The Legal Effects of European Soft Law and Their Recognition at National Administrative Courts
The Legal Effects of European Soft Law and Their Recognition at National Administrative Courts
The aim of this article is to analyse the extent to which European Union (EU) soft law is hardened in administrative court procedures, whether national courts recognise the legal effects of soft law, and can live with the coexistence of soft and hard law. References to EU soft law instruments at national level are most common in administrative law cases, especially in competition law, regulated markets, environmental law, and consumer protection law cases.
The first part of the study deals with the definition and classification of various EU soft law instruments. This part gives an overview of the origins of soft law and introduces the different types of soft laws that have become important and colourful threads in the acquis communitaire’s fabric.
Next, we summarise the rationale for the existence of soft law, looking at soft control within a hierarchy, enhancing transparency to the public, and maintaining consistent law enforcement. To better understand soft law, we provide analogies with the legal and economic principles of ius dispositivum.
Then, we explore whether formal and informal soft law instruments produce different legal and practical effects. The distinction between self-binding and external (soft) binding effects will be noted, and we will observe the extent to which law enforcers and judges can deviate from soft law norms. Sector specific insights from telecommunication regulatory cases and competition law enforcement lead us to conclude that, in the practice, EU soft law is treated as binding law before national administrative courts.
We suggest that the constitutional problems arising from this hardened role of EU soft law in national administrative courts could be cured by extending and improving the preliminary ruling procedure which would help national judges to decide cases.
II The Definition of Soft Law
1. Hard and Soft Law
When law enforcers, especially judges in EU Member States’ courts, come to decide cases, they are accustomed to consider various documents, involving (hard) law, soft law and sometimes other official policy documents that do not even reach the level of soft law. Obviously, these documents have different impact on the decision they make. Carefully trying to avoid the delicate issue of how to define (hard) law, this paper will focus on the role of soft law played in judicial procedures involving the review of administrative decisions at national level.
Soft law has its origins in international law, where it can have two meanings. On the one hand, in a formalistic way, soft law is not a source of international law, yet regulated individuals follow it as if it was law. On the other hand, it is acknowledged as a source of law, but without a normative content, meaning that neither rights nor obligations may be conferred by it.
In the creation of hard law, the forum (source), the wording (content) and the legal form play an important role. If one or two of these conditions is absent (suppose the forum does not have authority to enact that law, or no rights and obligations are created due to the ambiguous text of the law, or it was not enacted in an appropriate procedure, or was not published in the adequate form), its hard law character will be lost. Normativity is also a key concept for our study. Normativity is related to the content of the legal document. Legal norms are general rules created and enforced by the state to direct human behaviour. We will thus consider only normative, general rules of conduct as legal norms, disregarding otherwise binding court judgements.
Giving a clear definition of soft law is made difficult by the heterogeneous legal instruments that can be covered. Some authors describe soft law as an umbrella concept, and EU soft law as part of another umbrella, the acquis communautaire. We must also refer to the discourse in legal literature which claims that the term itself is self-contradictory, since law is either binding or not law at all, so non-binding law, which may be the primary definition of soft law, is a contradiction that should not exist.
Based on F. Terpan’s classification, it must be made clear that the term of soft law does not include hard law norms, the content of which is so ambiguous that they cannot confer rights or obligations on individuals (hard law with no or soft enforcement). Trepan uses the example of the Stability and Growth Pact, which contains hard obligations but its enforcement is dependent on the willingness of Member States to give effect to the obligations. Additionally, those cases when hard law cannot create rights or obligations must also be excluded, because it is not enforceable (hard law, no enforcement).Here Trepan refers to the Common Foreign and Security Policy, which is also meant to be legally binding, but without the judicial oversight of the European Court of Justice (ECJ).
For the purposes of this paper, we will focus on the distinguishing feature between hard and soft law from the perspective of national judicial decisions. Under non-binding hence soft law norms we will understand legal norms that cannot or usually cannot be enforced through judicial proceedings. This does not mean however, that they do not have a role in judicial proceedings. This interpretation is also supported by EU law as well, since Article 263 of the Treaty on the Functioning of the European Union (TFEU) explicitly excludes non-binding legal acts (recommendations and opinions) under Article 288 from the scope of the ECJ’s authority.
2. The Formal Classification of EU Soft Law Instruments
a) Official soft laws: recommendations and opinions
An obvious way to classify EU soft law instruments is to examine whether the given document was adopted in a form recognised by the TFEU itself. Accordingly, an act that was not adopted through an officially recognised procedure cannot become an official source of law.
Article 288 of the TFEU mentions recommendations and opinions as legal acts of the EU. These are adopted in a regulated procedure by certain institutions, based on the delegation of authority in the TFEU. Their soft law nature derives from the fact that, according to the TFEU, they shall have no binding force. In the following, we will focus on recommendations, as the application of this type of official soft law is prevalent both in the ECJ’s and national courts’ jurisprudence. In contrast, opinions are more like individualised documents relating to a specific legislative, accession or other decision and so their legal effect on third parties is not obvious.
b) Unofficial soft laws
The various types of annual reports, published legislative agendas, white books, green books and guidelines interpreting hard law provisions, notices, communications, etc. all constitute documents that are not official legal acts under the TFEU, yet they may have normative content. Since their creation is not regulated, their titles can be somewhat arbitrary, even though EU institutions publishing them try to give them a name matching their content. In the following, we will refer to these as communications.
Building on Senden’s approach, communications may be classified according to their legal effect into four groups. There are communications that are explicitly binding and confer rights and obligations [A]; others interpret hard law (interpretative communications) [B]; other documents restrict the law enforcement discretionary powers of EU or Member State authorities (decisional communication)[C]; and there are some which do not contain any general rule of conduct, thus cannot possess any legal force [D].
Communications of type [A] have some force of law due the explicit authorisation of a hard law provision that they should create rights and obligations. We will not analyse this type any further, since their enforceability at courts is the same as for any hard law provision.
As to types B and C, often there is no clear difference between interpretative and decisional communications. For example, the Communication of the Commission on waste is clearly an interpretative communication, because it interprets the provisions of a directive with references to judicial case-law. Here, the point of reference at courts will therefore not be the communication itself, but the directive and the related case-law. In other cases, however, the distinction is not so clear, so we will not differentiate between these two types of communications, and the judicial enforceability of such norms will also be analysed together.
III Exploring the Legal Nature of Soft Law
1. Legal Principles as an Analogy
The specific nature of soft law can be better understood by an imperfect analogy with legal principles. Scholars highlight the logical difference between rules and legal principles. The difference between the two is that rules function on an “all or nothing” basis, so in a single case it is either applicable or not – and if applicable, then it shall be applied. In contrast, legal principles are applicable to all cases or at least to a wide set of cases, yet they are not fully applicable and not without any restraints. This means that more than one legal principle may be of relevance to a given case, and the judge must assign a certain weight to each of these to deliver a balanced judgement. Legal principles can also be described as ambiguous abstract rules that should be realised through law enforcement.
Legal principles, just like soft law instruments have an important role in making abstract legal norms workable. Primary EU law usually consists of highly abstract norms. Take Article 102 TFEU as an example. None of the terms, such as undertaking, dominant position, abuse, or effect on trade, are defined by the text itself. There are no implementing regulations either. Even if the case law of the European Courts fills these norms with more precise content, there is still a wide margin of appreciation, where general principles of law and soft law instruments issued by the Commission will have a role to play.
Consequently, both legal principles and EU soft laws help to fill abstract norms with content. Soft law instruments often refer to general economic and legal principles, which may give them a legal flavour. The difference between legal principles and soft law is that the former fit better into the traditional hierarchy of legal norms, while soft law instruments seem to exist in an overlapping but different dimension of the legal universe. The Court held that interpretative communications cannot modify the rules contained in hard law. So, soft law documents either do not fit in traditional legal hierarchy at all, or, if they do, they occupy its very lowest position.
2. Ius Dispositivum as an Imperfect Analogy
In (hard) law, there are generally binding rules which always oblige persons (ius cogens) and there are legal rules which bind the parties unless they agree otherwise (ius dispositivum). We emphasise this distinction between hard laws, because the reach and nature of binding effects and the option to deviate from what is written into a soft law instrument will be of crucial importance. We shall see that the EU Commission may be bound by its own soft laws, but the option to deviate will be available, subject to giving reasons and avoiding the infringement of legal principles protecting third party interests. This makes soft law look like dispositive legal norms.
The main difference is that the legitimacy of deviation from dispositive law depends upon consensus between equal parties, whereas the deviation from soft law needs to be reasoned and to respect general legal principles in order to preserve the reliability of the legal system. We will mostly focus on the problem of deviation by the regulatory or competition authority, yet we should note that even third parties may choose not to follow soft law. However, if challenged later, they must be able to explain with persuasive authority why they believed that guidelines issued by the EU Commission to elaborate on competition concepts were flawed. Our experience is that the simplistic argument that ’soft law is not law so why should I ever bother reading it and acting in accordance’ would not work before a national court.
3. Soft Law as a Useful By-product of Law Enforcement: the Self-control Effect
To understand the nature and the role of soft law, it is important to realise that the emergence of soft law is almost a necessity, which cannot be avoided or prohibited, just as it cannot be legislated that the sun rises in the west instead of the east. Every efficient organisation that enforces broadly worded hard law tries to generalise the lessons of individual decisions. This is necessary in order to create a coherent and unified interpretation of laws. This process is justified by the pursuit of efficiency, so that not each and every decision has to be thought over and over again. Soft law can save resources and make law enforcement quicker and more consistent. Moreover, standardisation and generalisation are inevitable in bureaucratic organisations with a significant turnover of public servants in order to maintain a policy line and to eliminate arbitrary interpretation of laws. If a public authority makes individual adjudicatory decisions then sooner or later it will also create – even without any kind of authorisation or delegation – working documents that collect and normatively contain the generalised knowledge of the institution.
This is also true for EU institutions that are not the sole decision-maker in a certain field, but have oversight powers over the consistency or efficiency of law enforcement. Nevertheless, such normative documents will have a higher normative value if they are issued by the competent and experienced decision-making authority. The binding nature of their content “hardens” when the institution publishing them, mostly the Commission, also takes part in the enforcement, and in this way it is in a better position to specify hard law provisions. As an example, the fields of environmental law and competition law are worth mentioning.
The publication of soft law can also be seen as a self-restriction by the authority, since it makes the use of its discretionary power more legally bound. The more legally bound the discretionary powers of an authority are, the more they will be in line with the rule of law principle. It explicitly strengthens legal certainty if the decision-maker is bound by its own soft law.
Soft law documents interpreting the law are thus natural and useful by-products of law enforcement as exercised by executive institutions and courts. The judicial system is also characterised by case-specific decision-making. The publication of decisions with normative content (e.g. uniformity decisions), or the publication of jurisprudence analysis is fairly common in most national judicial systems. Some legal systems acknowledge binding decisions with normative content, which could be considered judicial law-making. The scope and formality of the binding power of such judge-made normative legal acts are usually an adequate indicator of judicial independence. Studies have confirmed that it does not really matter whether the highest court publishes such a normative document in the form of opinions or as binding precedent, since lower courts will be obliged to obey them in both instances.
Beyond this, there is one significant difference between authoritative court decisions and soft law instruments. A well-established and published highest court jurisprudence interpreting an abstract hard law rule will eventually become law and will be binding on everyone, since it will be enforceable at courts. For this reason, case-law and normative documents adopted by national courts shall here be excluded from scope of genuine soft law. Put differently, a published, standardised interpretation of a legal issue by a high court has a higher value than similar documents issued by administrative authorities. On the other hand, this kind of case law, just like some soft law rules, will never be applied on their own, just like the normative paragraph of an act by the Parliament. Court case law will connect to one or more paragraphs of an act. Soft law, to have any binding effect, will need the additional involvement of general legal principles, such as the protection of legitimate expectations.
4. Soft Law as a Tool of Soft and Transparent Control Mechanism
Normally, the addressees of intra-institutional soft law documents are employees of the given organisation or the supervised organisation, who are expected to follow them depending on the relative professional independence of their decision-making authority. Even if the legal system would disqualify these documents, they would still emerge, surviving the above mentioned functions, but hidden in desk drawers, accessible only to insiders. Were the legislator to prohibit all these documents (circulars, normative instructions), the organisation exercising control powers would lose one of its important tools. A prohibition would thus result in employees using non-public guidance to decide cases, which is a much worse situation than allowing the use of soft law documents, even if this might cause constitutional and theoretical problems.
Another driving force behind the creation of soft law is that, under the shared administration model of the EU, national authorities are responsible for enforcement, while ensuring consistency of law enforcement is the responsibility of the EU institutions (ultimately the ECJ), but also involving the Commission. The Commission, which does not possess hierarchical oversight powers, can give orientation to national authorities through soft law documents.
For example, the European Competition Network operates more like a family then a clearly regulated hierarchy of competition agencies. Soft law, coupled with the Commission’s exceptional role to step in and squeeze out an allegedly incompetent national authority from deciding a case of European relevance, softly regulates cooperation between otherwise independent authorities.
A number of soft law documents published by the Commission contain provisions on what is expected from national regulatory authorities in respect of the soft law document. For example, the Commission Recommendation on the Regulatory Treatment of Fixed and Mobile Termination Rates in the EU begins with a rather imperative wording: “NRAs should proceed in the way set out below”, which also shows that soft law documents can serve as a tool for the orientation of national regulatory authorities.
5. Soft Law as a Laboratory of Hard Law – The Implementation of Soft Law
Is there a need to implement EU soft law instruments into domestic hard law? Is this another process of witnessing the ‘hardening’ of soft law? For example, the Commission’s recommendation on the relevant product and service markets within the electronic communications sector was published in Hungary in the form of a decree.
We do not intend to go into any further details on the incorporation of soft law into national law, but in general it is worth noting that there exists no direct obligation to implement it: since soft law is formally not law, the questions of direct and indirect effect cannot arise either, although there are some authors who claim that soft law can have an indirect effect. As opposed to these views, we would rather claim that soft law can be applied directly, just like a regulation, even though formally it does not have direct effect. As such, the enforcement of recommendations and regulations is not markedly different. The single major distinction is their wording, since a regulation can create clear, precise and unconditional rights and obligations, while, at least in theory, a recommendation could not.
As far as EU competition soft law is concerned, their application is not subject to national implementation either. They can be observed by national law enforcers directly. Nevertheless, we can observe that EU soft law did influence the evolution of Hungarian substantive competition rules. For example, the de minimis rule for certain anti-competitive agreements was codified based upon the then existing Commission notice on de minimis agreements. The statutory rules on leniency policy were to a great extent also modelled on the relevant European soft law document. This was partly due to the law harmonisation obligation existing under the Europe Agreements. This soft law harmonisation, however, had nothing to do with the direct applicability of these instruments.
Regarding state aid, the Hungarian hard rules laying down the substantive and procedural rules on state aid under Article 107 TFEU include several provisions basically implementing or at least mirroring EU soft law documents. In addition to procedural rules, it includes some substantive rules that rely on EU soft laws. For example, Section 5 declares that the intensity of aid cannot be higher than the intensity laid down in Community rules. The word “rule” is a broad one, involving not only hard law. Furthermore, Section 6 lays down detailed rules based upon a Commission communication on the circumstances in which an undertaking can be regarded to be in such difficulty as to receive rescue and restructuring aid.
To conclude, EU communications can have hard legal consequences indirectly, through national legislation as well.
6. Is Soft Law Legitimised by Courts?
Key to the existence of soft law is the stance that courts take in relation to these legal instruments. If courts were to disregard them as no-law, simple policy documents issued by over-activist authorities, they would either disappear or become limited in their scope, regulating only inter-institutional relations within an authority.
The jurisprudence of the Hungarian Constitutional Court is an example of a cautious approach to soft law. The Court stated that documents that are not regulated as a source of law in the Act on legislation are in violation of the Constitution; however, it did not annul them, but made it clear that no legal effects can derive from them. This interpretation came from an overly protective reading of the separation of legislative and executive powers; nonetheless, its effect was instead to weaken the separation of these state functions. If these documents are not public and accessible to the affected individuals, the authority not only fails to fulfil its duty to provide information, but also creates a situation endangering the rule of law and legal certainty. Legal certainty is increased whenever the decision-making authority provides information on its decision making practice and on the conduct it expects undertakings and other persons to follow.
The ECJ elaborated its views on the legal effects of soft law in the Grimaldi case. The ECJ acknowledged that recommendations are not intended to produce binding effects, even as regards the person to whom they are addressed, and consequently they cannot create rights upon which individuals may rely before national courts. However, national courts are bound to take recommendations into consideration in order to decide disputes submitted to them, in particular where they cast light on the interpretation of national measures adopted to implement them or where they are designed to supplement binding Community provisions. 
This still valid definition of ‘no binding effect’ is regarded as the essence of soft law, under which ‘non-binding’ means, for a national judge, that the given legal rules are not enforceable in court. Nevertheless, the phrase ‘bound to take into consideration’ is in need of some explanation. It means that, in the course of enforcing the law, a recommendation or communication must be taken into consideration, regardless of whether the underlying hard law rule mandates this or not. This legal effect of soft laws, probably not limited to formal soft laws such as recommendations, could be described as a vertical indirect effect, just as with directly effective rules in an unimplemented directive. Individuals may rely on those soft law provisions whereby the EU Commission binds its hands; for example, when it comes to give significant reductions of fines for leniency applicants or for companies that prefer to settle their case. Just as for directives, third parties can rely on these rules against the state authorities, but not in a private (horizontal) dispute.
For Craig and de Burca, the indirect effect of soft laws means that they can help interpret hard law. This is certainly true, but communications sometimes go beyond this. By way of example, the Commission’s leniency and settlement notices introduced highly relevant figures for undertakings seeking a reduction of their fines. They seem to fulfil more of an executive role, laying down rules which, by their very nature, should have been codified in a Commission implementing regulation.
The ECJ stated that, as far as informal soft law acts (communications) are concerned, an action for annulment is not available in cases where the adopted legal measure is not intended to have legal effects on third parties. The same applies to recommendations under Article 263 of the TFEU. However, when a communication is binding or aims to create an obligation that does not exist in EU hard law then, based on a functional approach, it should be the subject of an annulment procedure. This means that if the wording of the document implies binding force then, regardless of its form, it can be subject to an action for annulment available for hard law acts.
More than one conclusion can be drawn from this practice. First, no obligations on third parties should be ever created by soft law. Second, even if a national judge would believe that a soft law rule has such binding effect, a national judge must take that into consideration as long as the European Court of Justice has not annulled it. This derives from that logical necessity that if a soft law document can be the subject of an action for annulment, this means the implied acknowledgement of its binding effect because otherwise an annulment would not be necessary. The same conclusion can be drawn from Advocate General Mengzi’s Opinion given in a recent preliminary ruling procedure in the field of electronic telecommunications.
Contrary to this, the Hungarian Constitutional Court does not annul soft law measures that are not formally regarded as law yet seem to have external effects. It is simply declared that these measures are unable to create any legal effects, in order to avoid even temporarily acknowledging their binding effect. As we have mentioned above, however, this approach has a number of downsides compared to the ECJ’s practice, as it can lead to hidden normative documents and less transparent law enforcement.
Despite the advantages of the ECJ’s approach, that it allows such documents to be created and published, it has disadvantages as well. The question of whether soft law is enforceable as law at national courts depends on the effectiveness of the procedure leading to their potential annulment. It is of crucial importance how effectively national judges can refer soft law documents for preliminary rulings on challenging their validity or to clarify their interpretation. This is also true in cases when the binding soft law norm is not in conflict with any hard law rules.
It is worth noting that, in Grimaldi, the ECJ held that a hard law rule may be substituted with a soft law one. It can be thus concluded that a soft law creating, or to be more precise, clarifying existing rights in relation to the public authority issuing the soft law document is not in violation of hard law; meanwhile, the one creating obligations on individuals is unlawful.
7. To Deviate or Not To Deviate: That is the Question
Laws, including cogently binding general provisions, should be applied without exceptions. Law enforcers cannot deviate; whatever creative interpretation they adopt, it must stay within the boundaries of the (hard) law rule. This principle does not hold for soft laws.
According to the settled case law of the ECJ, the Commission can be bound by its published communication or other types of soft law instruments. This self-binding effect is in line with general legal principles of non-discrimination, legal certainty, and protection of legitimate expectations. Nevertheless, deviation from published soft law is only allowed if it is well reasoned and does not contradict any of the above mentioned legal principles. This line of argument of the ECJ makes soft law documents binding through the general legal principles.
The binding nature of soft law on national authorities is a more delicate question. Given the carefully balanced share of sovereign powers between national and EU institutions, even if legal certainty and the unity of the single market would demand national authorities to follow the path of the EU Commission, this is far from being obvious. Without the special legal powers of the Commission, this would not work in practice.
Recommendations issued within the framework of telecommunication regulation aim to bind not only the issuing institution but also national authorities applying them. In the same way as the Commission can deviate from its own soft laws, national authorities can also deviate, if they have good reasons to do so. This rule is also valid when soft law does not expressly allow for any deviation. This also stands for national courts. In a recent Dutch telecommunication case referred to preliminary ruling, the Advocate General’s Opinion stressed that judicial deviation from a recommendation must be exceptional and the judge must be excessively cautious and can only deviate from the recommendation based on serious reasons.
EU communications interpreting Articles 101 and 102 TFEU often include a provision, according to which ‘Although not binding on them, this Notice is also intended to give guidance to the courts and competition authorities of the Member States in their application of Article 101 of the Treaty.’ These soft law instruments are well structured, detailed, lengthy documents, often relying on EU Courts’ case law and finalised following public consultation. In practice, it is hard not to take them seriously. This ‘not binding, but guiding’ wording of communications can be understood to reflect a soft binding effect. The Commission finds it important to give express mention to national law enforcers as targets of the communication and makes it clear that the Commission is in a position to guide them through the maze of EU competition rules. Certainly, the Commission cannot avoid mentioning that the communication is not binding on them, which is nothing new as long as we understand ‘binding’ in its traditional hard law sense: an absolute prohibition on deviating from the rule. According to our understanding, soft law instruments have a soft binding effect: law enforcers should do their best to follow them, but can deviate as long as it is explained in the decision and does not infringe general principles of EU law.
Apparently, the ECJ gives a wider margin of appreciation to national courts. The Court held in Expedia, responding to the French Court of Cassation that national authorities are not bound to apply EU soft law instruments, and they have complete discretion to take the thresholds mentioned in the de minimis notice into consideration. The Court rightly referred to the wording of the notice and that it was published not in the L, but the C series of the Official Journal.
The obvious difference between law and soft law becomes evident when considering behaviour contradicting the rules established therein. Any conduct against hard law will be regarded as unlawful behaviour with the legal consequences as prescribed by law. On the other hand, soft law cannot formally be infringed by undertakings. It will be always the underlying hard law provision as interpreted by a communication that can be the subject of any infringement. Only the issuing authority can be held liable for not respecting its own rules. Nevertheless, if the national authority, due to the soft binding nature of soft law, relies on the soft law rule in its decision, then the undertaking will also be bound by it in practice. Third parties may rely on them before courts reviewing the legality of the decision. Even here, EU law makes a reference to the protection of general principles of law, such as the protection of legitimate expectations.  The sole legal basis of invoking the infringement of soft law by an authority will not suffice to win a case. The prohibition of derogation does not come as a result of protecting carefully considered legal principles. This is simply the nature of hard law.
The article aimed at demonstrating the special legal character of EU soft law documents and their effects at national level.
We observed that the legal effects of soft law documents are somewhat analogous to general principles of law, as they both help to fill abstract norms with content but unlike hard law norms, neither of them applies on an “all or nothing” basis. Similarly, soft law can be compared to legal rules which bind the parties unless they agree otherwise (ius dispositivum). The EU Commission – and sometimes even national regulatory authorities and courts – may be bound by EU soft laws, but the option to deviate will be available, subject to giving reasons and avoiding the infringement of legal principles protecting third party interests.
We argued that the adoption and especially the publication of soft law can be seen as a self-restriction of the authority, since it makes the use of discretionary power more legally bound. The more legally bound the discretionary powers of an authority are, the more they will be in line with the rule of law principle. However, at EU level another driving force behind the creation of soft law is that, under the shared administration model of the EU, national authorities are responsible for enforcement and so the Commission, which does not possess hierarchical oversight powers, can give orientation to national authorities through soft law documents.
We argued that in the course of the national law enforcement, soft law documents are not intended to produce binding effects; consequently, they cannot create rights upon which individuals may rely before national courts. However, national courts are bound to take them into consideration in order to decide disputes. Even if a national judge would believe that a soft law rule has real legal effects, and it should be annulled, the a national judge must take it into consideration as long as the European Court of Justice has not annulled it. This derives from that logical necessity that if a soft law document can be the subject of an action for annulment, when it intends to have legal effects, then this means the implied acknowledgement of its binding effect, otherwise an annulment would not be necessary.
Thus the question whether soft law is enforceable as law at national courts depends on the effectiveness of the procedure leading to their potential annulment. How effectively national judges can refer to soft law documents for preliminary rulings to challenge their validity or clarify their interpretation is of crucial importance.
* Associate professor, Eötvös Loránd University, Administrative Law Department, administrative judge at Curia (the highest court of Hungary).
** Associate professor at Pázmány Péter Catholic University, Department of Environmental Law and Competition Law, attorney at Réczicza Dentons Europe LLP.
*** Assistant lecturer, Eötvös Loránd University, Administrative Law Department.
 Following the recommendations of Oana Stefan, Soft law in Court – Competition Law, State Aid and the Court of Justice of the European Union (Kluwer Law International BV, 2013, Alphen aan den Rijn) for further research topics. 242.
 See e.g. Stefan (n 1) 162-165. and 275-324.
 We should also note that the difference between these categories of documents is far from obvious at the margins.
 Starting our inquiry with recalling international legal principles is reasonable, since EU law, although it has evolved into a separate legal order, has its origins in international law.
 László Blutmann, ‘In the trap of a legal metaphor: International Soft Law’ (2010) 59 International and Comparative Law Quarterly 605-624, 606. or Damrosch et al., International Law, Cases and Materials (West Academic Publishing 2001, New York) 34.
 This approach follows G.M. Borchardt & K.C. Wellens, ‘Soft Law in European Community Law’ (1989) 14 European Law Review 267, 270, 298-301 See also: Zlatina Georgieva, ‘Soft Law in Competition Law and its judicial Reception in Member States: A Theoretical Perspective’ (2014) 16 German Law Journal 234. Similarly K.W. Abbott and Sindal who define ’hard law’ as the form of legalisation characterised by high levels of obligation, precision, and delegation, whereas they consider that soft law occurs whenever one or more of the three dimensions are weakened. K.V. Abbot et al, ‘The Concept of Legalization’ (2000) 54 International Organization 3, 17–35.; K. Abbott & D Sindal, ‘Hard and Soft Law in International Governance’ (2000) 54 International Organization 421-422. For the critic of this approach see: U. Mörth (ed), Soft Law in Governance and Regulation – An Interdisciplinary Analysis (Edward Elgar Publishing, 2004, Cheltenham) see also in O. Stefan (n 1) 9-10
 Terpan includes non-enforceable lex imperfecta rules in the realms of soft law, see: F. Terpan, ‘Soft Law in the European Union The Changing Nature of EU law’ (2014) 19 European Law Journal 1, 58-59. It must be noted that lex imperfecta rules exist in every national legal system; moreover, they must necessarily exist in any legal system (e.g. organisational rules). Here we refer to those binding norms which are not enforceable at courts, because they cannot even reach national courts due to the lack of an mechanism for implementation and enforcement. Another difference between lex imperfecta and soft law is that lex imperfecta rules are adopted in a legal form that is binding per se, even though their enforcement is mainly political in nature.
 It is well known that natural lawyers and legal positivists hold opposing views on the nature of law.
 Legalnormscan beclassifiedaccording to theirlegalforce or binding effect.Legalnorms can alsodividedintoimperativenorms,containingcompulsoryrules, i.e. antitrust prohibitions and telecommunication regulations,anddispositionalnorms,allowingtheparticipants to definetheirrightsanddutieswithinthe framework established by law (here the norm plays a subsidiary role in the law).
 There are some intermediate documents between single and normative acts, for example a single decision of an authority addressed to a distinct group of people containing theoretical, generalised conclusions.
 See for example: F. Terpan (n 7), O.Stefan (n 1) 11., E. Korkea-aho, ‘EU soft law in domestic legal systems: flexibility and diversity guaranteed?’ (2009) 16 Maastrict Journal of European and Comparative Law 271, 274.
 L. Senden, ‘Soft law, self regulation, and Co-Regulation in European Law: Where Do They Meet’ (2005) 9.1. Electronic Journal of Comparative Law 23. <http://www.ejcl.org/> accessed: 6 September 2015.; L. Senden, ‘ Soft law and its implication for institutional balance in the EC’ (2005) 1 Utrecht Law Review 2, 79, 81 similarly M. Medelson, ‘Formation of Customary International Law’ (1998) 272 Hague Academy of International Law, Collected Courses 155-410, 360. in L. Blutmann (n 5) 610.
 O. Stefan (n 1) 118, with respect to the T-115/94 Opel Austria v. Council case, EU:T:1997:3. However, a number of authors warn that international and EU soft law cannot be handled in the same way under the umbrella concept, since in EU law these instruments are somewhere between a general statement of policy and lawmaking, which as a tertiary source is at the bottom of the hierarchy of legal sources. See also: O. Stefan (n 1) 10-11, F. Terpan (n 7)55-56
 Soft law is a contradictio in terminis see: L. Senden, Soft law in European Community Law (Hart Publishing 2004, Oxford) 109, O. Stefan (n 1) 117., H. Hillgenberg, ‘A Fresh look at soft law’ (1999)10 European Journal of International Law 500. or L. Blutmann (n 5) 609-610.
 F. Terpan (n 7) 58-59.
 For example, in competition law procedures, soft law documents read in conjunction with legal principles may even create rights for third parties that courts should protect, see section 3.2.2.
 This approach conforms with the most often cited definition of soft law by Snyder, ‘The Effectiveness of European Community Law: Institutions, Processes, Tools and Techniques’ (1993) 56 Modern Law Review 64. See similarly the definition by Senden in L. Senden, ‘Soft Post-Legislative Rulemaking: A Time for a More Stringent Control (2013) 19 European Law Journal 1, 112. In adition, one of the elements of Chinkin’s soft law definition is that soft laws are based solely upon voluntary adherence, or rely upon non-juridical means of enforcement. See: C. Chinkin, Normative Development in the International Legal System in Commitment and Compliance: The Role of Non-Binding Norms in the International Legal System (D. Shleton ed., Oxford University Press, 2000) 30. in O. Stefan (n 1) 9; 11, So, all in all, they have no legally binding force but may have practical as well as legal effects or, as Senden says, may produce practical effects. Hence, sometimes in practice they have legal effect (e.g. at courts), but not always (not bindingly).
 See for example: Commission Opinion COM/2011/0667 final on the application for accession to the European Union by the Republic of Croatia, Commission Opinion COM/2011/0596 final on the requests for the amendment of the Statute of the Court of Justice of the European Union, presented by the Court, Commission Opinion COM/2008/0730 final on the request from the United Kingdom to accept Regulation (EC) no 593/2008 of the European Parliament and the Council of 17 June 2008 on the law applicable to contractual obligations (ROME I)
 L. Senden (n 17) 57–75 and 60-61 L. Senden (n 14) 219-220 For communications with binding force and without, see: Várnay Ernő- Tóth Tihamér, ‘Közlemények az Uniós Jogban’ (2009) 50 (4) Állam és Jogtudomány 417-472, 417-418
 It is, however, worth noting that such binding soft law is never put into recommendations or opinions, since their regulated legal status expressly precludes any binding effect.
 COM(2007) 59 final Communication from the Commission to the Council and the European Parliament on the Interpretative Communication on waste and by-products
 R. M. Dworkin, ‘The Model of Rules’ (1967) 35 University of Chicago Law Review 14, 23-29, 30-32. See also Jakab András, ‘A magyar jogrendszer alapelemei’ 2.4. footnote 45.<http://jesz.ajk.elte.hu/jakab13.html> accessed: 6 September 2015. See also: Jan-Reinard Sieckmann. Regelmodelle und Prinzipienmodelle des Rechtssystems (Mohr Siebeck, Tübingen, 1999)
 See for example C-27/76 United Brands Company and United Brands Continentaal BV v Commission of the European Communities, EU:C:1978:22
 As for regulations, see T-9/92 Peugeot v. Commission case, EU:T:1993:38 para 44, confirmed by C-322/93 Peugeot v Commission, EU:C:1994:257
 Naturally, it can be enacted as a rule, but it cannot meet the criterion of a norm or a single decision, because such a subject cannot be the subject of a regulation; the addressee is not a legal entity, etc., so it cannot be law.
 The soft law limits institutional discretion, encouraging the administrators to take consistent decisions. O. Stefan (n 1) 16
 In this case effectiveness as a goal also results in increased legal certainty, which is a separate goal in democratic states, although some aspects of the two terms are in a dichotomous relation, similarly to the principles of fair procedure and reasonable time.
 In environmental law it is only an interpretative tool to case-law, while in competition law it prescribes rules. See: COM(2007) 59 final Communication from the Commission to the Council and the European Parliament on the Interpretative Communication on waste and by-products. Compare with footnote 21. However, the soft law documents mentioned in footnote 18 rarely appear in judicial decisions.
 For example, in Hungary, as a civil law legal regime, the Supreme Court (Curia) may publish uniformity decisions in cases of theoretical importance in order to ensure the uniform application of law within the Hungarian judiciary. Such decisions are binding on all Hungarian lower instance courts (see: http://www.kuria-birosag.hu/en/uniformity-decisions-jurisprudence-analysis). Ont he other hand, in common law legal regimes, precedents may be considered judge-made law with normative content (see e.g. Cross, R. – Harris, J., Precedent in English Law (4th edition, Oxford University Press 1991, Oxford).
 However, these indicators not necessarily show the independence from the executive, but rather indicators of professional qualities, personnel decisions and material resources. The reason for this is that one of the main components of judicial independence is professional autonomy, in order for the judge to be able to reach autonomous professional decisions. The lack or uneven character of this (lack of proper selection procedures and organisational resources) is often compensated by judicial law-making competences, in order to create uniformity in the judicial practice. It also shows the level of judicial independence whether these uniformity decisions are adopted by the judges themselves, or by a special committee or group of people outside the judiciary.
 See for example: Ződi Zsolt, ‘A korábbi esetekre történő hivatkozások a magyar bíróságok ítéleteiben’ (2014) MTA Law Working Papers. <http://jog.tk.mta.hu/uploads/files/mtalwp/2014_01_Zodi_Zsolt.pdf> accessed: 24 October 2015.
 This exclusion is clear based on the jurisprudential debate exactly on the hardening of soft law by courts. See: O. Stefan (n 1) 152-155, E. Korkea-aho (n 11) 279.
 Here the question is whether executive rule-making with or without a delegation of authority should be allowed.
 Often this is the case in competition law and sectoral market regulation, where a large number of soft law documents exist. For the same reason, soft law is prevalent in those fields of law where EU has responsibilities but lacks real competences and so it tries to counterbalance it by the adoption of soft law documents, such as employment and social policy, especially if the hard law rules international law features, such as the Lisbon Strategy, or if the division of competences between EU and member states is not entirely clear. See O. Stefan (n 1) 12.,15.
 Commission Recommendation 2009/396/EC of 7 May 2009 on the Regulatory Treatment of Fixed and Mobile Termination Rates in the EU  OJ L 124 para 1
 In the common regulatory framework for electronic telecommunications, it must be mentioned that not only soft law documents may orient NRAs, but similar questions arise in relation to letters addressed by the Commission to NRAs in individual cases. (See: T-109/06 – Vodafone España and Vodafone Group v Commission, EU:T:2007:384) Although the main issue regarding such letters is also the legal effect they produce, we will still exclude these from the analysis, as these are not generally applicable documents in the same way as soft law norms. Instead, they raise another question, namely the normative and binding force of individual decisions (or even comments).
 There are cases when the case law of the EU Courts gives more room for national authorities. In case C-226/11 Expedia Inc v Autorité de law concurrence and Others, the Court explained that the Commission’s de minimis notice does not bind competition authorities.
 Commission Recommendation 2007/879/EC of 17 December 2007 on relevant product and service markets within the electronic communications sector susceptible to ex ante regulation in accordance with Directive 2002/21/EC of the European Parliament and of the Council on a common regulatory framework for electronic communications networks and services  OJ L 344
 16/2004 (IV. 24.) IHM rendelet a piacmeghatározás, a piacelemzés és a jelentős piaci erővel rendelkező szolgáltatók azonosítása, valamint a rájuk vonatkozó kötelezettségek előírása során alkalmazandó alapelvekről (Decree (16/2004 (IV. 24.) IHM of the Hungarian Minister of Information and Telecommunication on the applicable principles regarding market determination, market evalutaion, the identification of service providers with significant market powers and regarding setting obligations on them)
 It must be noted that the Commission’s guideline, which interpreted the content of the recommendation without the authorisation of the directive was implemented into Hungarian law, into a national communication (8001/2004. (IHK 8.) IHM tájékoztató a piacmeghatározás, a piacelemzés és a jelentős piaci erővel rendelkező szolgáltatók azonosítása, valamint a rájuk vonatkozó kötelezettségek előírása során a hatóság által alkalmazandó alapelvekről (vizsgálati szempontokról) (8001/2004. (IHK. 8.) IHM guideline on principles that the authority must use when determining the relevant market and SMP service providers and their obligations). This legal document is not a formal legal source, because it cannot have outside effect based on the act on the hierarchy of legal sources, however, it bound the employees within the organisation. Even though, the guideline was not public, it was still relied on in judicial cases. This is the same guideline that was the subject of the ECJ’s judgement in C-410/09 Polska Telefonia Cyfrowa (PTC) sp. z o.o. v Prezes Urzędu Komunikacji Elektronicznej, EU:C:2011:294. Here the ECJ stated that “the 2002 Guidelines do not lay down any obligation capable of being imposed, directly or indirectly, on individuals. Accordingly, the fact that those guidelines have not been published in Polish in the Official Journal of the European Union does not prevent the NRA of the Republic of Poland from referring to them in a decision addressed to an individual” (para 34). This case shows that publication has no significance in respect of referencing a guideline. However, since it is no longer in force, we will not deal with it any further, even though it raises interesting question regarding constitutional principles. For more on the controversies of the PTC case, see: O. Stefan, ‘European Union Soft Law: New Developments Concerning the Divide between Legally Binding Force and Legal Effects’ (2012) 75 (5) The Modern Law Review 879, O. Stefan (n 1)193-198
 Craig–de Burca, EU Law Text, Cases and Materials (Second Edition, Oxford University Press 2008, Oxford) 277-278
 It must be noted that there are certain regulations that are directly applicable by definition but do not have direct effect, because no rights and obligations can be created based on their text. L. Blutmann, Az Európai Unió joga a gyakorlatban (HVG-ORAC 2013, Budapest) 350.
 In 1996, the then applicable de minimis notice set a single 10% limit, which was codified by the Hungarian legislator. Later, when the European percentage was raised to 15% for vertical restrictions, the Hungarian competition act did not follow this move. It is also true that the Hungarian statutory exception is less sophisticated; it prohibits only horizontal price fixing and market sharing for small undertakings, whereas the European soft law instrument involves a more extensive list of hard core restrictions
 T. Tóth, ‘The reception and application of EU competition rules: an organic evolution’ in Várnay and Varjú (ed), The Law of the European Union in Hungary (HVG Orac 2014, Budapest) 244
 37/2011. (III. 22.) Korm. rendelet az európai uniós versenyjogi értelemben vett állami támogatásokkal kapcsolatos eljárásról és a regionális támogatási térképről (Government decree 37/2011. (III. 22.) Korm. on the procedural rules and regional map of state aid) Examples include definitions and the pre-notification procedure.
 Communication from the Commission — Guidelines on State aid for rescuing and restructuring non-financial undertakings in difficulty, OJ C 249, 31.7.2014, 1–28 p.
 See: Decision 39/2010. (IV. 15) AB of the Constitutional Court of the Republic of Hungary, ABH 2010, 1048 on the unconstitutionality of the 8001/2004. IHM guideline on principles that the authority must use when determining the relevant market and SMP service providers and their obligations; Decision 35/2009. (III. 27.) AB of the Constitutional Court of the Republic of Hungary, ABH 2009, 1097 on the normative nature of the circular published by National Police Force on how to apply wheel clamps; Decision 23/2007. (IV. 19) AB of the Constitutional Court of the Republic of Hungary, ABH 2007, 311 on the guidelines of the National Tax Authority on certain tax relieves; Decision 121/2009. (XII. 17.) AB of the Constitutional Court of the Republic of Hungary, ABH 2009, 1013 or 41/1993. (VI. 30.) AB of the Constitutional Court of the Republic of Hungary, ABH 1993, 292 or 52/1993. (X.7.) AB of the Constitutional Court of the Republic of Hungary, ABH 1993, 407 generally on non-binding norms.
 Which is understandable in the light of historical experiences.
 Here we will not deal with the instance when only referring to it is prohibited, or it can be referred to but not made accessible.
 In EU law context this appears in the principle of legitimate expectations (see: C-98/78 Racke v. Hauptzollamt Mainz, EU:C:1979:14, 205-215/82 Deutsche Milchkontor GmbH and others v. Federal Republic of Germany, EU:C:1983:233 or C-24/95 Land Rheinland-Pfalz v. Alcan Deutschland GmbH (Alcan II), EU:C:1997:163. In Hungarian context this principle is detailed in the hard law provisions of the general administrative procedure act.
 C-322/88 Grimaldi case, EU:C:1989:646. Although the judgement is about recommendations, the legal literature interprets it generally for the legal effects of soft law. See: Craig-de Burca (n 38) 210. Rideau J, Droit isttitutionnel de l’union européene et des Communatés européennes (4.e édition L.G.D.J. 2002, Paris) 162-166. It must be noted that related to comfort letters the ECJ had reached a decision even before the Grimaldi case, stating that national courts must take them into account. See: joined cases of 253/78 and 1/79-3/79 Procureur de la République and others v Bruno Giry and Guerlain SA and others, EU:C:1980:188 and the judgement in C-99/79 SA Lancôme and Cosparfrance Nederland BV v. Etos BV, EU:C:1980:193 See also: O.Stefan (n 1) 162.
 Grimaldi, EU:C:1989:646, para 16
 Grimaldi, EU:C:1989:646, para 18.
 Also cannot create obligations, as it will be seen later.
 These delegations can be phrased in various ways. The study does not differentiate between ‘take into account’ or ‘take the utmost account of’, although it may show some distinctions in legal effects based on hard law. However, this could be the subject of a separate analysis.
 See: Craig-de Burca (n 38) 210. and Rideau J. (n 47) 162–166.
 Craig-de Burca (n 38) 277–278.
 C-301/03 Italian Republic v Commission of the European Communities, EU:C:2005:727, para 19 and 24
 Case 22/70 Commission v. Council, EU:C:1971:32 para 42 (ERTA case), C-366/88. French Republic v Commission of the European Communities, EU:C:1990:348, para 8, C-303/90 French Republic v Commission of the European Communities, EU:C:1991:424 para 24; C-325/91 French Republic v Commission of the European Communities, EU:C:1993:245 para 14. This latter judgement shows that it also cannot create obligations, substituting binding rights, see para 22, 23 and 31. The judgement of C-57/95 French Republic v Commission of the European Communities, EU:C:1997:164 para 23. is specifically about interpretative communications; Case C-27/04 Commission v. Council, EU:C:2004:436 para 44., also the Joined Cases C-189/02 P, C-202/02 P, C‑205/02 P to C‑208/02 P and C‑213/02 P,Dansk Rørindustri A/S and others v. Commission of the European Communities, EU:C:2005:408.
 French Republic v Commission, EU:C:1991:424, para 26.
 See: Koninklijke KPN and Others, C-28/15, EU:C:2016:310, para 39, 55-57, 65.
 See: n 28.
 This is called regulation by publication in the jurisprudence. See: F. Snyder ‘Soft Law and Institutional Practice in the European Community’ in S. Martin (ed), The Construction of Europe: Essays in Honour of Emile Noel ( Kluwer Academic Publishers 1994, Dordrecht/Boston)199-201.
 See: C-167/04 P JCB Services v Commission of the European Communities, EU:C:2006:594, para 207. referring to C-189/01 Jippes and others, EU:C:2001:420, also to T-17/99 Ke Kelit v Commission, EU:T:2002:73 and Dansk Rørindustri, EU:C:2005:408, para 209. Also T-210/01 General Electric v. Commission, EU:T:2005:456 para 516 (cannot deviate from rules binding itself), the Lizin cartel case (Case C-397/03 P Archer Daniels Midland Company v. Commission of the European Communities, EU:C:2006:328) on the conditions of deviation, Case C-112/77 August Topfer & Co. GmbH v. Commission of the European Communities, EU:C:1978:94 (legitimate expectations), see also T-15/02 Judgement in case BASF v Commission of the European Communities, EU:T:2006:74, T-71/03 Joined Cases of Tokai Carbon v. Commission, EU:T:2005:220, T-279/02 Degussa v. Commission, EU:T:2006:103, T-52/02 Union Pigments, EU:T:2005:429, C-3/06 Groupe Danone v. Commission, EU:C:2007:88. Generally on the conditions of deviation see O. Stefan (n 1) 201-227. O. Stefan gives a comprehensive analysis of these, including the principles of non-retroactivity, rights of defence, human rights and transparency, the conflict of which may lead to the deviation from soft law documents.
 This non-absolute self-binding effect is also recognised by the Hungarian Supreme Court (Curia). It held that the competition authority is bound by its guidelines, but may also deviate in a certain case if it gives valid reasons for that (Curia Kfv. II. 37.076/2012/28)
 C-207/01 Altair Chimica SpA v. ENEL Distribuzione SpA, EU:C:2003:451
 Lizin cartel case (Case C-397/03 P Archer Daniels Midland Company v. Commission of the European Communities, EU:C:2006:328 and joined cases of C-80/81-83/81 and C-182/82-185/82 Adam and others v Commission, EU:C:1984:306, para22.; joined cases of C-181/86-184/86 Sergio Del Plato and others v Commission, EU:C:1987:543, para 10.; C-171/00 P Liberos v Commission, EU:C:2002:17 para 35. See also in O. Stefan (n 1) 139.
 C-28/15 Koninklijke KPN and Others, EU:C:2016:310, para 53, 64 and 66
 It is also added that the communication is without prejudice to any interpretation which may be given by the Court of Justice of the European Union. See, for example: Notice on agreements of minor importance which do not appreciably restrict competition under Article 101(1) of the Treaty on the Functioning of the European Union (De Minimis Notice), (2014/C 291/01)
 The latest agricultural guidelines occupy 37 pages in the Official Journal. COMMISSION NOTICE, Guidelines on the application of the specific rules set out in Articles 169, 170 and 171 of the COM Regulation for the olive oil, beef and veal and arable crops sectors (2015/C 431/01).
 C-226/11 Expedia Inc. v Autorité de la concurrence and others, EU:C:2012:795 para 30-31
 AG Kokott also argued that although national courts are not obliged to apply soft law, they should nevertheless consider the Commission’s assessment and also give reasons for any divergence. Expedia opinion, EU:C:2012:544 para 39. For a similar argument see: Oana Andreea Stefan, ‘Relying on EU Soft Law Before National Competition Authorities: Hope for the Best, Expect the Worst’ (2013). CPI Antitrust Chronicle July 2013 (1).
 Tridimas considered general principles constituting abstract propositions of law on which the legal systems are founded. T. Tridimas, The General Principles of EU Law, (2nd edition, Oxford EC Law Library, Oxford University Publishing 2006, Oxford)
 It is a different problem raising constitutional issues when the hard law itself violates general principles of law, for instance a hard law provision infringing the principle of non-discrimination.