2.1. REACH: Law in Progress

The European Chemicals Agency (ECHA) organized its seventh annual Stakeholders’ Day on 23 May 2012 in Helsinki (Finland). At this meeting, representatives of the ECHA, industry associations and non-governmental organizations (NGOs) gave short presentations arranged around each of the four pillars of REACH (Registration, Evaluation, Authorization, and Restriction of Chemicals), after which members of the audience, mostly businesses and NGOs, could ask questions about specific aspects of EU chemicals legislation.^{Footnote 11} The answers given by the ECHA rarely made reference to the legislative text of REACH.^{Footnote 12} Instead, on several occasions ECHA representatives referred to REACH guidance documents that had been developed to assist in the implementation of chemicals legislation. The intriguingly low profile in discussions of the legislative text itself certainly was not because of a dearth of provisions. REACH is printed on 849 pages of the Official Journal and has 140 long and layered articles as well as 17 elaborate annexes.

Despite an impressive thicket of legal rules, as a law REACH leaves open many important questions. This has led some commentators to depict REACH as ‘a framework regulation that leaves many political questions to the discretion of the bodies created for implementation’.^{Footnote 13} The preamble to REACH appears to constitute the following bodies for its implementation. According to recital 24, the Commission establishes so-called REACH Implementation Projects (RIPs), which comprise relevant experts and stakeholders. Under Commission leadership, these groups develop guidance documents intended to assist the Commission, the ECHA, Member States, manufacturers, importers and downstream users of substances to comply with obligations arising from chemicals legislation.^{Footnote 14} While the Commission initiated guidance drafting within RIPs, following the establishment of the ECHA, responsibility for guidance changed hands from the Commission to the ECHA, which now coordinates guidance drafting and has updated many documents initially drafted within RIPs.

By describing and explaining good practice on how to meet legal obligations, these guidance documents promote the implementation of REACH and supplement chemicals legislation by filling in the gaps left in the text of REACH. At present there are altogether 21 guidance documents, taking up approximately 7,000 pages.^{Footnote 15} Faced with this staggering number of pages, some even cry foul play: guidance documents bring in through the backdoor the ‘detailed re-regulation’ that REACH was designed to substitute.^{Footnote 16} Guidance documents are primarily technical notes on different aspects of REACH, but this says little about how they are applied in normative and practical contexts. The main text of REACH makes no reference to guidance documents;^{Footnote 17} they are classic soft law material. They shirk binding commitments: the cover page of each REACH guidance document reads ‘the information in this document does not constitute legal advice’. Nonetheless, they rule in practice. This pragmatism is reflected in recital 35 of the preamble, which provides that ‘the Member States, the Agency and all interested parties should take full account of the results of the RIPs’.^{Footnote 18}

Guidance writing is a shared effort, drawing upon the input of Member State authorities, the Commission, the industry, and other stakeholders, including NGOs.^{Footnote 19} To smooth the path of decision-making, the ECHA has published a (once again non-binding) Consultation Procedure on Guidance, which sets out a detailed manual for guidance drafting.^{Footnote 20} The preparation of a new guidance or amendment, or the updating of existing guidance, starts with the ECHA producing a draft, usually in collaboration with external experts. The ECHA will also seek the Commission’s opinion, in particular in cases that involve legal interpretation of REACH. The draft is then submitted for consultation to a Partner Expert Group (PEG)^{Footnote 21} and the two scientific committees of the ECHA and/or a Forum,^{Footnote 22} – the scientific committees and the Forum become involved only when their contribution is seen to be necessary. With consultations completed, the ECHA will prepare a final draft, taking the comments on board. The next step is the concluding consultation between the ECHA, the Commission, and national competent authorities. These concluding talks aim to ensure that the new guidance, amendment, or its revision is accepted and implemented in a coordinated manner by all authorities. REACH guidance drafting is a consensual practice, where agreement is attempted in the first place through a written procedure. If consensus is achieved, the guidance is adopted at the stage of concluding talks; if not, a meeting will be arranged. Here the majority prevails if the participants fail to agree on a joint decision on the content of the guidance document.^{Footnote 23} Finally, the ECHA will place the guidance and dissenting opinions on its website. The chart above summarizes the procedure.

Figure 1 Procedure for REACH guidance drafting

2.2. A Closer Look: Guidance on Requirements for Substances in Articles

As every REACH guidance document is the outcome of a tailored process, born out of the need to bring clarity to an unclear situation, it is instructive to look more closely into the drafting history of one particular guidance document. This is all the more vital since, despite the manifest importance of these networks to legal practice in the EU, empirical knowledge of their activities is still scarce.^{Footnote 24} The following analysis is intended to redress this glaring gap and to describe how REACH rule-making occurs in practice. I have chosen the Guidance on Requirements for Substances in Articles (SiA Guidance) for a case study.^{Footnote 25} It is an interesting, though controversial, choice for a case study as several Member States opposed the definition of an ‘article’ adopted in the guidance. In view of this opposition and efforts to resolve the stalemate that has arisen, the SiA Guidance is not a typical example of guidance writing. At the same time, it is precisely for this reason that this guidance is worth studying: the tensions and mechanisms for resolution can teach us a lot about the ways in which post-legislative rule-making creatively adapts and evolves in the face of challenges. Because the SiA Guidance relates to a term that has an actual legal definition in the REACH Regulation, it is an informative example of how rules change. Article 3(2) of REACH provides that an article ‘means an object which during production is given a special shape, surface or design which determines its function to a greater degree than does its chemical composition’. The case study opens a small window into how rule-making translates and transforms underlying framework norms. In so doing, it makes real the ‘dialectic between the enunciation of a new object of government in a legal text, and the administrative and technoscientific instruments that give it material meaning’.^{Footnote 26} By analyzing this dialectic, the case study foregrounds the discussion of accountability.

Behind the seemingly technical controversy regarding the definition of ‘article’ lies a problem with real consequences. Article 7(2) of the REACH Regulation requires a producer or importer of articles to notify the ECHA if a substance is a ‘substance of very high concern’ (SVHC) and if an article contains more than 0.1% of this dangerous substance.^{Footnote 27} In addition to complying with this relatively cumbersome notification procedure, a producer or importer of an article must provide information to customers and consumers on safe use of the article. Now much depends on how an article is defined. Taking the example of footwear being imported into the EU, the question arises as to whether a shoe is an article. The intuitive answer is ‘yes’. However, another equally plausible interpretation presents itself. A shoe can be perceived as an ‘umbrella’ article, giving rise to a regulatory visualization of a shoe that is made up of shoelaces, insoles, heels, eyelets, etc., as articles in their own right. Because the presence of a SVHC in an article is determined by its proportion of weight, the smaller the article, the more easily the notification requirement kicks in. Producers and importers are therefore significantly better off if shoes rather than shoelaces are defined as an article.

In essence, REACH rule-making was confronted with a task of statutory interpretation: a need to define what constitutes an ‘article’. Confronted with such a herculean task, the adoption of the SiA Guidance was set in motion. The work was initiated within REACH Implementation Project RIP 3.8, led by the Commission services as preparation of the guidance had started before the ECHA was up and running (before 1 June 2007). In addition to the Commission, a total of 19 participants were involved in the work: seven Member States, 11 industry representatives, and a Japanese representative.^{Footnote 28} A difference of opinion emerged between a group of Member States and the Commission. Of the participating Member States, Denmark, France and Germany supported the view that the 0.1% threshold should be made applicable to shoelaces, challenging the Commission’s preference to apply the threshold for shoes. Member States were not opposed just on principle. They were particularly worried about the loss of information caused by the interpretation championed by the Commission. Namely, the stance adopted by the Commission means in practice that producers or importers of articles with SVHC content above 0.1%, which usually triggers notification and information requirements, escape the requirements if these articles are put together as one large article in which a SVHC remains below the critical threshold of 0.1%.

In its opinion of 2007, the Commission, with reference to internal consultations with its Legal Service, rejected the interpretation by the three Member States in favour of the interpretation that the threshold applies to the whole article, not to its individual parts. The first version of the SiA Guidance was adopted in May 2008 along the lines of the Commission’s view, carrying a disclaimer for the position of the opposing Member States.

In the summer of 2009, the ECHA initiated the process of updating the SiA Guidance with a view to ironing out the disagreement that had flared up. With the first deadline for notification and communication obligations approaching (1 June 2011), the ECHA wanted to further clarify the matter and, remarkably, to publish the guidance without reference to the dissenting Member States. In line with the Consultation Procedure on Guidance, the ECHA prepared the draft, which it then modified after receiving comments from outside experts. The PEG consultation took place in two stages, application of the 0.1% threshold forming part of the second round. This consisted of 20 stakeholders: nine Member States, eight business representatives, one NGO and the Commission Directorates–General (DGs) Enterprise and Environment.^{Footnote 29} Altogether 83 comments were submitted, of which 12 touched upon the 0.1% issue.^{Footnote 30} The summary of comments reveals that the initial coalition of Member States was reinforced by new members. In addition to Denmark, Germany and France – which renewed their objection – Sweden, Austria and Norway, as well as one NGO (the World Wildlife Fund) appeared. Pleading new reports, they suggested a number of amendments to the part of the guidance concerning application of the 0.1% threshold.^{Footnote 31} By contrast, five industry representatives expressed their satisfaction with the interpretation of the draft guidance and observed that the contrary interpretation would contradict the legal text of REACH. Providing a brief reply to each comment, the ECHA noted that the Commission’s view had been imparted to Member States at the PEG meeting. There, the Commission had invoked the convenience of a hypothetical legislator and argued that ‘[i]f the legislator had wanted [the 0.1% threshold to apply to homogenous materials only] … the legislator would have explicitly mentioned this’.^{Footnote 32}

The matter came up again in the consultation of the Forum and Member State Committee in spring 2010. In addition to the Member States that had opposed the Commission interpretation in the PEG round, Belgium joined the ranks. These Member States questioned the Commission’s point of view, illustrating inconsistencies in its interpretation. In their opinion, the Commission was virtually arguing that an article ceased to be an article when it was assembled into another article, the two items forming something that the Commission would describe as a ‘complex article’.^{Footnote 33} The Member States were particularly interested (almost to the point of absurdity) in knowing whether the article would become an article again when it was disassembled. The business was further discussed in the meeting of the Competent Authorities (CARACAL). The Commission ended up sticking to its guns and the Member State argumentation fell on deaf ears. The Commission argued that, in view of the approaching deadline for notification, ‘the principle of legitimate expectations might be questioned if the previous guidance on the 0.1% threshold were overturned just a few months before the deadline for notification without any change in legislation’.^{Footnote 34} This again led some Member States to complain that the Commission had delayed the announcement of its final view, leaving Member States with insufficient time to reflect on it.

In April 2011, the ECHA published the updated SiA Guidance, accompanied by a fact sheet informing readers that not all Member States endorse the interpretation of the guidance. Austria, Belgium, Denmark, Germany and Sweden left dissenting opinions which were added to the ECHA website.^{Footnote 35} References to the controversy or dissenting views have been deleted from the final text of the SiA Guidance.

2.3. REACH Rule-making: Some Critical Comments

From what can be inferred from the guidance drafting process, the ECHA is a major player. Firstly, it coordinates guidance writing and maintains an online repository for guidance documents as well as for other information on chemicals legislation. In the electronic corridors of the vast database that the ECHA manages, information surrounding the guidance writing process is relatively easy to retrieve. Yet the ECHA’s success should not be measured by its capacity to store information but by how readily it discloses that information. Two points for improvement should be addressed: (i) in the case of enquiries for further information about a particular guidance document, the enquirer is at a loss; guidance documents do not designate anyone as a contact person who can answer questions or provide additional information regarding the process of guidance writing or document content, and (ii) although step-by-step information on the consultation procedure can be found on the ECHA website, final guidance documents do not include summary information on the process (who was consulted, how many comments were received, and so on). This stands in regrettable contrast to guidance documents adopted in the framework of implementation of the Water Framework Directive (WFD):^{Footnote 36} each of these WFD guidance documents, which currently number 28, includes a list of participants and their organizations.^{Footnote 37}

In spite of these shortcomings, by and large information on REACH guidance has been made available and is documented so as to allow for later scrutiny. For instance, documents from the consultation rounds are saved on the ECHA website as Excel spreadsheets in which positions are listed and controversial issues are flagged, with notes in the spreadsheet on reasons why comments are accepted or rejected.

Secondly, the ECHA has an exclusive right both to start the process of drafting a new guidance and to update existing guidances. In this respect, it holds an agenda-setting position. The influential role of the ECHA in the guidance writing process has not gone unnoticed and has spurred suspicions that may potentially undermine implementation projects. Scholars comment that ‘it might be an unfortunate construction that the specific operationalization and application of the generic legal standards (such as adequate control) has to be dealt with by the same bodies that prepare and hence essentially shape the risk related decisions’.^{Footnote 38} Its influence is, however, constrained by the Commission’s role in REACH implementation.

Anecdotal evidence available from the drafting history of the SiA Guidance seems to support the view that guidance-making opens the door for the Commission to wield its power. Notably, it is the Commission that benefits from disagreement. The Consultation Procedure on Guidance that was applicable at the time provided that, should Member States remain in disagreement, the ECHA Executive Director decides on the final text of the guidance document, after consultation with the ECHA Management Board.^{Footnote 39} However, the Commission’s view proved to be decisive for finalizing the text of the document. In the summary of PEG comments, the ECHA endorsed the authoritative status of the Commission opinion and noted that ‘the legal opinion of the European Commission is determining for ECHA unless overruled by a decision of the Court of Justice’.^{Footnote 40} One of the opposing Member States, Sweden, pointed out that during the SiA Guidance drafting the Commission virtually created new terms, such as ‘component’, ‘autonomous article’, and ‘complex article’.^{Footnote 41} The special status of the Commission is also clear from an analysis of comments supplied during consultations. For instance, in the case of the PEG consultation, the Commission participated but did not submit comments, which implies that it can get its point across in other ways and can, as a result, forgo commenting altogether. The lack of traceable imprints of influence is worrying, as the Commission may not always speak with one voice and guidance drafting can end up being plagued with latent controversies. For instance, in the case of the 0.1% disagreement, the PEG consultation included two DGs: Environment and Industry. As a coalition of Member States opposed to the 0.1% interpretation on environmental grounds, it is difficult to believe that the views of both DGs were unanimous.

The strong, if not dominant, role the Commission plays in guidance drafting is also problematic from the perspective that it destroys the level playing field (or at least in principle ‘egalitarian’ peer culture) among network participants.^{Footnote 42} The development of the SiA Guidance shows how those Member States that adopted views in opposition to those of the Commission were identified early in the process as ‘troublemakers’ that needed to be proselytized. The strategy which was deployed here, as well as in network decision-making more broadly, is the power, and politics, of repetition.^{Footnote 43} We recall that the update of the SiA Guidance was not triggered by the need to take into account recent scientific or technical developments. Instead, it was launched in order to reach a common understanding of the application of the 0.1% threshold and to publish the guidance without dissenting footnotes. The involvement of the Commission Legal Service, in delivering an opinion on the 0.1% limit, raises a question of procedural fair play. As is clear from the above, the Member States felt that more than a half year during which the Commission spent elaborating the issue was too long, especially given that its opinion was made available to them just before the CARACAL meeting.

Finally, the role played by non-business stakeholders and NGOs remains shrouded in uncertainty. Despite the occasional glimpses of societal legitimacy in the Consultation Procedure on Guidance,^{Footnote 44} non-institutional interested parties, with the exception of industry partners, do not play a noticeable role in the guidance writing process. For instance, the SiA Guidance was drafted, negotiated and adopted with the participation of one NGO only. Non-institutionalized stakeholders may be involved in guidance drafting only through PEGs.^{Footnote 45} According to the revised Consultation Procedure on Guidance, the PEG is set up on an ad hoc basis for each guidance document, although the creation of a permanent PEG is also anticipated.^{Footnote 46} Getting into a PEG is premised on prior institutional friendship with the ECHA, as PEG members are chosen from those partners accredited by the ECHA.^{Footnote 47} When preparing a new guidance, the ECHA invites accredited stakeholders to nominate experts, one for each organization. Those thus nominated and notified to the ECHA may participate in the PEG without further involvement with or assessment by the ECHA. On the face of it, there are no institutional hurdles that would put non-institutionalized stakeholders in a disadvantaged position to compete for influence and visibility.

More worryingly, however, the perception of what the mandate of the PEG is and how it is supposed to function in the context of REACH rule-making appears to have become lost somewhere along the way. Generally speaking, the Consultation Procedure on Guidance justifies consultation on two grounds. In the first place it explicitly acknowledges that guidance documents are ‘of a highly technical nature’. However, in a moment of rare frankness about the nature of these guidance documents, it continues that the adoption of guidance documents is not just dispassionate technical wrangling but ‘require[s] interpretation of the underlying regulation(s)’, thus acknowledging the policy choices inevitably involved in guidance drafting practices.^{Footnote 48} This again affirms that, at least in theory, the ECHA recognizes the need to gain both expert knowledge and societal legitimacy. However, recognition does not translate into a guarantee that societal legitimacy de facto enters into guidance drafting. As noted, non-institutionalized actors – meaning those with potentially other than technical rationality and training – may access guidance drafting only through the PEG group, whose mandate is construed quite narrowly on ‘technical content issues’. This renders mostly symbolic the role of the PEG as a platform to present, for instance, NGO views.^{Footnote 49} Accredited NGOs may feel hindered in nominating an expert because of a lack of technical content expertise where, in reality, scientific issues disguise problems of a different kind. Answering questions such as whether something is workable for industry, environmentally friendly, or accessible and safe for consumers is more often than not an exercise in balancing between conflicting objectives.

Two years after the adoption of the second version of the SiA Guidance, the situation is very much open. Enforcement of the REACH Regulation falls on national authorities, which, in the case of the six disagreeing Member States plus Norway, is a concern for non-EU importers who raised the issue of trade distortions at the meeting of the Asia-Pacific Economic Cooperation (APEC) in Moscow on 4 February 2012. The Executive Director admitted that importers may end up having to adopt varying practices in different Member States, making it a compliance nightmare for foreign companies. He continued, however, that ‘in many cases’ where the importer is providing spare parts, it would in any event have to provide information on the components of the product to comply with the notification requirements.^{Footnote 50} In September 2012, the ECHA representative confirmed that no further development had been made.^{Footnote 51}

3.1. Demarcating Network Accountability

In the academic literature, implementation networks, such as those found in the realm of REACH rule-making, are called ‘forums’.^{Footnote 54} They are characterized by having an informal institutional skeleton, emerging and operating in an area that has traditionally been reserved for domestic implementation actors but is now transnational, incorporating European and global actors. The participant pool is diverse and network working methods are usually associated with forms of learning and benchmark setting, reflecting a ‘reciprocal and heterarchical relationship between participants’.^{Footnote 55} For lawyers, networks pose two general yet fundamental questions: (i) what is the significance of these networks for law and legal practices, and (ii) how can they be subjected to (judicial) oversight?

The relevance of networks for law eludes definitive understanding and categorization. Crucially, this is down to the wide variety of contexts in which networks operate, thus prescribing rigorous empirical evaluation. The importance of empirical analysis cannot be overstated: the more essential network activities become in establishing themselves as reference points for all those working in the relevant field, the more important it becomes to know how their authority is constructed and maintained and whether it can effectively be challenged from outside. Indeed, the second crucial question arises with regard to accountability: what is the extent of autonomy that these networks and other such loose organizations enjoy and how can their external control adequately be arranged? It is observed that ‘networks often exhibit autonomy in their promulgation of norms of best practice and benchmarking in ways that are not clearly susceptible to control’.^{Footnote 56} There are no readily identifiable principals to whom actors report, and the most common accountability mechanism is peer review, in which network actors explain their actions and have their reasoning scrutinized by network peers. In the discussion and analysis of networks, a ‘peer’ is understood widely. A peer is a member for ‘mutual monitoring of one another’s performance within a network of groups, public and private, sharing common concerns’.^{Footnote 57} Peer control is considered as affording a rather weak form of (democratic) accountability, to the extent that it is democratic at all. It is often dismissed as a form of free-floating accountability, not really contributing to the common pool of public accountability. Underlying this view is a suspicion of an impenetrable community, which begs the question of how to ensure that those outside have a sufficiently informed understanding of what is going on inside.^{Footnote 58} In so far as networks are marginal actors, this does not have to be a particular concern, yet the situation is radically different when networks put down roots and gain authority. As an instance of such regulatory entrenchment, this article has analyzed network activities transforming REACH framework rules. It is precisely this transformation that fuels the need for external, perhaps most importantly judicial, control.

Judicial review of new regulatory forms has been subject to heightened interest in recent years. Observers seem to be united in the inability of judicial review to control and curb autonomous network actors.^{Footnote 59} A wide variety of reasons have been given for this, ranging from the transnational character of network activities that result in decay of the traditional delegation doctrine to the soft law character of many network ‘outputs’ and lack of substantive expertise on the part of the courts to make any meaningful intervention. The shared conclusion is that the traditional command-and-control approach needs to be updated to better reflect the reality of network activities.^{Footnote 60}

One such modernized approach is democratic experimentalism, as elaborated primarily by Charles Sabel and Jonathan Zeitlin. It is sceptical of any version of constructions that rely on either delegation or the principal–agent relationship to hold actors to account. It is also fundamentally distrustful of courts in a world of ‘what-ifs’: increasing complexity, turbulent scientific progress, and transnational connections.^{Footnote 61} In a recent contribution, Sabel and Zeitlin make an interesting link between type of norm and judicial accountability. In their view, courts are increasingly in deep water and cannot establish or police lines of accountability, because when actors seek solutions to problems of different kinds, rules will inevitably change. Not only does this apply to framework norms that provide very little clarity but equally to norms that are given a very specific definition in legislation. The concept of ‘article’ is verbosely enshrined in REACH, yet it cannot guide actors to a sufficient degree and is thrown into a normative limbo. Consequently, the courts are deprived of, or confused about, rules against which to measure the behaviour of actors.^{Footnote 62}

The suggested solution by Sabel and Zeitlin suits the context of framework regulation and networks. They highlight internal mechanisms prevalent in networks, inducing peer review which ‘becomes in turn dynamic accountability’.^{Footnote 63} As an alternative to principal–agent accountability, dynamic accountability accepts the transformation of rules ‘in use’ and exploits as a form of review those particular processes that occur when norms are ‘used’. REACH rule-making serves as one of many examples of such processes. Although theoretically and empirically convincing,^{Footnote 64} this kind of dynamic accountability is, however, somewhat unsatisfactorily connected to other established forms of (external) accountability, leaving peer control as a legitimate and effective mode of accountability undetermined and vulnerable to doubt. Therefore, the question about the relationship between external and internal accountability mechanisms in relation to networks involves an unexplored point of synergy. Developing the idea further and taking into account the context in which framework norms coexist with a maze of lower level rule-making, I suggest that peer control can be connected with judicial review. Just as post-legislative rule-making accompanies much framework regulation, peer control can be incorporated as part of judicial control.

Immediately below I present some concrete examples of how peer control and judicial supervision can pool resources and expertise to make sure that gaps in framework laws do not translate into gaps in accountability.

3.2. Peer Control and Judicial Control

The incorporation of peer control into the review function of courts can take many forms, ranging from: (i) the use of guidance documents to fill in gaps left by framework norms, through (ii) assisting courts to select the optimal level of intensity of review, to (iii) keeping conflicts out of the purview of the courts. I address each of these separately.