of the legislation that have become obsolete or invalid. A first initiative in this respect concerns the SLIM project: Simpler Legislation for the Internal Market. The outcome of this pilot project led the Commission to conclude that the project should be extended to other areas of European law as well, and that European legislation should be screened in a more structural way. Given the rather slow progress that was achieved in this respect, the Commission set itself a very clear target in 2001 of reducing the volume of the acquis by 25 per cent by the end of 2004, a reduction which corresponds to about 22,500 pages in the Official Journal. In particular as a result of the enlargement process and the translation efforts this entailed for the acquis communautaire, it has recently become clear that this goal is not likely to be reached. Yet, the Commission deems it feasible to realise this ‘within a reasonable period thereafter’. 2.3 The second pillar: Diversification of modes of governance As regards new legislation, the striving for deregulation entails a search for other possible modes of governance or alternative means of regulation besides legislation: the second foundation of the Union’s legislative policy. The European Council of Edinburgh laid the basis for this policy in December 1992, making clear that the foundations of the new legislative policy are rooted in the principles of conferred powers, subsidiarity and proportionality, laid down in Article 5 EC. These principles determine not only the competence of the EC to act, but also the intensity of its actions. As such, they are guiding principles for the conduct of its institutions, in particular for their choice of instruments. Every new legislative proposal is thus to be preceded by a review in terms of competence, subsidiarity and proportionality. The application of the principles of subsidiarity and proportionality may further require the preference for other ways of regulation over legislation. More in particular, the Edinburgh European Council Conclusions stated that, whenever possible, action has to be taken at the national level, be it by other ways of cooperation between the Member States, the use of voluntary codes or self-regulation. If European measures are deemed necessary, then non-binding measures such as recommendations should be used, if possible. If legislation is considered necessary, resort should preferably be taken to - framework - directives, not to regulations. Since the European Council of Edinburgh, this point of view has been confirmed on various occasions, and most importantly in the Protocol on subsidiarity and proportionality attached to the Treaty of Amsterdam. This Protocol explicitly confirms the guidelines laid down in the Edinburgh Conclusions, emphasising also that, consistent with the achievement of the objective, ‘the form of Community action shall be as simple as possible’, that ‘the Community shall legislate only to the extent necessary’ and that ‘Community measures should leave as much scope for national decision as possible’. The Treaty establishing a Constitution for Europe also contains such a Protocol (Protocol no. 2), in which it is stated that decisions are to be taken as closely as possible to the citizens. Furthermore, as was observed in the Introduction to this article, in the 2001 White Paper on European Governance, the 2002 Commission Action plan and the 2003 Interinstitutional Agreement on better law-making the determination to resort, where possible, to self-regulation, co-regulation and soft-law instruments and mechanisms has been confirmed. 2.4 The underlying aim of enhancing the Union’s legitimacy Application of the notions of flexibility and differentiation and of the principles of subsidiarity and proportionality is not an end in itself; this is considered to contribute to enhancing the effectiveness, legitimacy and transparency of Union action. The White Paper on European Governance makes this explicitly clear in respect of differentiation of the Union’s modes of governance and legal instruments. It is thus understood that this enables rulemaking closer to the citizen and realisation of the aims of the Lisbon strategy, respectively contributing to the legitimacy and the effectiveness of EU action. Focusing here on the aim of enhancing the Union’s legitimacy, we should first establish what is actually meant by legitimacy before we can consider what effects the use of self-regulation, co-regulation and soft law may have on this. The notions of ‘democracy’ and ‘rule of law’ are of crucial importance in this respect. Even if the European legal system constitutes a legal order in its own right, it is clear that it is based on these notions, just like the national legal systems represented in it. This means that not only the existence and division of European power must be acceptable to the citizen but also the exercise thereof. Democratic organisation and exercise of power is usually considered to be the basis for this acceptability and hence for the democratic legitimacy of the Union. So, enhancing legitimacy is primarily understood as an effort to increase the influence, control and participation of the European Parliament and, more generally, of the citizen in the European decision-making process. Yet, acceptability of - the exercise of - state power requires more than that, and in particular reliance on the rule of law. The essence of a state, or any other entity vested with the exercise of government power based on the rule of law, is that government action is bound by the law. This can be said to require, on the one hand, governing sub lege and, on the other, governing per lege. Governing sub lege does not only mean governing on the basis of the law, i.e. that there is a competence-conferring legal basis (the principle of legality), it also means governing within the boundaries of the law, that is in conformity with certain principles on which a constitutional state is based and which indicate the limits to the powers to be exercised, with a view to ensuring the freedom and liberty of citizens. Governing per lege means that power should be exercised through the adoption of laws, inter alia with a view to ensuring legal certainty and equality. These requirements must also be considered to apply to the European legal system. Given that this system is founded on the principle of conferred powers, the intention was expressly not to endow it with general competence to act, and consequently the Treaty provisions indicate when the Community may act, by what institution, in what form and according to what procedures. This principle thus functions as the Community principle of legality and emphasises the limits and control of state power. As such, it reflects the classical, rather formal and procedural, conception of the rule of law, which puts the legitimacy of government action on a par with the legality thereof. Yet, like in the national legal contexts there has been a shift also in the European legal context from such a classical, liberal conception of the rule of law to a more democratic and social conception, in which the realisation and protection of general principles of law and fundamental rights have increasingly gained attention. Various authors have spoken in this respect of ensuring social legitimacy, of legitimacy granted by the rule of law and substantive legitimacy. Pescatore asserts that true, substantive legitimacy ensues from the adequate performance of the functions of government; legitimate power is understood to be the power that responds best to the expectations and needs of the public and that is capable of resolving the problems affecting it, i.e. that is best for the general interest. I too understand legitimacy in this broad way, which can in particular be said to imply the duty to ensure good governance, demanding compliance with principles such as legal certainty, equality and legitimate expectations. 3. The legal framework for European self-regulation and co-regulation Given the increasing references to the use of self-regulation and co-regulation at the European level, the first question that arises is how self-regulation and co-regulation should be conceptualised here (subsection 3.1). Starting from this conceptualisation, some important manifestations of European self-regulation and co-regulation that have occurred up to now can be considered and the specific legal framework, if any, in which they occurred; in particular, on what legal foundation, under what conditions and control mechanisms they occurred (subsection 3.2). Next, the questions whether a general legal framework for this use is desirable and what action has been taken in this respect will be dealt with (subsection 3.3). This issue will also be considered from the point of view of legitimacy, by looking at the safeguards that have been created in this regard. 3.1 Conceptualising European self-regulation and co-regulation In legal writing, European self-regulation and co-regulation have been described as ‘forms of interaction between Community processes and private actors’ and the common feature has been considered to be ‘the existence of some form of relationship between binding legislation and voluntary agreements in a particular area’. Yet, it appears from the Commission’s White Paper on European Governance, subsequently adopted Commission documents and in particular also from the Interinstitutional Agreement on better law-making, that the European decision-making institutions hold a slightly different view on this, in particular as regards their actual involvement in the establishment of self-regulation. In drawing on the views expressed in the White Paper and the Action plan, the Interinstitutional Agreement marks a step forward in the sense that it not only establishes a joint definition but also the contours of the legal framework within which co-regulation and self-regulation may be used, agreed upon between the three decision-making institutions. Point 18 of the Agreement defines co-regulation as . . . the mechanism whereby a Community legislative act entrusts the attainment of the objectives defined by the legislative authority to parties which are recognised in the field (such as economic operators, the social partners, non-governmental organisations, or associations). [emphasis added] Point 22 of the Agreement defines self-regulation as . . . the possibility for economic operators, the social partners, non-governmental organisations or associations to adopt amongst themselves and for themselves common guidelines at European level (particularly codes of practice or sectoral agreements). [emphasis added] It is further explained that this latter type of voluntary initiative does not generally imply that the Community institutions have taken a particular position, in particular if the area at issue is not covered by the Treaties or if the Union has not yet legislated in such an area. In the Commission’s Action plan 2002, it is explicitly held that ‘[u]nlike co-regulation, self-regulation does not involve a legislative act’. Where voluntary agreements already exist and may be appropriate to achieve the Union’s objectives, the Commission may consider it preferable not to make a legislative proposal or it may suggest, by means of a recommendation, for example, that such an agreement be concluded by the parties concerned to avoid having to pass legislation. These voluntary agreements are considered to constitute a form of self-regulation, unless concluded on the basis of a legislative act. In that case, they are considered to be a form of co-regulation, enabling the parties concerned to implement a specific piece of legislation. It thus appears that, in the context of the EU, co-regulation is regarded rather as an implementing mechanism, presupposing the prior adoption of a piece of European legislation. Consequently, co-regulation also presupposes the direct involvement of a public actor in this regulatory process, which is usually not the case with self-regulation. In line with this, one may also speak of co-regulation as a regulatory mechanism that is primarily considered or understood to take a ‘top-down’ approach. That is to say, its use implies that the European legislature first sets the essential legal framework, that the stakeholders or parties concerned then fill in the details and that public authorities, usually the Commission, monitor the outcome or that sometimes the European legislature validates those more detailed rules by turning them into binding legislation. Co-regulation can thus be seen rather as a complement to legislation than as an alternative to this. Finally, it can also be said to situate itself somewhere between legislation on the one hand and ‘pure’ self-regulation on the other, by constituting in fact some form of ‘conditioned self-regulation’. ‘Pure’ self-regulation, as this also flows from the above definition in the Interinstitutional Agreement, clearly follows a rather bottom-up approach, as it concerns a regulatory mechanism which is primarily initiated by the stakeholders themselves, resorted to independently of the prior adoption of a legislative act. Thus the use of self-regulation can also more readily be considered to be an alternative to the use of legislation, which is not to say at the same time that it can be seen as being detached from the law. Both the use of co-regulation and self-regulation take place ‘under the shadow of the law’, by having to meet certain legal standards, as the Interinstitutional Agreement also makes clear (see subsections 3.2, 3.3 and 3.4). One may wonder to what extent the use of the open method of coordination (OMC) fits into the above definitions of co-regulation and self-regulation, as it is sometimes considered a form of co-regulation in legal writing. Quite surprisingly, the Interinstitutional Agreement itself does not touch upon this method at all. The Commission’s White Paper on European Governance also treats it clearly as separate from co-regulation and self-regulation. The White Paper specifies that it is ‘a way of encouraging co-operation, the exchange of best practice and agreeing common targets and guidelines for Member States . . .’. It further states that the OMC may be backed up by national action plans and that it relies on the regular monitoring of progress to meet the targets set. Member States can thus compare their efforts and learn from each other’s experiences. It may be combined with a legislative approach, such as in the areas of employment, social policy and immigration policy, or be adopted where there is little scope for the adoption of legislation. As such, the Commission considers the OMC to be rather a complement to than a replacement for Community action. The distinction with co-regulation and self-regulation as described above seems to lie in the fact that the OMC concerns forms of non-binding policy coordination involving primarily public actors, both at the European and at the national level. Hence, the OMC distinguishes itself from co-regulation and self-regulation in the sense that these regulatory mechanisms involve, to a lesser or greater extent, rulemaking by private parties. 3.2 Some important manifestations of European self-regulation and co-regulation Over the last decade, we have seen various manifestations of European self-regulation and co-regulation, without an overall legal framework having been developed for such use until recently. Yet, it is also clear that in some areas a specific legal framework has been created for the use of self-regulation and co-regulation. We will take a brief look at these here and see how they fit into the above conceptualisation. A first manifestation of this occurred in the mid-1980s, as a result of the so-called New Approach Strategy, leading to the adoption of European - framework - directives laying down only the basic requirements regarding, for instance, safety and consumer protection and leaving the establishment of specified, technical standards to private European standardisation bodies (CEN, CENELEC). This occurred in particular in the area of the free movement of goods. If a product has been manufactured in adherence ro these standards, it is assumed that the directive requirements have been met. This process clearly concerns the further execution of prior legislation, and can thus be qualified as a form of co-regulation or at least conditioned self-regulation in the sense described above. Subsequently, inspired by the use in at least some Member States of social or collective agreements and of environmental agreements, a debate has started on the potential usefulness of such self-regulation and co-regulation mechanisms in the areas of European social and environmental law and policy. Meanwhile, this debate has led to a number of concrete results. Looking first at the area of social law, the European social dialogue has developed as from the mid-1980s, leading in 1991 to the insertion of Articles 138 and 139 into the EC Treaty as a result of the adoption of the EU Treaty. Article 138 EC provides in particular that the Commission is to promote the consultation of management and labour at the European level and that it will facilitate the social dialogue. Pursuant to this consultation process, management and labour may express the wish to initiate the procedure provided for in Article 139 EC, which enables them to establish contractual relations, including the conclusion of agreements. Such agreements can be implemented either by a Council decision, following a proposal from the Commission, if the agreement touches upon matters covered by Article 137 EC, or by management and labour and the Member States themselves, according to their own procedures and practices. In a recent Communication, the Commission took stock of this European-level social dialogue process to date, highlighting its achievements and the areas where improvement can be made. Quantitatively, the Commission notes that the social dialogue has resulted in the adoption of over 40 joint texts drawn up by cross-industry social partners and approximately 300 drawn up by sectoral social partners, whose topics are diverse and have widened over the past few years. These texts take a variety of forms, ranging from joint opinions to guidelines, codes of conduct, etc., but it is clear that only a few of them took the form of an agreement in the sense of Article 139 EC. In total, only eight such agreements have been concluded up to now, six of which were or are to be implemented by a Council decision (in practice, by means of directives). The responsibility for ensuring that these agreements are transposed and implemented lies with the Member States, also in the case of implementation through collective bargaining by the social partners. The Commission is responsible for monitoring these agreements. Two other voluntary agreements are implemented in accordance with the procedures and practices specific to management and labour and the Member States, which are typified as ‘autonomous agreements’: the cross-industry framework agreement on telework of July 2002 and the cross-industry agreement on work-related stress, concluded in May 2004. The social partners themselves are responsible for implementing and monitoring these agreements. Given that Article 139(2) EC states that the Community level agreements ‘shall be implemented’ (emphasis added), there is a legal obligation to implement these agreements, and the signatory parties will have to exercise influence on their members with a view to ensuring this. However, in the case of autonomous agreements which are the result of an Article 138 EC consultation, such as those at issue, the Commission has a particular role to play since the social partners’ decision to negotiate an agreement temporarily suspends the legislative process which the Commission initiates in this area. The Commission will publish autonomous agreements and inform the European Parliament and the Council of Ministers, after undertaking an ex-ante assessment as it does for agreements to be implemented by Council decision. Upon the expiry of the implementation and monitoring period, the Commission will also undertake its own monitoring of the agreement, to assess the extent to which it has actually contributed to the achievement of the European objectives. The Commission may exercise its right of initiative, also during the implementation period, if it is of the opinion that either management or labour are delaying the pursuit of these objectives. In conclusion, both types of agreements can be considered to constitute some form of co-regulation in the sense described in subsection 3.1, by involving at least the Commission in an initial, preparatory stage and at the monitoring stage or the Council at the implementing stage. Furthermore, the legal foundation for the use of this co-regulation mechanism is not provided by a secondary legislative act, but by an even more fundamental, primary Community law provision: Article 139 EC. It is also clear that the social partners have to act within the procedural and substantive framework set by Articles 137, 138 and 139 EC, which the Commission has further substantiated in its Communication, in particular by reserving its right of initiative if it is not satisfied with the action undertaken by the social partners and by stating that ‘where fundamental rights or important political options are at stake, or in situations where the rules must be applied in a uniform fashion in all Member States and coverage must be complete, preference should be given to implementation [of autonomous agreements] by Council decision’. In so observing, the Commission confirms in fact some of the more general legal standards which the three decision-making institutions have agreed upon for the use of co-regulation and self-regulation in the Interinstitutional Agreement (see subsection 3.3). In the area of environmental law, the Commission has expressed the wish to encourage the preparation of European environmental agreements in a wide range of sectors, not only in those where it has announced its intention to propose legislation. Yet, unlike the situation in the area of social law, it is clear that for this use not only a Treaty foundation is lacking, but also a general legal foundation in secondary legislation. Despite the fact that, already in its 2000 work programme, the Commission proposed the creation of such a legal basis in the form of a Community regulation, only a communication has been adopted until today: the Communication from the Commission of 17 July 2002 on Environmental Agreements at Community Level within the Framework of the Action plan ‘Simplifying and improving the regulatory environment’. This Communication seems to have been issued in particular in response to criticism voiced by the European Parliament and the Council, both urging the Commission to present a clearer procedural framework for the use of environmental agreements. The Parliament did so already in a resolution dating back to 17 July 1997, and later in a resolution of 3 April 2001, in which it called on the Commission ‘to present as soon as possible a proposal for framework legislation on environmental agreements, which lays down the relevant criteria with regard to conditions, monitoring arrangements and penalties’. It also held that the definition of precise objectives at Community level is a matter for the legislature and is not to be left to the goodwill of the industry within the framework of a voluntary commitment. In its resolution of 7 October 1997 on environmental agreements, the Council held ‘that . . . environmental agreements should be negotiated in accordance with procedures to be agreed’. Obviously, these observations have been induced by the concern that the decision-making competence which these two institutions have been assigned in the area of environmental law and policy on the basis of (now) Article 175 EC run the risk of being ‘hollowed out’ by the use of - voluntary - European environmental agreements. This risk is not imaginary, given the fact that such agreements are to be directed towards achieving the environmental objectives set in Article 174 of the Treaty. In view of this, the questions arise what legal framework the Communication actually establishes for the use of voluntary agreements and whether the Communication itself can be said to provide a sufficient legal basis for the use of such agreements, given that it constitutes a form of soft law. In this respect, it is of importance that the Communication distinguishes different forms in which environmental agreements can be cast, considering them either as instruments of self-regulation or as instruments of co-regulation. Agreements may in the first place be spontaneous decisions, initiated by stakeholders in an area where the Commission has not proposed legislation and has no intention to do so. The Commission may acknowledge such bottom-up self-regulation by issuing a recommendation or by a simple exchange of letters. A recommendation is considered to be a non-binding act by its very nature, which can only encourage economic operators that have committed themselves to reach an environmental objective as established in Article 174 EC. If the Commission and the European legislature have an interest in the results of an environmental agreement and, in particular, if they wish to monitor it closely, the recommendation may be accompanied by a Parliament and Council decision on monitoring. In fact, this is what occurred with the best-known examples of European-level environmental agreements concluded up to now, i.e. the agreements of the European, Japanese and Korean carmaker associations on the reduction of carbon dioxide emissions from passenger cars. The Commission acknowledged these through recommendations, and a decision of the European Parliament and the Council supplemented them, by establishing a monitoring scheme. The only sanction imposed is the threat of future legislation if the objectives are not achieved. The co-regulation mechanism which the Communication provides for follows the - top-down - Community method, by which the European Parliament and the Council adopt a directive upon the proposal of the Commission. Such a directive is to set out the essential aspects of the legislation, including the objectives to be achieved, implementation deadlines, the monitoring of the application of the directive and any sanctions. The legislature determines what implementing measures can be left to the parties concerned and thereby sets the limits for them to do so by the conclusion of environmental agreements. This form of co-regulation is thus initiated by the Commission, which it may do on its own initiative but also in response to voluntary action on the part of industry. Clearly, there is considered to be a need for a fairly strong legislative framework, setting out the targets and monitoring mechanisms for ensuring compliance, but it is left to the parties in the field to take the measures necessary for achieving these targets, by way of concluding an environmental agreement. The Communication also sets a number of basic conditions for the use of environmental agreements and assessment criteria. Environmental agreements must be in compliance in particular with the EC Treaty, international commitments of the Community and multilateral trade rules, respect the institutional balance in decision-making and ensure judicial control of compliance with the agreement; information on environmental agreements must be made accessible to the public and they must have added value in the sense of aiming at a high level of protection. It is also stated that the Commission will give ‘due attention’ to the following criteria: cost-effectiveness of administration, representativeness, quantified and staged objectives, involvement of civil society, monitoring and reporting, sustainability and incentive compatibility. If the agreement fails to deliver the expected results, the Commission can always make use of its right of initiative and propose legislation. Apart from these most prominent manifestations of European self-regulation and co-regulation in the areas of social and environmental law, occurring within a legal framework that was created specifically with a view to the use thereof, other manifestations of in particular European self-regulation can be identified, for which no such specific European-level legal foundation and framework has been called into being. The EASA code of conduct, mentioned in the Introduction to this article, is a clear illustration of this. Most of the initiatives taken within the framework of the European social dialogue, such as the adoption of joint opinions, guidelines, codes of conduct, etc., also fit in with this. 3.3 The development of a general legal framework: The Interinstitutional Agreement on better law-making The discussion in the previous subsection shows that European co-regulation and self-regulation take place in different settings which are situated between two extremes; they may be used outside any European-level legal framework and, on the basis of a specific legal framework, even in the EC Treaty itself. Co-regulation and self-regulation also occur on the basis of specific rules and conditions set out in secondary legislative acts (usually directives) or in ‘tertiary’, soft law acts such as communications. The question that arises in view of this situation is how to fit the use of self-regulation and co-regulation more generally into the overall European legal framework. The Interinstitutional Agreement on better law-making provides an answer to this question. For the first time, this Agreement has created a general legal framework for the use of co-regulation and self-regulation at the European level, providing a number of general rules and conditions their use has to comply with. To a certain extent, the Agreement reflects the experiences already gained with European co-regulation and self-regulation, as discussed above. Some of the general conditions which the use of both self-regulation and co-regulation has to meet are of a substantive nature, including the requirements that their use has to be consistent with Community law, that they must have an added-value for the general interest and that it may not affect the principles of competition or unity of the market. Of a more procedural nature are the requirements that their use must meet the criterion of transparency, implying in particular the publishing of agreements, and that the representativeness of the parties involved is ensured. Very importantly, there are two other conditions that go further than the aforementioned ones as they actually rule out the use of self-regulation and co-regulation in two situations: 1) where fundamental rights or important political options are at stake and 2) where the rules must be applied in a uniform fashion in all Member States. In those circumstances, self-regulation and co-regulation are held to be ‘not applicable’ (point 17). In addition, specific conditions may apply to the use of co-regulation mechanisms, as possibly defined already by the legislative act they purport to implement. The Agreement provides in this respect that co-regulation ‘may be used on the basis of criteria defined in the legislative act so as to enable the legislation to be adapted to the problems and sectors concerned, to reduce the legislative burden by concentrating on essential aspects and to draw on the experience of the parties concerned’ (point 18). The legislative act will also indicate the possible extent of co-regulation in the area concerned. Agreements between social partners must comply with the provisions laid down in Articles 138 and 139 EC. The parties affected by the basic legislative act may conclude voluntary agreements for the purpose of determining practical arrangements in the context defined by the basic legislative act. The Agreement also provides for some form of control of the use of self-regulation and co-regulation mechanisms and in particular of its consistency with the above conditions. In the case of co-regulation, the Commission must explain to the competent legislative authority its reasons for proposing the use thereof, in an explanatory memorandum to its proposals. Furthermore, the Commission must forward draft voluntary agreements to the legislative authority, which it will have checked as to their compliance with Community law (and, in particular, with the basic legislative act). A provision may also be included in the basic legislative act, at the request of, inter alia, the European Parliament or of the Council, providing for a two-month period of grace following the notification of a draft voluntary agreement to these institutions. They may then either suggest amendments, raise objections as to its entry into force or ask the Commission to submit a proposal for a legislative act. Apart from that, the basic legislative act will also define the measures to be taken in order to follow up its application in the case of non-compliance by one or more parties or if the agreement fails. Such measures may include a duty upon the Commission to inform the legislative authority on follow-up to application, a revision clause under which the Commission must report within a certain period of time or any legislative measure that is deemed appropriate. As regards self-regulation, the Commission must scrutinise whether self-regulation practices comply with the provisions of the EC Treaty and it will also notify the European Parliament and the Council as to whether it regards this use to be compatible with the general criteria of representativeness and of added value. Here also, the Commission explicitly reserves the right to put forward a proposal for a legislative act, in particular at the request of the competent legislative authority but also where it deems that self-regulation practices have not been observed. As observed in subsection 3.1, the Interinstitutional Agreement did not deal with the use of the OMC as an alternative method of regulation, so the foregoing does not concern the OMC. Yet, in the White Paper on European Governance the Commission listed a number of criteria for its use which bear some resemblance to the aforementioned ones. In particular, the use of the OMC may not upset the institutional balance and exclude the European Parliament from the European policy process nor dilute the achievement of common objectives in the Treaty. Overall accountability must be ensured in line with the following requirements: it must be used to achieve defined Treaty objectives, there must be regular mechanisms for reporting to the European Parliament, the Commission should be closely involved and play a co-ordinating role and the data and information generated should be widely available. Furthermore, it is considered that the OMC should not be used when legislative action under the Community method is possible. Interestingly, various working groups of the European Convention have dealt with the question as to whether the OMC should be defined and regulated in the new Treaty establishing a Constitution for Europe. Yet, no agreement could be achieved on this and thus the new classification of the Union’s legal instruments in Articles I-33 to I-37 of this Treaty does not contain such a definition nor does any other of the proposed Articles of the Treaty. The Treaty establishing a Constitution for Europe does not expressly deal with the use of co-regulation and self-regulation either, but confirms, on the one hand, existing forms thereof in, for instance, its Articles II-88 and III-212; on the other, it seems to enlarge its scope as a result of introducing different categories of Union competence in its Articles I-12 to I-17, in particular by providing for ‘areas of supporting, coordinating or complementary action’ in Article I-17. 3.4 Effects in terms of legitimacy Both in the area of social law and of environmental law, it would seem that the way in which co-regulation is given shape does not pose problems in terms of legitimacy and that in fact it may indeed be seen as a contribution thereto. Since co-regulation presupposes the existence of a legal basis in primary or secondary European law and the prior adoption of a legislative act in which the conditions for its implementation by social and environmental agreements are set out, not only the Commission but also the European Parliament and the Council are actually strongly involved in this process. This process also allows a balanced evaluation of what is to be established in binding legislation and what can be dealt with at a level which is closer to the stakeholders and actors concerned. In my view, this brings about that the fact that up to now no EC regulation has been adopted, setting out the general conditions and framework for the use of environmental agreements, is not harmful. Whether the use of voluntary social and environmental agreements in practice complies with the requirement of good governance should ultimately be assessed on a case-by-case basis, but it appears that there is certainly attention for this requirement, given the boundaries and additional criteria that have been formulated for the use of both social and environmental agreements. In the case of self-regulation acts, such as certain environmental agreements and codes of conduct in the area of social law and policy, the case may be different since these may be precisely directed towards achieving objectives defined in the EC Treaty (e.g. in Article 174) and in respect of which the EC has thus been attributed certain decision-making competence. The European Parliament and the Council should then at least be informed about and involved in the Commission’s intention and decision to leave particular matters to be dealt with by the stakeholders themselves, since this decision will affect their decision-making powers. What then, does the Interinstitutional Agreement add to this? It can be seen as a step forward and as a contribution to the legitimation of European law- and rulemaking in various respects. First of all, the Interinstitutional Agreement represents a joint and strong commitment of the three decision-making institutions to better coordination and cooperation not only in the legislative process, but also in the use of alternative methods of regulation. Secondly, the Interinstitutional Agreement provides a number of general safeguards against the uncontrolled use of alternative methods of regulation in whatever area - including those in which a specific legal framework for the use of co-regulation and self-regulation is lacking - which might affect the position of the legislative authority and in particular that of the European Parliament. Even if an automatic right of ‘call-back’ was not established in the Agreement, the legislative authority is involved in such a way that it can express its objections to the use of co-regulation which the Commission cannot easily circumvent. As such, the Interinstitutional Agreement confirms and completes at a more general level the boundaries and conditions under which use can be made of self-regulation and co-regulation, as already defined in the areas of social and environmental law and policy. With regard, for instance, to the use of self-regulation environmental agreements, the Interinstitutional Agreement can be said to provide more safeguards. Apart from that, establishing explicitly the cases in which self-regulation and co-regulation may not be used can also be seen as a contribution to upholding democratically legitimate law-making where this is essential. A final remark must be dedicated here to the legal status of the Interinstitutional Agreement itself; even if it represents a joint and strong commitment, does this also mean that it is to be considered a legally binding commitment? It is not possible to state in general terms that interinstitutional agreements have legally binding force or not, since under present Community law their status is not regulated and much therefore depends on the interinstitutional agreement at issue. As regards the Interinstitutional Agreement on better law-making, its having at least binding force inter partes can be defended on the basis of two arguments. Firstly, it contains a number of rather compelling terms (‘agree’, ‘will’), which can be said to express the intention of the institutions to enter into a binding commitment. A confirmation of this intention can also be seen in its points 37 and 38 on the implementation and monitoring of the Agreement, providing, inter alia: ‘The three Institutions will take the necessary steps to ensure that their staff have the means and resources required for the proper implementation of the provisions of this Agreement’ (point 38). Secondly, where ‘agreed acts’ are specifically intended to reinforce interinstitutional cooperation such as the Interinstitutional Agreement at issue here, it can be argued that there is a specific duty of cooperation which in conjunction with the duty of sincere cooperation laid down in Article 10 EC may actually lead to the conclusion that such an agreed act must be considered binding upon the concluding parties. The Declaration on Article 10 EC, attached to the Treaty of Nice, clearly also establishes a link between this duty and the adoption of interinstitutional agreements, by providing expressly for the possibility of the European Parliament, the Council and the Commission of adopting interinstitutional agreements by common accord whenever this appears necessary for the facilitation of the application of the EC Treaty provisions, even if not stipulating - yet - that these are to have binding force. Interestingly, the Treaty establishing a Constitution for Europe now appears to contain this conclusion, by not only providing a legal basis for the use of interinstitutional agreements but also by recognising their possible binding force in Article III-397. Yet, this provision does not make unequivocally clear when an interinstitutional agreement will be binding and when it will not. 4. Interconnectedness with Community soft law In the preceding sections, a number of instruments have been mentioned that can be labelled soft law: communications, codes of conduct, guidelines and recommendations. In certain instances at least, it has thus appeared that soft-law instruments are used to shape the processes of self-regulation and co-regulation. When we focus on how the Council and the European Commission make use of soft-law instruments in these processes in areas within the first pillar of the EU (i.e. the EC), it appears that the instrument of the recommendation shows best where the use of Community soft law may meet European self-regulation and co-regulation. Before dealing with this, the concept of soft law will be clarified, and in particular the different ways in which it presents itself and the different functions it can be said to fulfil. 4.1 The concept, classification and functions of Community soft law Clarification of the concept of soft law is needed because the term ‘soft law’ is used in many contexts and for a huge variety of instruments. In the context of the EC, legal writing points to three core elements of soft law. The first element is that it is concerned with ‘rules of conduct’ or ‘commitments’. Secondly, there is agreement on the fact that these rules or commitments are laid down in instruments which have no legally binding force as such, but are nonetheless not devoid of all legal effect. Thirdly, they aim at or may lead to some practical effect or impact on behaviour. On the basis of these elements, I propose the following definition of soft law: Rules of conduct that are laid down in instruments which have not been attributed legally binding force as such, but nevertheless may have certain - indirect - legal effects, and that are aimed at and may produce practical effects. The huge variety of instruments that may fall within the scope of the above definition makes it impossible to make general comments on the nature, function, possible legal effect and other characteristics of soft law, such as addressees, possible legal basis, etc. This does not mean that a classification of soft-law instruments would also be impossible. Even if soft-law instruments are not defined or regulated in any way, it appears possible to establish their core features by looking at the instrument itself, its actual contents and the intention of its drafters. With a view to assessing their possible use as an alternative or complement to legislation, a classification is best made on the basis of the function and objective the various soft-law instruments can be said to have. Even if we have to recognise beforehand that such a classification has its limitations, it enables us to identify the different roles that soft-law instruments play in the Community legal order and the instruments that can be connected to co-regulation and/or self-regulation. A first major category of instruments which are usually considered as soft law can be designated preparatory and informative instruments. Within this category fall in particular Green Papers, White Papers, action programmes and informative communications. These instruments are adopted with a view to preparing further Community law and policy and/or providing information on Community action. Notably, the European Parliament has stated that the instruments falling within this category are being used as alternatives to legislation. However, it may be questioned whether they actually constitute soft law at all, for they do not establish any rules of conduct themselves but only prepare the future adoption thereof in the sense that they are an element in the assessment of their desirability or necessity and possible contents. As such, they can also be regarded as fulfilling a pre-law function. A second category of soft-law instruments are the interpretative and decisional instruments. These instruments aim at providing guidance as to the interpretation and application of existing Community law. The interpretative instruments restate or summarise the interpretation that should be given to Community law provisions, also on the basis of the case law of the ECJ. The decisional instruments go further than mere interpretation by indicating in what way a Community institution - usually the Commission - will apply Community law provisions in individual cases when it has implementing and discretionary powers. To this category belong notably the Commission’s communications and notices and also certain guidelines, codes and frameworks frequently adopted in the areas of competition law and state aid. These instruments bear resemblance to the administrative rules or policy rules that exist in quite some national legal systems. They are usually not intended to substitute for legislation, but rather to complement it. As such, they can be considered to fulfil primarily a post-law function, being adopted subsequent to already existing Community law with a view to supplementing and supporting it. The third category covers what one could call - formal and non-formal - steering instruments. These aim at establishing or giving further effect to Community objectives and policy or related policy areas. Sometimes this is done in a rather political and declaratory way - in declarations and conclusions - but often also with a view to establishing closer cooperation or even harmonisation between the Member States in a non-binding way, as occurs in particular in recommendations, resolutions and codes of conduct. The recommendation constitutes a formal steering instrument; it is presented as a Community legal instrument in Article 249 EC. The other instruments are non-formal instruments, in the sense that they occur only in daily practice. To a certain extent at least, the latter instruments are used as alternatives to legislation and, in view of this, they can often be said to fulfil a para-law function. A variety of situations may hide behind this function. Sometimes, the adoption of such a soft-law instrument is intended to serve only as a temporary alternative to legislation and will in any event be replaced by legislation, which may in fact already be in the pipeline. Sometimes, however, it is intended to function as a permanent alternative to legislation, in particular where legislation is merely held out as a prospect in the event of unsatisfactory compliance or non-compliance with the soft law act. Yet, they may also merely perform the pre-law function, in the sense that they pave the way for the adoption of future legislation. That is to say, they may facilitate the subsequent adoption of legislation by providing or increasing the basis of support for the rules contained therein. 4.2 The case of recommendations and their link with self-regulation From the brief description and classification of soft-law instruments in the preceding subsection, it follows that the term ‘soft law’ is in fact an umbrella concept which encompasses a huge variety of different instruments with different characteristics and fulfilling different functions. This discussion makes clear that the third category of steering soft-law instruments, aimed at giving further effect to Community objectives and policy or related areas, shows links with the process of self-regulation as described in section 3. In particular, it has come to the fore that the instrument of the recommendation is a soft-law tool being used in this process, notably as a means of confirming or supporting self-regulation that is being undertaken at the European level. Apart from the statement of ‘no binding force’, Article 249 EC merely indicates that recommendations may be adopted by both the Commission and the Council, thus leaving the nature, objective, function and other characteristics of this Community instrument largely an open question. On the basis of the actual wording and contents of recommendations adopted hitherto, it can be concluded, however, that both Council and Commission recommendations aim at laying down general rules of conduct and are directed at influencing the behaviour of notably outside parties. They are thus of a general, normative and an ‘external’ nature. Furthermore, the aim is usually to lay down new rules, which are not necessarily linked to existing legislation or Treaty provisions and cannot be said to be inherent to the existing legal framework, or at least are not limited to this. Often, they are thus not only intended to function as an alternative to legislation, but they also display the characteristics required for actually being able to function as such. This conclusion fits in with the use the Commission makes of the recommendation as a means of confirming European self-regulation, which has also been seen to be directed foremost to functioning as an alternative to legislation (subsection 3.1). Yet, quite importantly, this does not mean that they pursue the same purposes as Community legislation. Only if recommendations seek to establish harmonisation of national law can they be said to be used with a view to realising the same objectives as legislation. In this respect, it has been found that, except for certain Commission recommendations, recommendations are usually confined to realising closer cooperation, coordination or a concerted approach among the Member States. More in particular, it seems that the choice for the instrument of recommendation not only says something about the legally binding force of the Community action thus put forward, but to a certain extent also on the intensity thereof; the aim of coordination established in a regulation or directive thus usually boils down to the unification or harmonisation of law, whereas coordination established by way of a recommendation is often merely aimed at the establishment or promotion of a better coordination of national policies. The foregoing also fits in with the use the Commission may make of recommendations in the context of European self-regulation, which has been seen to take place outside the legislative - and harmonisation of laws - framework. In my view, it would have been appropriate to specify the more limited role and purpose of the recommendation in the new sources catalogue in the Treaty establishing a Constitution for Europe. Such a clarification would be welcome from the point of view of increasing transparency and legitimacy of the Union’s action, given that the Union’s legal acts would then reflect better the different levels of action the Union may actually engage in (unification, harmonisation and coordination) and the competence it has been assigned (in particular exclusive competence, shared competence and supporting, coordinating and supplementing competence). Unfortunately, however, the European Convention did not seize the opportunity to make a proposal of this kind, and the new Article I-33 on the legal acts of the Union thus still merely states: ‘Recommendations and opinions shall have no binding force.’ Apart from that, it is very clear that Council recommendations are adopted within the legal framework of the Treaty; a particular legal basis is established for virtually all recommendations and the prescribed decision-making procedures are consistently followed. In most instances, it is thus clear that the Council acts on the basis of a proposal or draft recommendation of the Commission, generally followed by the opinion of the European Parliament, often also that of the Economic and Social Committee and occasionally that of the Committee of the Regions, which boils down to applying the consultation procedure. In some recommendations even the cooperation procedure of (now) Article 252 EC has been declared applicable and was thus followed. Furthermore, the Council also appears to take account of views expressed by outside parties, although this is not done in a consistent way. The Council may feel compelled to justify in each individual recommendation that it has thus been adopted in accordance with the Treaty and in particular that it falls within its scope, in view of the fact that the recommendation is a formal European instrument. Furthermore, there are far fewer Treaty provisions providing expressly for the adoption of recommendations by the Council than by the Commission, and there is no provision comparable to Article 211, second indentation, which is and can indeed be said to provide the legal basis for the adoption of Commission recommendations. In respect of the latter, it can thus be argued that there is less need for such express reference. Yet, the rather general power provided for in Article 211, second indentation, does not prescribe any involvement of the European Parliament, which may explain why Commission recommendations are not adopted following any particular decision-making procedure. Although the recommendations may be drafted on the instigation of another EC institution or body, calling upon the Commission for a particular action, the outside involvement in the determination of the rules actually laid down is thus limited to consultations that may have taken place in this respect. So, it appears that the way in which the Council proceeds in the adoption of its recommendations is surrounded by more democratic guarantees than the Commission’s way of acting, now that no general involvement of the European Parliament is provided for in the case of adoption of Commission recommendations. In my view, even if consultation of the European Parliament might be considered to run counter to the interest of flexibility of decision-making that may precisely underlie the Commission’s choice for this soft-law instrument, consultation of the European Parliament is justified from the point of view of the recommendation being a formal European instrument, which in most cases has been found to fulfil the para-law function. This leads to the conclusion that where the Commission thus intends to confirm or support European self-regulation by way of a recommendation, it should first consult the European Parliament on this. Such a requirement of consultation seems to be implied also in the Interinstitutional Agreement on better law-making, where it provides that the Commission is to notify the European Parliament and the Council as to whether it regards the use of self-regulation as compatible with the general criteria of representativeness and of added value, and where these institutions may actually request the Commission to submit a proposal for legislation (see subsection 3.3). In terms of accessibility, there appears to be no problem as both Commission and Council recommendations are published in the Official Journal of the EU, even in the L-series. This may be explained by the fact that the recommendation is a formal European instrument. 5. Concluding remarks The above discussion has brought to light a number of differences and links between the phenomena of European co-regulation, self-regulation and soft law. European co-regulation and self-regulation have in common that they involve private actors and public authorities, the latter to a larger extent in the case of co-regulation and to a lesser extent in the case of self-regulation. European co-regulation presupposes the prior establishment of a general legislative framework by the European legislature and thus also takes place within the scope of the Union’s competence. It merely leaves the further execution and implementation of this framework to the various private actors in the field concerned. Hence, co-regulation primarily aims at complementing legislation. European self-regulation occurs outside such a legislative framework, where there is deemed to be no need (yet) for legislation or where a European legal basis for legislation may be lacking. As such, it may rather be perceived as an alternative to legislation. Against this background, it has been established that in particular the soft-law instrument of recommendation may be used by the European Commission as a means of confirming European self-regulation processes, such as the adoption of environmental agreements. The Commission’s White Paper on European Governance and its Action plan ‘Simplifying and improving the regulatory environment’ culminated in the adoption of the Interinstitutional Agreement on better law-making, which provides inter alia a legal framework for the use of co-regulation and self-regulation. In this respect, it certainly contributes to enhancing the legitimacy of the Union’s action, by setting out in a general way the legal standards which use of the instruments of co-regulation and self-regulation has to comply with and in what situations these instruments may not be resorted to. It also provides for a stronger position of the European Parliament and, even if somewhat implicitly, for its consultation in the case of the use of self-regulation. At the same time, it may be regretted that the European Convention did not seize the opportunity to propose more transparent use of the instrument of recommendation in the Treaty establishing a Constitution for Europe by elucidating the role and purpose of this instrument in Article I-33.