7.1 Introduction

So far, empirical research in EU law has focused on courts and is heavily dominated by quantitative approaches, as if legal debates took place in judicial institutions and could be analysed numerically. This focus disregards qualitative empirical analysis of law and institutions, ignoring the political and administrative processes, from preparatory stages to implementation, where the law is interpreted and operationalised every day. Key actors in these processes include lawyers and legal experts operating in various roles in the EU institutions.Footnote ¹ This chapter zooms in on legal data, such as legal opinions and other types of advice, and soft law documents involving the interpretation of law, outside the judicial context, in particular in the Commission and the Council.Footnote ² Their legal services are particularly powerful in EU policy-making. Their mutual relationship is competitive yet symbiotic.Footnote ³ Legal services shape political decision-making in their institutions acting in the ‘background’, exercising their power of framing.Footnote ⁴ I understand EU law primarily as a language of argumentation to justify the choice in the service of a particular purpose. I treat internal legal work within the EU institutions as an exercise in professional legal speech that makes certain outcomes seem credible, necessary, or even unavoidable.Footnote ⁵ When legal doctrinal analysis reveals a range of possible interpretations, legal experts add professional judgment, which makes some interpretations plausible and others implausible according to ‘the conventions of their expertise and the pratices of their profession’.Footnote ⁶ Qualitative research methods can be used to make these interpretative practices visible and mirror them against alternative readings of the law that may lead to other consequences and priorities becoming dominant, enabling a critical examination of the politics involved. Access to data is key to knowledge. This point is regrettably often ignored when scholars construct research agendas and ask research questions based on the availability of data. I believe that there is a need to look deeper than the data that the institutions provide on their websites or make proactively available through their public registers.

In this chapter, I argue that systematically used access to documents (ATD) requests can be a source of data about the law-making and governance practices for a study of how law operates in the EU’s political and executive institutions. Unlike EU primary or secondary law or case law, this kind of legal data is not traditionally understood as an authoritative source of EU law. However, for practical purposes in EU governance processes internal legal documents are influential. They provide an authoritative reading of legal requirements for the purposes of institutional decision-making. In social sciences, ATD requests are used to empirically document the workings of elite power, authority, and governance.Footnote ⁷ In such research methodologies the information-brokering process is also considered a valuable source of research data.Footnote ⁸ The objective is to gain access to streams of information that illustrate everyday practices of government agencies and produce records that can be further analysed using various qualitative data analysis techniques. Such records include various texts used in government processes or produced as part of governing, including also unofficial data never intended for public circulation, such as notes, internal memos, and emails.Footnote ⁹ As a method, ATD is ‘bound to attract all who are curious, passionate about research, and who are interested in investigating practices and processes of governing’.Footnote ¹⁰

Despite this promise, ATD and freedom of information (FOI) requests have been little used as a systematic method in the context of EU legal scholarship. Article 15 TFEU and Article 42 CFR establish a right to access documents held by the EU institutions.Footnote ¹¹ It is unclear whether the limited use of this right is more due to the lack of knowledge about its existence or difficulties in gaining access to this material. Yet, given the narrative of the centrality of law and lawyers in the European integration process, there remains a great deal to study. Various historical studies have demonstrated how through the process of intepreting the Treaties, lawyers working in the EU institutions have played a deeply political role in European integration, picking, for a purpose, the winning arguments. This professional legal speech is largely invisible outside the EU institutions and can only be studied based on internal data. When a legal issue reaches a policy-maker for determination, a legal expert has typically framed the possible alternatives, thus exercising ineradicable influence on policy outcomes. It is easy to find such examples in the EU’s daily law-making practices, where a competent legal expert may frame the same question as a matter of environmental policy or social policy;Footnote ¹² or as an exercise of cohesion policy or economic and fiscal policy (which this chapter uses as an example). This choice settles the EU’s Treaty competence to legislate and the impact its legislative action has on national policy choices. This framing is conducted in legal language, and guided by institutional preferences.Footnote ¹³ This kind of legal data includes in particular legal advice, which the Court has defined as

The data may also include legal documents originally drafted outside the institutions. The Court has accepted that even informal ‘exchanges of legal views between the legal services of three institutions in order to reach a compromise regarding a legislative text in the context of a trilogue may, where appropriate, be described as legal advice’.Footnote ¹⁵ While the role of lawyers and legal argumentation has so far been primarily studied in historical contexts, their impact has become particularly crucial during the past ten years when invisible ‘legal engineering’ within the institutions, in particular in the context of crises, has become a substitute for Treaty reform.Footnote ¹⁶

I build on a combination of methods, including the use of expert interviews.Footnote ¹⁷ Like many others, I would caution against overreliance on interviews as a single source of empirical data on government practices. Officials – and EU officials very specificallyFootnote ¹⁸ – are often ‘restricted in what they can say, calling the validity of such data into question’.Footnote ¹⁹ There is a risk of ‘spin and bowdlerization’, where interviews easily turn into an exercise in impression management.Footnote ²⁰ While ATD requests do not exclude the risk of spin entirely, they are directed at disclosure of internal government recordsFootnote ²¹ and thus provide a way to ‘peer into the everyday worlds of these insulated communities and organisations’.Footnote ²² Ideally, the resulting data is triangulated with interviews and analysis of official organisational discourses, used to complement and contextualise the data.Footnote ²³ Since ATDs also enable the disclosure of even politically sensitive and potentially contentious data,

My background is in the critical school, which seldom engages with ‘empirical’ evidence of the social ‘impact’ of law or the behaviour of legal actors.Footnote ²⁵ Legal data gained through ATD can be used for qualitative, critical, and post-structural analysis of policy-making, law, and institutions. In particular discourse analysis on legal data provides additional perspectives in allowing moving ‘outside the text’Footnote ²⁶ to study the ‘subtext, and representation to uncover issues of power relationships that inform what people think and do’.Footnote ²⁷ It can be used to ‘identify the regulatory frameworks within which groups of statements are produced, circulated, and communicated within which people construct their utterances and thoughts’; and ‘uncover the support or internal mechanisms that maintain certain structures and rules over statements about people … as unchallengeable, “normal” or “common-sense”’.Footnote ²⁸ From the internal perspective of law, the way of organising and selecting legal arguments to support particular political aims is particularly interesting. In addition to studying arguments, the method also studies the silences of texts, as well as questions of when, where, how, and why a text was produced.Footnote ²⁹

This chapter starts by describing the legal data which exists in the Commission and the Council and the key processes in which it is produced. The application of ATD as a method requires

The chapter describes the process of gaining public access to legal data. My requests have usually led to confirmatory applications and action before the EU Courts and the European Ombudsman. I will also present examples of legal data gained through ATDs and how they can be used in research. Finally, the chapter will discuss how more recently, in particular, the Commission has restricted access further by making ATD dependent on what legal analyses have been entered into formal registers, excluding more informally provided advice from the scope of ATD. This section relies on an ATD request made specifically for the purposes of this chapterFootnote ³¹ and raises concern about the future possibilities of conducting research on the handling of legal questions within the institutions, which is threatened by an increasing problem of empty archives.

7.2 What Legal Data?

Lawyers take many roles within the EU’s political institutions. Officials with legal training often work in Directorate-Generals (DGs) and as policy advisors. The most important legal work, however, takes place in the legal services of the institutions. These units have the formal institutional position and the related authority to solve legal questions for the purposes of institutional decision-making.Footnote ³² Most of their legal work is technical routine and takes place below the political radar. A key part is anticipating how the Court would evaluate the situation at hand, which de facto serves as the ex ante constitutionality control conducted in the EU institutions.Footnote ³³ Few matters are ever appealed to the Courts. If they are, lawyers from the legal services will assume centre stage defending their institutions, building on legal argumentation carefully developed during the preceding stages of policy-making. They will also exercise authority to determine what policy implications rulings of the Court will have after a judgment has been delivered.

The Commission Legal Service is closely integrated into Commission decision-making. Invisible to outside audiences, it is present whenever decisions are made and new EU policies developed. According to the Commission website, during 2021 the Legal Service replied to 17,318 consultations of which 10,536 were on legislative drafts.Footnote ³⁴ It is big and well-resourced with around 445 staff members.Footnote ³⁵ Its power stems from its formal role in the Commission decision-making procedure. Other Commission services have a duty to consult its Legal Service ‘on all drafts or proposals for legal instruments and on all documents which may have legal implications’.Footnote ³⁶ On most technical matters the Legal Service has a full veto, so the relevant DG must work in cooperation with the Legal Service.Footnote ³⁷ Advice is given orally or in emails between DG officials and members of the Legal Service.Footnote ³⁸ The Legal Service does not dominate political decision-making in the Commission, but often shapes its decisions and determines their reach. The Commission is a collegial body, and the Legal Service’s job is to defend its agenda. Its formal role is particularly strong when the College takes decisions by way of a written procedure,Footnote ³⁹ as the procedure is available only with the approval of the Legal Service. While many Commission Legal Service members are well-known profiles in the legal academia,Footnote ⁴⁰ they seldom make appearances outside the Commission in inter-institutional contexts. However, if fundamental disagreements appear with the Council lawyers, the Commission may send its legal advisers to defend its position and even produce a non-paper. With these rare exceptions, most Commission legal advice is intended for its internal use and is invisible in public registers.Footnote ⁴¹ What therefore is usually most interesting for legal research is not the formal position of the Legal Service in the final inter-service consultation required by the Commission Rules of Procedure (often half a sentence in an Excel sheet), but the advice given when Commission proposals or other actions are planned, in other words, how the Legal Service has interpreted the legal requirements forming and shaping future EU action.Footnote ⁴² This advice is offered in emails, memos, or as track changes to Word documents.

The Council Legal Service (CLS) is equally powerful, but in a different way.Footnote ⁴³ It is a much smaller unit that provides a first-instance peer review for the legal solutions suggested by the Commission. When Council or inter-institutional decision-making runs into trouble, the CLS smooths political fractures through legal language. Its proposals constitute the foundations of Council decision-making, and it is present in drafting sessions used to prepare Presidency proposals. It is the most powerful Legal Service at the point where final decisions are taken. Unlike the Commission, the Council is not a collegial institution, and its political will is more heterogeneous, which grants the Council’s Legal Service a great deal of leeway in defining its goals. It is always present in Council bodies and exercises an independent, strong voice. As the key consumers of its opinions, the Member States, screen and assess its work. However, the view of the Legal Service also influences the position of many Member States. Much of its advice is provided orally, some of which may be recorded in other Council documents. The CLS also gives formal written opinions when requested to do so by the Council. These contributions can usually be found in the Council’s public register, but are seldom proactively disclosed.

Even if legal opinions are well protected, I have gained access to many. The notes and opinions of the legal services describe the state of the law, with references to case law and EU legislation, sometimes from a completely different field. They may include concrete suggestions for amendments, as well as recommendations against certain approaches and in favour of others. In external relations, legal services engage with the process of signature, provisional application, conclusion and implementation of international agreements.Footnote ⁴⁴ Legal opinions outline the choice for a correct legal basis, and assess the compatibility of envisaged measures with already existing internal rules or the EU Treaties.Footnote ⁴⁵ Legal Services may also be asked for reports, studies, research notes or documentation, ‘a systematic presentation of the case law of the ECJ or, failing that, of the precedents in the institution’.Footnote ⁴⁶ Whether a document is called ‘legal opinion’ is irrelevant. According to the Court, ‘[o]ver and above the way a document is described, it is for the institution to satisfy itself that that document does indeed concern such advice’.Footnote ⁴⁷

The expertise of institutional lawyers is collective and cumulating knowledge, which is recorded in these opinions and approved collegially. Legal opinions are discussed within each legal service to ensure historical and horizontal consistency. Conceptual distinctions are stabilised in path-dependent ways. Through this process, a doctrine is developed and refined, which will then be reflected and referenced in future opinions.Footnote ⁴⁸ This method also ensures a continuous and coherent line of interpretation and gives findings a sense of permanency and credibility. This language is different from the ‘law in books’ taught in law schools. It includes bureaucratic expertise of knowing the EU’s inner workings. It is produced by EU officials, paid to see things from the institutional perspectiveFootnote ⁴⁹ – a perspective that they either prepossess or assume efficiently. As Jean-Claude Piris, the former highly influential Director-General of the Council’s Legal Service put it, ‘you have to know precedents, but you have to be creative. So you do not consult books or scholars, but political people involved in the matter. You have to follow the fight.’Footnote ⁵⁰ ‘Following the fight’ may also involve realities that persuade a legal service to engage in serious re-interpretation of its own doctrine.Footnote ⁵¹ These instances (where the illegal may suddenly become legal) are a particularly interesting source of critical analysis. They may build on highly selective use of case law or silences as regards the choice of sources or arguments, gearing conclusions to support specific political outcomes.

In addition to studying legal advice, writing ‘about law’ requires taking into account its actual operation. In addition to actual legal advice, the institutions possess various legal documents that take the form of internal working documents and arrangements, often formally non-binding. An example is my recent attempt to examine the practical implications of the conclusions of the July 2020 European Council to task the Economic and Financial Committee (EFC), to oversee disbursements from the Recovery and Resilience Facility.Footnote ⁵² My ATD request filed with the Council resulted in three documents, which the Council first agreed to disclose on appeal. These guidelines were worthy of trouble, as they reveal the practical operation of the Facility: scrutiny is abstract, conducted under tight deadlines, and very limited by scarce resources. The documents also reveal a process that most resembles a formality: the idea of ‘very tight control’ reflected in the European Council Conclusions is watered down by bureaucratic routine. This is an empirical reality that could not possibly be grasped by looking at the secondary legislation and will form the basis of our forthcoming monograph.

7.3 How to Access Legal Data in the EU Institutions

Beyond leaks that sometimes appear in Politico or the Financial Times, access to EU institutional legal opinions remains highly restricted and relies on public access requests under the EU’s public access legislation.Footnote ⁵³ Regulation No. 1049/2001 establishes the principles and rules concerning access to documents held by the Commission, Council, and the European Parliament. Unlike in some national systems,Footnote ⁵⁴ ATD for research purposes has no specific status in Regulation No. 1049/2001. It is also possible to file more general FOI requests using the ‘Europe Direct’ platform, which may be a quick and efficient means to gather information that is not recorded in any document.Footnote ⁵⁵ This section describes the process under Regulation 1049/2001 (the Regulation) while also offering some examples of legal issues that have proved interesting for my own research.

The Regulation establishes that all documents held by the institutions are, as the main rule, public. They can however be fully or partially withheld in case their substance relates to a protected interest. As far as legal documents are concerned, Article 4(2) enables the institutions to refuse access in case disclosure would undermine the protection of ‘court proceedings and legal advice’. The institutions may also refuse access to documents ‘drawn up by an institution for internal use’ and relating ‘to a matter where the decision has not been taken by the institution’, in some cases even after the relevant decision has been taken. However, both of these exceptions require the institution to balance the potential harm from disclosure against an overriding public interest in disclosure. Under established case law, the risk to protected interests must be reasonably foreseeable and not purely hypothetical.Footnote ⁵⁶ Under Article 12(2), legislative documents, understood as those ‘drawn up or received in the course of procedures for the adoption of acts which are legally binding in or for the Member States’, should be made directly accessible through public registers.

For those documents that have not been proactively disclosed (as is typical with legal data), the Regulation lays down a two-stage administrative procedure requiring the institution to examine the requested documents and provide an answer within fifteen working days. Alternatively, the institution can provide the reasons for the total or partial refusal, which enables the applicant to ask the institution to reconsider its position in a confirmatory application. Further, ‘[i]n exceptional cases, for example in the event of an application relating to a very long document or to a very large number of documents, the time-limit … may be extended by 15 working days, provided that the applicant is notified in advance and that detailed reasons are given’ (Article 7–8).

It has been difficult to convince the institutions about the merits of the regime, and the lawyers working for them are no exception. Lawyers very much like to keep the circle where the evaluation of their work takes place small, and insist that their advice should be kept confidential. They rely on the principle of non-accountability, which is a classic part of legal professional rules even if it is questionable how it applies to officials working in public institutions.Footnote ⁵⁷ In responding to ATD requests and subsequent challenges before the Courts, the institutions emphasise how ‘public interest requires that the EU institutions should be able to benefit from the advice of its legal service, given in full independence’; how their legal advice is to be understood as ‘purely internal exchanges’ that should ‘be as a rule protected as part of the institution’s “space to think”’, and justify this conclusion with reference to ‘the specific dual nature of the Legal Service, as both the sword and the shield of the legality of Union acts’. Their ‘advice should always be “frank, objective and comprehensive”’.Footnote ⁵⁸

What advice counts as ‘frank, objective and comprehensive’ is primarily a matter for other legal professionals to assess with reference to their shared professional standards and skill.Footnote ⁵⁹ The institutions have been unified in their opposition to disclosing their work to broader outside scrutiny, and insisted that these standards can only be upheld in case their legal advice remains confidential. The EU Courts have not been convinced, but instead clearly established that public access should, as the main rule, be granted in the legislative context:

The legal services further argued that disclosure of legal advice might ‘lead to doubts as to the lawfulness of the legislative act concerned’. However, the Court has stressed:

The Court accepted that access to legal opinions can be denied temporarily and in exceptional cases if the advice is ‘of a particularly sensitive nature or having a particularly wide scope that goes beyond the context of the legislative process in question’.Footnote ⁶² With this passage, it created a fresh battleground for later court cases.

Both the Commission and the Council have put a great deal of effort into exploring the boundaries of this exemption. When Samuli Miettinen applied for access to certain legal opinions for the purpose of his PhD thesis in EU criminal law, which examines the contours of the EU’s criminal law competence under Article 83 TFEU, including how the institutions’ themselves have analysed their competence to act, the Council was unwilling to support his critical exploration. It argued that ‘particularly sensitive’ could apply to whole policy areas (such as EU criminal law) and that there was institutional disagreement on the legal basis. Particular sensitivity could also be justified with how the relevant legal basis was ‘new’, not yet well defined, involved horizontal questions, and had not yet been addressed by the Court, making it likely that the directive would be challenged before the Courts.Footnote ⁶³ All of these factors obviously also highlight a research interest in some of the most fundamental questions: how legal actors make law. The General Court rejected these arguments, emphasising the need to evaluate sensitivity based on substantive content.Footnote ⁶⁴ The Court framed the choice of legal basis as ‘an essential question in the legislative process’ underlining how a ‘proposal is designed to be debated, in particular as regards the choice of legal basis’.Footnote ⁶⁵ The Court noted that it is precisely when ‘dealing with novel questions that the institutions request advice from their legal service’, therefore, ‘conferring particularly sensitive character on all legal advice concerning a novel question would result in impeding in practice the disclosure of a large proportion of that advice’.Footnote ⁶⁶ This jurisprudence has not in practice led to more legal data being proactively disclosed; instead, the institutions recycle their old arguments already defeated in Court when responding to new ATD requests.

The institutions often justify their refusals to share legal data with reference to potential legal challenges, as if these institutions would routinely fail to comply with the Treaties. This suggests that some of this data could be understood as ‘dirty data’ in the meaning of Gary Marx, as ‘information which is kept secret and whose revelation would be discrediting or costly in terms of various types of sanctioning’. While data can be ‘dirty’ in different ways, according to Marx, this kind of data

In the context that I study, ‘dirty data’ could be advice that has not been followed, and which therefore could be – if disclosed – compromising for the decision-maker, and result in demands for accountability. But data could also be ‘dirty’ specifically because it fails to fulfil the criteria of being ‘frank, objective and comprehensive’, and thus raise questions about the professional ethics or standards of the lawyers producing it, and whether they had given in to political pressure. Both of these are important research perspectives where public debate could significantly strengthen the EU’s democratic governance or lead to reforms in the EU institutions.

ATD requests often result in lengthy application procedures with highly uncertain outcomes. The applicant is placed in a position of arguing against the lawyers who produce those documents and wish to maintain their confidentiality.Footnote ⁶⁸ Writing confirmatory applications is cumbersome, as is keeping track of procedures and deadlines, which the institutions do not follow rigorously. While the Council remains more respectful of deadlines, the Commission employs systematic delays as a means of avoiding disclosure obligations.Footnote ⁶⁹ Moreover, the institutions often provide only partial access to their data, such as the few introductory paragraphs of opinions describing the legislative proposal. One example involves the recent debate among the Member States, NGOs, and legal academia concerning the EU’s possibility of leaving the Energy Charter Treaty, which led me to seek public access to the Commission’s legal opinion on the matter.Footnote ⁷⁰ The response arrived six months later, after all the vital decisions had been taken, and was of little use.Footnote ⁷¹ ‘Partial access’ was limited to the highly interesting title and equally interesting subtitles of the opinion while every sentence of the legal analysis was redacted, strategically preventing any public evaluation of this legal advice.

These difficulties of gaining access resemble other ATD contexts where the process of ‘data production’ provides an active role for both the researcher and the official in charge in shaping the outcome of the request.Footnote ⁷² The researcher needs to be ‘moving from observer to participant and back, stretching our capacities, using our emotions, working actively with those barriers that confront us – we may learn more from those very obstacles than if they had never presented themselves’.Footnote ⁷³ By using both active and passive research strategies barriers can not only be overcome, but can also be used as sources of important data.Footnote ⁷⁴ When debating disclosure with the institutions, the researcher also engages in co-creation of research material, as the correspondence is reflected in the institutional responses. Their decisions on non-disclosure typically explain the stage of decision-making, the substance of the requested documents, and their broader relevance for the development of the EU legal framework, including pointing out potential horizontal issues. This information also provides insight into the development of legal debate within the institutions, which would not otherwise be publicly available.

In legal opinions, silences can be powerful, and a significant characteristic of the research data. These include strategic silences in terms of matters that are not addressed by opinions, and choices between arguments that are used and not used. Such silences are often used to smooth the decision-making procedure, in particular to diminish the risk of democratic complications, especially at national level. One example of this is the EU–UK Trade and Cooperation Agreement.Footnote ⁷⁵ Agreements of this kind have generally been concluded as mixed agreements, requiring a separate national approval round. This time, the decision to sign the agreement as an EU-only agreement was taken between Boxing Day and New Year’s Eve, relying on a confidential but leaked CLS opinion, which openly admits that it ‘does not provide an in-depth examination of all of its aspects, nor does it provide a comprehensive and detailed competence analysis’.Footnote ⁷⁶ The opinion illustrates how the CLS understood its function as promoting the ‘Council’s interest’; in this case, saving the EU from a no-deal Brexit and enabling the conclusion of the EU–UK Trade and Cooperation Agreement (TCA) without additional procedural hurdles that a more serious engagement with the limits of EU competence would have required. In practice, the CLS empowered the national governments against their own parliaments by providing them with an authoritative analysis justifying the existence of ‘exceptional circumstances’, which are more political than legal.Footnote ⁷⁷ The Council lawyers define the EU’s economic and political interest and find legal arguments to promote it. This legal opinion demonstrates the deeply political role assumed by Council lawyers, and how the relationship between legal advice, democratic politics, and parliamentary scrutiny is far from straightforward.

Much of my recent research has circled around the legal and constitutional transformation around the EU’s COVID-19 response, NextGenerationEU (NGEU).Footnote ⁷⁸ Questions involving the use of these funds have also occupied investigative journalists, which has enabled ‘pooling resources, findings, and analyses’,Footnote ⁷⁹ and resulted in ‘qualitative inquiry that breaks down barriers between academics and other knowledge communities’.Footnote ⁸⁰ In addition to questions of financial accountability, I examined the history of the legal construction. Just before the approval of NGEU, the Council and the Commission – highly exceptionally, strategically, and proactively – disclosed a legal opinion and a Q&A document, which demonstrate a complete U-turn of earlier and well-documented legal interpretations by these two institutions. However, while providing legal justifications for why certain innovative solutions should not be seen as problematic, these opinions remained silent on several key aspects of the legal framework. An ATD request for the legal advice preceding the NGEU construction to both institutions in March 2023 resulted in the Commission’s disclosure of a few documents involving minor technical amendments to draft proposals from the final inter-service consultation preceding their formal approval, a discussion I return to below. However, the Council disclosed its advice in full after some initial hesitation.Footnote ⁸¹ These opinions provide ample ground for discourse analysis, since they effectively show how the difference between cohesion policy, on the one hand, and economic and related fiscal policies, on the other, as separate policy fields was diluted based on a highly selective reading of the case law. The opinions include effective silences on matters that would have spoken against the line of argumentation that the CLS opted for, certainly under considerable political pressure. The opinions also demonstrate how a new legal interpretation was first introduced and then gradually sedimentedFootnote ⁸² without public debate and only with limited analysis of risks and consequences.

ATD may require confrontational tactics to work around the official response tactics aiming at stonewalling requests.Footnote ⁸³ First-round rejections are the institutions’ standard way of testing whether the applicant is indeed determined, and need to be followed up by a confirmatory application that they hopefully engage more seriously with. If the institution insists on refusing access, the applicant can appeal to the General Court or the European Ombudsman. My first appeal to the EU Courts involved the European ParliamentFootnote ⁸⁴ and was related to the transparency of trilogues,Footnote ⁸⁵ with discussion evolving around the pending Court case de Capitani v. the European Parliament. I sought access to the administrative decision with which the European Parliament had denied access, to study its legal reasoning and its relationship to political pro-transparency rhetoric. But instead of formally disclosing the decision, the EP referred me to de Capitani’s personal website. The Court of Justice agreed with me that I had retained ‘a genuine interest in obtaining access to an authenticated version of the requested document … guaranteeing that that institution is the author and that the document expresses its official position’.Footnote ⁸⁶ As the Court passed the case back to the General Court for final ruling, overall, Court litigation in this case took five years, one month, and twenty-two days.Footnote ⁸⁷

My second court case originated in a research project on the legitimacy of the ongoing climate transformation and involved a Council legal opinion assessing whether the Commission proposal updating the EU’s Aarhus Regulation lives up to the requirements of the Aarhus Convention, in particular as regards the negative findings of the Committeee supervising compliance with it.Footnote ⁸⁸ The Council’s claim that public debate on EU compliance by the applicants – an environmental NGO and a legal academic – constituted ‘external pressure’ seemed particularly questionable, given that the Convention concerns access to information and public participation. The ‘external pressure’ had taken the form of a blogpost that I had published in the European Law Blog.Footnote ⁸⁹ The General Court annulled the Council decision repeating why openness in the legislative process matters:

Despite these judicial successes, the difficulties involved in engaging the Court are obvious. Even if a decision refusing access provides the applicant with standing to bring the case, the time frame is short. In practice, there is a need to appoint a legal representative;Footnote ⁹¹ there is a risk of costsFootnote ⁹² and the length of proceedings prevents using this route for any short-term research interests. Overall, however, using ATDs has enabled collecting unique data of legal sources that would otherwise remain unavailable to the public, including the data from the actual ATD processes that frequently underpins my research.

7.4 Open Access: Empty Archives

Collecting data via ATD from the Commission legal advice relating to the NGEU legal model also highlighted a deeper systemic issue with ATD requests: what documents do the institutional registers actually contain and how should the relationship between a register and public access be understood? The Regulation establishes a right to access to documents, defined broadly as ‘any content whatever its medium … concerning a matter relating to the policies, activities and decisions falling within the institution’s sphere of responsibility’. The European Ombudsman has consistently maintained that whether or not a document is registered in a document management system has no bearing on whether they fall within the scope of the public access rules.Footnote ⁹³

A request for Commission legal advice on the NGEU model led the Commission to maintain that its registers contained nothing beyond a few technical documents issued in the final interservice consultation. I triangulated this position with the formal role of the Legal Service under Article 21 of the Commission Rules of Procedure and a working paper published by a member of its Legal Service in 2019.Footnote ⁹⁴ In this paper, Leo Flynn – charged with the relevant files at the time – explains the engagement of the Legal Service with these very same questions over a period of several years. He describes how cohesion policy, primarily geared under Article 174 TFEU at ‘reducing disparities between the levels of development of the various regions and the backwardness of the least favoured regions’, had been firmly identified by the EU institutions as a way to fill the ‘gaps’ in the ‘incomplete policy side of EMU’. Treaty constraints on that side were considered ‘challenging’ and ‘limit the possibility for the Union institutions to adopt measures that are binding as to how the Member States conduct their economic policies’. It seemed highly unlikely that the Legal Service would have provided no written advice on this legal transformation, which took place over a period of time stretching over several years preceding the completion of the various Commission proposals in this area. But how to turn from speculation to evidence?

Many of the documents that were initially identified by the Commission as a response to my request were blanked out.Footnote ⁹⁵ However, a comment box in the margin of one disclosed documentFootnote ⁹⁶ revealed that more would have existed, as the ‘SJ referred to its full set of drafting suggestions as well as observations in writing that had been made ahead of the FT ISC meeting’ and that further suggestions had been elaborated in an ‘accompanying email’. Neither of these documents were listed as falling under my request. In my cofirmatory request I clarified that my interest related to the Commission analyses on the applicable legal bases, since

Since the Commission reply was delayed, I contacted the European Ombudsman.Footnote ⁹⁷ The Commission’s confirmatory decision arrived five months after the initial request was filed,Footnote ⁹⁸ and claimed that my argumentation presented a forbidden attempt to broaden the initial request. No further documents could be identified, and no new documents needed to be created. Finally, the Commission observed, if an institution claims not to hold any documents, there is a presumption of lawfulness attached to the declaration,Footnote ⁹⁹ unless the applicant can rebut it by relevant and consistent evidence.Footnote ¹⁰⁰ The Commission argued, the ‘mere suspicion that there must be a document does not suffice to put in question the presumption of legality of the institution’s statement’. Instead, its advice was to file a new request for ‘additional documents’. Unlike the Ombudsman has required in its practice, the Commission made no attempt to ‘engage with requesters openly and constructively at all stages, and ensure they are provided with all necessary information that may enable them to clarify their requests (for example by providing a list of documents that potentially fall within the scope of the request)’.Footnote ¹⁰¹ I consulted the Ombudsman who found a new request ‘the more appropriate course of action. We have asked the Commission to treat any such follow-up public access request promptly.’Footnote ¹⁰²

The new request specified that my requested access related to

The wording of the request was based on the Grand Chamber ruling in ClientEarth v. the Commission, where the Court not only recognised the Commission as a ‘key player in the legislative process’ but also underlined the need of transparency in the preparatory phase of legislative proposals, which ‘enables citizens to understand the options envisaged and the choices made by that institution and, thus, to be aware of the considerations underlying the legislative action of the European Union’.Footnote ¹⁰³ However, in response, the Commission reiterated that despite ‘a thorough search in their corporate document management systems’, the Legal Service and the competent DGs had failed to identify any further documents.Footnote ¹⁰⁴

Upon a confirmatory application, which the Commission again failed to consider in due time, I submitted another complaint to the Ombudsman in December 2023.Footnote ¹⁰⁵ The Ombudsman proposed broader access of documents that the Commission claimed did not exist. The internal discussions between the Commisison and the Obudsman were not disclosed, with the reasoning that ‘disclosure of this document is prevented by the need to protect the Ombudsman’s ongoing inquiry’.Footnote ¹⁰⁶ Overall, the Ombudsman seemed at a loss with the Commission’s refusal to answer my request.

Internal consultations within the Commission took altogether sixteen months from the initial access request.Footnote ¹⁰⁷ However, in June 2024 it identified twelve new and more informal documents from the relevant time period. While the Commission did disclose some of the queries from the DGs that its lawyers had been reacting to, it redacted all the actual legal advice, because it ‘concern[ed] purely internal exchanges related not even to any draft versions of future proposals (and therefore not part of [any] legislative file) and thus should be as a rule protected as part of the institution’s “space to think”’.Footnote ¹⁰⁸ When triangulated with publicly available information such as Commission work programmes, Euro Summit conclusions, and letters by the Commission President, this statement seems factually wrong: legislative drafting was definitely ongoing. A final appeal is pending with the European Ombudsman. While the Commission is unlikely to volunteer to disclose the information, the process reveals the institutional strategy: first, the claim that documents do not exist; then stall Ombudsman investigations; and finally play with time. This strategy does not become the ‘Guardian of Legality’, suggesting that it sees the data as something worthy of protection.

The queries from DGs disclosed by the Commission demonstrate that the Commission lawyers were working their ways around key Treaty constraints, which suggests that the data could indeed be ‘dirty’. But it is difficult to see what consequences this might have at this stage beyond the public image of the Commission and the professional reputation of its lawyers; both paid for by European taxpayers who also have a Treaty-based right to know the legal foundations of decisions that concern them. The Commission proposals concerned by the above requests either stalled in the legislative process or were replaced by other proposals that have already entered into force, thus making annulment cases a highly unlikely scenario.Footnote ¹⁰⁹

The Commission insistence that its registers contained no documents matching the search criteria raises the question of how the search was conducted and what the registers contain. The past couple of years have seen a great deal of discussion about investigative journalists’ attempts to gain access to the Commission President’s text messages relating to the so-called Pfizergate.Footnote ¹¹⁰. According to Vice-President Jourova, the Commission follows three registration criteria:

1. 1. The information relates to the policies, activities or decisions falling within the institution’s sphere of responsibility;

2. 2. The information concerned is important and not short-lived;

3. 3. The information concerned has been drawn up or received by the Commission.Footnote ¹¹¹

The same idea of ‘short-lived’ documents is also repeated in Council rules, to which I sought access for the purposes of this chapter. However, the Council rules also stress that the purpose of data retention is to ‘guarantee the medium- and long-term preservation and availability of documents and official files, in view of their administrative, legal or historical value’, but also to ‘enhance transparency and make it easier for the public to gain access to documents’.Footnote ¹¹² While the focus of the debate has been on text messages that can probably more seldom be understood as containing ‘legal data’, the question of what is registered and what happens to non-registered documents remains valid. Legal advice is given about matters that fall under the institution’s duties. Spiegel reports that

Six months is a short time for legal research, given that the Commission routinely takes over a year to respond to requests and strategically delays their handling. Since legal advice is often considered informal in the Commission, my ATD requests suggest that it is seldom registered. However, even such non-existing advice seems retrievable, presuming that the Commission can be convinced to engage in such an exercise. This highlights the Commission’s own discretion in identifying the relevant documents and possibly deleting ‘dirty data’, which may result in a new version of the ‘open access – empty archives’ argument. According to the classic version of this argument, disclosure of documents deters actors from writing their positions, which is feared to damage the quality of records and impoverish the content of future archives.Footnote ¹¹⁴ In the new version, it seems that the institutions may also be actively engaged in destroying existing records to avoid public scrutiny. This means that in the process of discovering dirty data, the researcher

7.5 Conclusions

It is in the interest of EU institutions but also EU citizens that the legal data that the EU institutions rely on is indeed ‘frank, objective and comprehensive’. Being able to anchor its actions into the Treaties has been significant for the EU’s authority over the years.

Legal debates in the institutions help to explain how law operates and how it is understood to frame the choices of policy-makers. While the institutional legal services routinely claim that ensuring the quality and objectivity of their advice demands that they remain confidential, the opposite might be true. When examining the doctrine that legal service opinions create, there is no doubt that the function of the EU is to move forward, and the job of the legal experts is to remove obstacles to progress. Given this paradigm, it is evident that even the most confidential advice sometimes fails to fulfil the criteria of being ‘frank, objective, and comphrehensive’. In this process of interpretation, democracy, transparency, or stronger citizen involvement are frequently sacrificed to pave the way for deeper integration, stronger control from the middle, and swift approval processes. It would be important for EU legal scholarship to make these interpretative choices visible and subject them to critical analysis. Access to data is key to any research, qualitative or quantitative, because it effectively determines what questions we can ask – and if we are in fact asking relevant questions and launching relevant critique of the practice. This also ensures that legal academia and legal practice engage in a constructive conversation. I believe that the institutions have a responsibility to contribute to this conversation.