Allies in Transparency ? Parliamentary , Judicial and Administrative Interplays in the EU ' s International Negotiations

International negotiations are an essential part of the European Union’s (EU) external affairs. A key aspect to negotiations is access to and sharing of information among the EU institutions involved as well as to the general public. Oversight of negotiations requires insight into the topics of negotiation, the positions taken and the strategies employed. Concurrently, however, some space for confidentiality is necessary for conducting the negotiations and defending EU interests without fully revealing the limit negotiating positions of the EU to the negotiating partner. Hence, attaining a balance between the necessities of oversight and confidentiality in negotiations is the subject of a dynamic debate between the EU institutions. This paper provides a joint analysis on EU oversight institutions’ position on transparency in international negotiations. We set out to answer whether parliamentary, judicial and administrative branches of oversight are allies in pursuing the objectives of transparency but also examine when their positions diverge.


Introduction
International negotiations are an essential part of the European Union's (EU) external affairs.Only in the past decade, the EU concluded over 140 agreements1 on diverse and highly salient issues including security, trade and climate change. 2Article 218 of the Treaty on the Functioning of the European Union (TFEU) sets out a "single procedure of general application" 3 for the negotiation of international agreements and provides the different roles in the negotiating process for the Commission, the Council and the European Parliament (hereinafter: EP).
A key aspect of negotiations is access to and sharing of information among the EU institutions involved as well as to the general public.Oversight of negotiations requires insight into the topics of negotiation, the positions taken and the strategies employed.Concurrently, however, some space for confidentiality is necessary to conduct the negotiations and defend EU interests without fully revealing the limit negotiating positions of the EU to the negotiating partner.Attaining a balance between the necessities of oversight and confidentiality in negotiations is therefore the subject of a dynamic debate among EU institutions.
Scholars have extensively discussed the role of the EP in the context of EU international agreements (Ripoll  Servent, 2014; Van den Putte, De Ville, & Orbie, 2015), including issues of the "democratisation" of external affairs (Meissner, 2016), and how its role in external relations impacts the EU's constitutional fabric (Cardwell, 2011; Eckes, 2014; Krauss, 2000).These studies in principle point to the executive and semi-executive institutions, the Commission and Council respectively, as actors with a preference for space for confidentiality rather than openness of negotiations.By contrast, the EP, for carrying out its oversight function, often features as the protagonist pushing for greater transparency (Curtin, 2013).
The Court of Justice of the European Union (hereinafter: CJEU) and the European Ombudsman (hereinafter: EO) also play crucial institutional roles in maintaining a balance between this space for confidentiality and requirements of transparency.For example, the CJEU safeguards the EP's right to information through the interpretation of Art 218(10) (Abazi, 2016; Peers, 2014)  and limits the space for confidentiality due to the exercise of the right to public access to information (Abazi  & Hillebrandt, 2015).The EO, in turn, tries to mediate in cases of administrative malpractice that include questions of access to information (see also Neuhold & Năstase, 2017, in this issue). 4et what we lack is a more integrated analysis of whether and how parliamentary, judicial and administrative oversight together constrain EU executive secrecy in international negotiations. 5This paper aims to fill this lacuna.Seemingly allies in transparency, we question whether parliamentary, judicial and administrative branches of oversight share similar views of transparency in international negotiations.Are these institutions allies in pursuing the objectives of transparency and/or when do their positions diverge?The notion of an alliance is used in this paper to imply an alignment of preferences and not a concerted joint action by independent institutions.
The paper analyses cases on public and institutional access to documents in the EU's international negotiations.Based on these cases, it derives institutional preferences on the support of access to documents and the degree of openness in publishing documents.The latter dimension is conceptualised as a scale, ranging from a situation where an institution provides public access to documents on the assumption that all documents should be made public, unless there is an overriding justifiable reason for some level of secrecy.The other end of the scale pertains to the institution arguing in favour of confidentiality of documents unless there is a convincing argument for public disclosure.The paper shows that most institutional preferences on the degree of openness fall between these extremes.
The paper is structured as follows: Section 2 presents the overall research design through which we derive the institutional preferences on transparency.Based on the consulted sources, Section 3 discusses the role of judicial, parliamentary and administrative oversight in furthering transparency in international negotiations.Section 4 builds upon this analysis in order to map the alignments among the institutions along their preferences.We offer conclusions in Section 5 by reflecting on the implications of our findings.

EU Institutional Preferences on (Limits to) Transparency in Foreign Policy
Transparency is a much-debated issue both in academic and public discourse.It is often treated as conceptually close to the notions of accountability and legitimacy, but we still lack a consensus on its specific meaning.Yet one aspect of transparency is straightforward: it is a precondition for conducting oversight, whether public or institutional.Without some transparency, how can one see what ought to be kept in check?(Bentham, 1839).This paper aims to contribute to the expanding discussion on the "accountability landscape" by examining the preferences of the institutions on the conditionality of oversight.Oversight is understood as a loose relation of checks, as opposed to an accountability arrangement that addresses a relation of principal and agent with elements of sanctions in the hands of the principal (Bovens,  2010; Brandsma & Adriaensen, 2017).
The specific contribution of this paper is to provide insights on the space where contestations on the necessary levels of transparency and secrecy between institutions are discussed.To this end, we study the positions of different EU (oversight) institutions on the limits of transparency in foreign policy.The purpose is not only to identify the institutions' positions but also to gauge whether these positions are shared or whether divergences can be observed.Preferences on secrecy and transparency are not derived in isolation but are continuously shaped and reshaped through inter-institutional interplays.Consequently, it is impossible to separate institutional preferences from the broader context in which these institutions operate.Indeed, the new-institutionalist literature has long acknowledged this conceptual struggle when searching to assess the "true preferences" of a political agent (Immergut, 1998).Hence, we first discuss the contextualisation of institutional preferences and then explain the methodology through which we seek to identify these preferences.

Contextualising Institutional Preferences
The EU oversight institutions studied in this article each play a significantly different role through their core functions.Their differences give rise to questions of whether one could expect them to have similar positions on trans-parency.In this paper, a position or preference for transparency implies that the institutions would support disclosure of the document.An institution may favour the disclosure of all documents requested, may favour disclosure of certain parts of a document (otherwise known as partial disclosure), or may argue in favour of secrecy.
Taking an inductive approach, we want to identify and map (potential) similarities and differences not through a predetermined theoretical framework but on the basis of what the institutions themselves have stated in cases involving public and institutional access to information.On the one hand, one may expect that a court could be in favour of disclosure of documents in light of defending principles of transparency and good governance, which the courts are tasked to safeguard.On the other hand, executive institutions would have an interest in maintaining some secrecy in the foreign policy arena in which they function and hence may have a preference for space of confidentiality.These differences in positions would not be fully unexpected in light of existing literature that focuses on transparency in foreign policy (Abazi  & Hillebrandt, 2015), and on the broader constitutional role of the institutions studied here.However, whether this is indeed the case in EU institutional practice is a matter that this paper will explore.
We do not seek to conceptualise institutional preferences as a simplistic divide between institutions favouring transparency and others that do not.Instead, our efforts are focused on the broad range of arguments provided by the institutions with the goal of improving our understanding of why institutions take their respective positions.The common thread among the oversight institutions we examine is that they deal with transparency but have a different institutional role towards transparency.The mandate of the EO pertains to maladministration and issues of public access to information.As its mandate is limited in terms of substance and scope (the institutions it can scrutinise), the EO has a track record of informally pushing the envelope of its reach.On the EO's interpretation of Regulation 1049/01, it has been noted that a "broad definition of 'document' has been adopted, matched by a narrow one as to the exemptions, thereby giving substance to the right of access" (Vogiatzis, 2018, p. 153).The Court has a role in ensuring that EU law is observed.It conducts judicial review, making sure that the EU institutions act in conformity with EU law.Hence, the role of the Court is to uphold the principle of transparency and ensure that if an exception is applied that takes place in accordance with the established limits between transparency and secrecy.The role of the EP as co-legislator is also complex.It has political significance for determining transparency levels not only through its own practice but also through its role as legislator.

Deriving Institutional Preferences
The objective of this paper is to understand the oversight institutions' preferences on transparency in the EU's international negotiations.To that end, we study the arguments they present in cases on public and parliamentary access to information, i.e. cases pertaining to Regulation 1049/01 and Article 218(10) TFEU respectively.The focus on public and parliamentary access to information as opposed to other definitions and conceptualisations of transparency (see e.g.De Fine Licht, Naurin, Esaiasson, & Gilljam, 2014; Pozen, 2010), follows from the legal regimes on which our data-gathering is based.Public access to documents enables public awareness and debate of the EU's negotiations, which are necessary for public trust in the negotiations and their potential successful outcome (Abazi, 2016).Institutional access by the EP is necessary for democratic oversight of EU negotiations (Curtin, 2014).Whilst we do not subsume these two types of access to documents under the same rationale since they serve different purposes, one of public accountability and the other of democratic oversight, both these types of access are a necessary condition for any accountability or oversight process to occur.Therefore, we find them both relevant when addressing transparency in the EU's international negotiations.
Regulation 1049/01 stipulates the regime for public access to documents. 6Article 4(1)(a) therein provides the exceptions to public access to documents with regard to international relations.This exception is considered of mandatory nature since the institution is obliged to first explain that the disclosure of the requested document could specifically and actually undermine the protected interest and, second, that the risk deriving from the disclosure is reasonably foreseeable and not purely hypothetical. 7If these two conditions are met, the institution is not obliged to examine whether there is an overriding public interest in disclosure, which is the case with the exceptions under Article 4(2) of Regulation 1049/01.
In the EU's international negotiations, the appointed negotiator is obliged to inform the EP "immediately and fully…at all stages of the procedure".8Art.218(4) TFEU stipulates the role of the Council during the negotiations and provides that-in addition to the possibility of adopting negotiating directives-the Council may "designate a special committee in consultation with which the negotiations must be conducted". 9The difference in wording as to the obligations of the negotiator toward the Council and the EP is already indicative of the different treatment of these institutions with regard to institutional access to information.
Preferences with respect to institutional and public access to documents are drawn from existing case law.To complement these measures, we derived a third indicator, which is more political in nature and tries to gauge the institutions' preference on the degree of transparency in international negotiations in more general terms (beyond the contested cases).Does an institution adhere to the idea that all international negotiating documents should in principle be accessible unless convincing evidence can be presented to suggest otherwise?Or should the exception of international negotiations apply to all documents related to international negotiations unless a compelling case can be made that their release to the legislative institutions or the public would not harm the negotiations?The proactive publication of documents (in the absence of an external demand for it) reflects an administrative practice informed by the acknowledgment that the logic of secrecy may not apply to certain documents.The preference on the degree of openness goes beyond such an administrative practice and appeals to the underlying rationale.This rationale is-in part-reflected in the preferences on institutional and public access as derived from case law but aims to transcend these cases by looking at the line of reasoning presented within and beyond the selected cases.
The oversight institutions' preferences cannot be studied independently from their mandated roles in the EU's constitutional fabric, as was mentioned above.Interviews as a method of data-gathering were hence deemed inadequate to gauge the position of a singular institution.As a representative, the EP covers a great plurality of voices and its role as a (co-)legislator implies its preferences cannot be fully understood without reference to the Council.Hence, it would be challenging to derive representative claims for the whole institution drawing on a series of personal interviews.The EO's involvement in matters of transparency is bound by the legislative act through which it has been created, and the EO's position in its transparency decisions is linked to case law (Neuhold & Năstase, 2017, in this issue).The Court is guided by the principles of impartiality and independence, yet at times is also seen as "activist", which may to some extent have effects on its jurisprudence.Taking this into account, it is unlikely that the CJEU or the OE would divulge a clear preference beyond stating the legislation they are to uphold.As our objective is to provide a comparative mapping of preferences across the oversight institutions, we therefore focus on revealed preferences in cases on institutional and public access to documents.This paper offers an analysis which applies to international agreements across EU policy fields, i.e. international agreements regarding trade or related to CFSP, tak-ing into account that accessibility to information is a procedural requirement applicable to all international agreements under Art 218 TFEU and of mandatory nature. 10he focus on international negotiations implies that the paper does not examine the lack of transparency in EU's legislative process, specifically the trilogue negotiations (see Leino, 2017, in this issue).More clearly than in other primary law reforms, the Lisbon Treaty draws a distinction between external diplomacy and internal legislation and the latter is conditioned by requirements of lawmaking that uphold a higher level of transparency and should not be treated as a diplomatic setting of negotiations. 11y focusing on contested access to document requests, one could argue that the agreements covered in our analysis are self-selected and therefore do not constitute a representative sample.Yet, we consider this less problematic for two reasons.First, the preferences expressed in such contested cases either reaffirm or tilt the institutions' view on the balance between transparency and secrecy towards the former.They shape expectations in other (future) negotiations.Second, the cases we study cover a broad range of the EU's external negotiations.We incorporate cases on trade (ACTA, TTIP), transatlantic security cooperation (TFTP) and CFSP (Tanzania, Mauritius).Therefore, the focus on international agreements in which transparency is a contested issue does not lead to a selection bias.

Public and Institutional Access in the EU's International Negotiations
In the EU's international negotiations, it is particularly useful to distinguish between the negotiating mandate on the basis of which the negotiations take place and other documents which emerge during the process of negotiations.We focus on these sets of documents separately since different principles govern access to each.

Negotiating Mandate
International negotiations are initiated on the basis of a mandate drafted by the Commission and adopted by the Council.Whilst informal contacts take place between the Commission and the negotiating partner prior to the Commission's drafting of a mandate (Gastinger, 2015;  Stein, 1988), this section focuses on the adoption of the negotiating mandate as the initiation of the negotiations.We elaborate whether and to what extent the mandate is disclosed through public and institutional access.
The mandate of negotiations is deemed to have a constitutional significance. 12The Court maintains a distinction between the specific content of the mandate relating to the substance of negotiations and the choice of legal basis regarding those negotiations. 13The latter does not form part of the substance of the negotiations and as such may be considered separately.Furthermore, the institutions do not have discretion to withhold the mandate merely because it pertains to international negotiations, but rather an assessment must be conducted in line with the exceptions under Article 4(1)(a) of Regulation 1049/01.
Regarding public access, recent case law has clarified the level of proof that the institutions must establish in order to defend secrecy in international negotiations (Abazi & Hillebrandt, 2015).Namely, the institution must first establish that the disclosure of the requested document could specifically and actually undermine the protected interest and second, that the risk deriving from the disclosure is reasonably foreseeable and not purely hypothetical.This test has long been part of judicial review regarding the exceptions in Regulation 1049/01.Yet, since the case of Council v. Sophie In't Veld it is clearly applicable in the context of international relations.
The originator of the information has the authority to decide whether the document will be disclosed following the exceptions stipulated in Regulation 1049/01.With regard to the mandate, the fact that it is the Council that gives the mandate to the Commission for the negotiations is sometimes lost in the public debates.In the case of TTIP, the Commission received criticism that it lacked transparency and refused to share this document publicly.14However, it falls under the authority of the Council, which issues the mandate, to publicly release it.In the case of TTIP, however, the mandate was leaked.On the one hand, the EO targeted the Council with an own inquiry to demand the release of the negotiating mandate for TTIP. 15The EO argued that the leaked document shows that there are no clear reasons why a publication would jeopardise the public interest.The EO argued that the release would not invalidate the applicability of Article 4 of Regulation 1049/2001 to negotiating mandates at large. 16It would instead add credence to the view that the Council is concerned with transparency and seeks to strike a balance between efficiency and transparency.On the other hand, the Council in its opinion accentuated that it is not obliged to reveal the mandate as it concerns a non-legislative document and thus there is no instance of maladministration.17Throughout the exchange of letters both the Council and the EO were conscious of the risk of setting a precedent for future agreements. 18nder growing public pressure, the Commission has advocated the public release of the mandates by the Council.Former commissioner Karel De Gucht pushed for the release of the TTIP negotiating mandate on multiple occasions, among others in a plenary debate in the EP. 19s part of the new trade strategy's emphasis on transparency, Commissioner Malmström indicated she would "[invite] the Council to publish all negotiating mandates immediately". 20ltimately the Council released the mandate.However, more general conclusions cannot be drawn that the Council would continue to make the mandate public for other international agreements.Since the publication of the TTIP negotiating mandate, the Council has been more cooperative in releasing similar documents in highly politicised negotiations.Still, in most cases the publication took place long after the launch of the negotiations (cf.TTIP, CETA and TiSA) or through a partial release omitting critical sections of the document (e.g.EU-China investment agreement).The Commission also requested the release of the mandates for the free trade negotiations with Japan, Mexico and Tunisia in September 2016. 21Commissioner Malmström reiterated her request in May 2017 a few weeks before Greenpeace leaked documents related to the EU Japan free trade negotiations. 22The Council discussed the issue during a meeting of the Trade Policy Committee but could not obtain the common accord required for their public release.In sum, the Council has thus far only shown a willingness to publicly release negotiating mandates at an advanced stage in the negotiations, amidst intense pressure from civil society and the European institutions, and when the lack of transparency may jeopardise the negotiations.Hence, while they provide some tentative support for the public access to the mandate (scored low → medium), their preference regarding the degree of openness remains low as summarized in Table 1.
With regard to the position of the EP on public access to information, a stream of cases reveal that the EP has not intervened in support of the party requesting public While the EP is not directly involved in the drafting and adoption of the mandate, in line with Art 218(10) TFEU it is supposed to be informed at all stages of the procedure.Hence, it should receive the mandate of negotiations.The Lisbon Treaty provided increased prerogatives to the EP in international negotiations by granting veto powers.Practice suggests that the EP uses its increased powers for access to information in order to affirm its institutional role.For example, on the Terrorist Finance Tracking Programme Agreement with the US, the EP first vetoed the Agreement by raising concerns on data protection safeguards to only then give its consent at a later stage although there were "no remarkable differences between the first and second agreements" (Vara, 2013, p. 20).Rather, the difference was that in the second round of negotiations, the EP was fully informed at all stages of the negotiations.This has raised questions of whether the EP's position is too focused on inter-institutional power dynamics (Eckes, 2014).A similar change of position after having received more information is also notable with regard to the Passenger Name Record Agreement with the US (Ripoll Servent &  McKenzie, 2011).
In sum, the Court and the EO favour public access to the mandate.Disclosure does not necessarily apply to the entire document.The Court draws a distinction between the legal basis, which should be disclosed, and other substantive parts of the mandate that may remain confidential.The Court does not accept an argument in favour of secrecy by default for the mandate simply because this document pertains to international relations.Rather, the Council and the Commission must substantiate their reasoning for non-disclosure in light of require-ments established by the Court.While the EP is generally supportive of access to information, for the mandate specifically we lack sufficient data to draw solid conclusions as to whether this support is high.Yet, when it comes to institutional access, the EP has made significant efforts to ensure that its prerogative to access information is respected and has also utilised judicial recourse towards such aims.
The Court is also in favour of an informed EP throughout all stages of international negotiations, as is evident from a few recent salient cases regarding institutional access in line with Art 218(10).Taking into account that the EO is generally aligned with the case law of the CJEU as far as transparency is concerned (Abazi & Tauschinsky,  2015), it could be noted that the EO favours institutional access to the mandate.The position of the Commission is less clear.Although in the case of TTIP the Commission argued in favour of opening up the mandate by the Council, whether this continues as the new practice and position of the Commission remains to be seen.The recent reminder to publish the mandates for the negotiations with Japan, Mexico and Tunisia supports this view.Still, the litmus proof consists of continuing this practice once politicisation subsides.

Negotiating Process
Public access to information is limited during the negotiating process.The Council and the Commission continuously defend secrecy as necessary for the negotiating process, emphasising the need for trust between the negotiating partners (Abazi & Hillebrandt, 2015).Interestingly, the EP posits similar arguments in favour of nondisclosure when citizens file public access requests.For example, the EP refused access to documents regarding the ACTA arguing in favour of secrecy not only for ongoing negotiations but also for future negotiations, stating that: There is a concrete risk that disclosure of preparatory documents would prejudice not only relations with third countries in the context of ACTA, but also any other negotiation to be conducted by the EU in the future.Indeed, any future negotiating partner of the EU could doubt the EU's reliability with regards to the confidentiality of negotiations, if preparatory documents concerning the position of one of the EU's contracting partners were released to the public. 24e EO, although generally critical of the lack of transparency, shares the view that some level of secrecy is justified in international negotiations.The EO reaffirmed this position when handling an access to documents request vis-à-vis the EP in the context of the ACTA negotiations. 25Similarly, regarding the Council, the EO stated that: Releasing the documents in question, [would] reveal the negotiating position of the US and Japan, [and]  would be highly likely to be detrimental to the EU's relations with those countries.The EO also agrees that, as further argued by the Council, it is likely that such disclosure would have a negative effect on the climate of confidence in the on-going negotiations, and that it would hamper open and constructive co-operation. 26vertheless, the EO seems to pursue a balance between the opposing needs of transparency and secrecy in international negotiations.The EO opened an owninitiative inquiry levied against the Commission for access to document requests in relation to TTIP and the existence of a potential bias in the Commission's disclosure of negotiating documents to a limited group of "privileged stakeholders". 27The EO provided five recommendations to increase transparency, including not only the hosting of all disclosed documents on their website, the creation of a public register and the publication of all meetings with civil society, but also the reinforcement of measures to ensure confidential documents stay confidential.In much of its communication, the EO also men-tioned TTIP as a specific case and was cautious in drawing wide-ranging conclusions on other cases.
The Commission has taken a more proactive approach towards publishing negotiating positions, not only with respect to TTIP but also other ongoing negotiations (Coremans, 2017, in this issue).Yet whether these efforts will be maintained once politicisation subsides can be questioned (Gheyle & De Ville, 2017, in this issue).
Regarding institutional access, practice shows that the Council delays significantly in meeting its obligation to inform the EP.For example, in the case of European Parliament v. Council,28 the Council shared the document for the EU-Mauritius Agreement more than three months after the adoption of that decision and the signing of that agreement. 29In this case the CJEU held that the information requirement laid down in Article 218(10) TFEU applies to the entire process of international negotiations and significantly, that this procedural requirement also applies to agreements falling exclusively under CFSP. 30Therefore, informing the EP is a mandatory procedural requirement within the meaning of the second paragraph of Article 263 TFEU and its infringement leads to the nullity of the measure. 31Similarly, in the case of the EU-Tanzania agreement, the Court held that even a delay of 9 days implies that the EP was not informed "immediately" and hence the Council had failed to comply with its obligations to share information in all stages of the procedure fully and immediately. 32he different positions are summarised in Table 2 below.The data confirms the differences in positions on institutional access to documents between the Council and Commission, as (quasi-)executive bodies on the one hand, and the institutions responsible for legislative, judicial and administrative oversight, on the other hand.When we take a closer look at support for public access and the institution's preference on the degree of open- ness, the picture becomes more complex.The Commission in particular has undergone a notable shift in its stance.In the next section, we explain these preferences by focusing on the impact transparency may have on the institutional balance.

Alliances in Transparency?
This section offers a three-step analysis of the oversight institutions based on their institutional preferences on transparency as derived above.First, we will discuss preference alignment among the oversight institutions as well as the executives.Second, we turn our attention to what we classify as "unexpected alliances", i.e. cases of preference alignment between the executive institutions and institutions of oversight.Finally, we provide broader reflections on the contextual factors that help explain these alliances, thereby presenting an insight into how the EU's transparency regime may develop in the future.

Oversight Alliances
EU oversight institutions differ on their preferences on the three dimensions of transparency we have analysed: public access, institutional access and degrees of openness.In general, the EO, the Court and the EP are well aligned as far as preferences on institutional access are concerned.This applies both to the mandate as well as the release of documents during the negotiating process.Especially for the EP, institutional access is core to its function of democratic scrutiny, a view supported by the EO and CJEU.For example, in the consultations for TTIP, the EO stated that the MEPs have a special democratic responsibility to scrutinise the negotiations on behalf of their constituents. 33Indeed, this is in line with the Treaty of Lisbon, which has cemented the EP's role as (co)legislator in the EU's constitutional structure, and the interpretation of Art 218(10) further supports this recognition.
When we shift our focus to public access to information, two patterns emerge.First, all oversight institutions support the public disclosure of the negotiating mandate.The case of Council v. In't Veld helped to clarify the distinctions in the document between the constitutional elements, which should be disclosed, and the substantive parts of the document, where the negotiating position may be revealed, and hence there is justification to maintain some confidentiality.Moreover, the explicit request to provide evidence that public access to the mandate would jeopardise the negotiations further raises the barrier to maintain secrecy.The EO's own-initiative inquiry with regard to the TTIP negotiating mandate builds upon this case law.The EO further added to this argument the democratic need for the Council to acknowledge the relevance of transparency in international negotiations.Second, all oversight institutions also accept-to varying degrees-the reasoning that external negotiations war-rant a certain degree of secrecy to ensure that the EU's bargaining position is not compromised.The exceptions introduced by Regulation 1049/01 indicated clearly that the legislators accepted the limitations to public access to information in external negotiations.And as the Court and the EO operate within the confines of the EU's legal order, they follow the transparency regime laid down by the legislators.
Finally, with respect to the degree of transparency, significant variations could be found.The EO, in particular in recent practices, has taken a more proactive stance towards the publication of documents.This is not followed by the EP, which has accepted a higher degree of secrecy for negotiations, even showing concern for future agreements of the EU as we saw above with the ACTA documents and thereby potentially creating a political space for confidentiality of negotiations in the EU.By contrast, the CJEU bases its judgment on the disclosure of documents by requiring the institutions to show an actual (as opposed to a hypothetical) risk from disclosure.
We expect to find the (quasi-)executive bodies opposing the oversight institutions, as they hold information that may be disclosed to the interest of the public or its representatives.Our findings suggest that the Council and the Commission, do align preferences to a certain extent, except when cases concern public access to the mandate and the degree of openness to be pursued, as will be explained in the following section.

Unexpected Allies in Transparency?
The alliance among oversight institutions, despite their differences, may be to some extent expected.However, it is less common to see views shared by oversight and executive institutions which often find each other on opposite sides on transparency issues.Two such "unexpected" alliances emerge from our analysis regarding public access to documents.We consider these 'unexpected' as explained in line with the institutional position and role of the oversight institutions and executive actors.Generally, the literature suggests that oversight institutions are in favour of transparency and access to information, whereas due to their negotiating position in the international negotiations arena the executive institutions rely on arguments of trust and confidentiality.
In the first instance, the Commission aligned with the oversight institutions in favour of public disclosure of the mandate.It is a Council document that to a certain extent dictates the Commission's actions.As such, the mandate's release allows the Commission to deflect some of the public criticism back to the Council.Moreover, requests for the mandate's release have often arisen when the negotiations were already under way and when its contemporary strategic value had been diminished.For example, in the case of TTIP, the mandate had already been leaked which made it clear that public disclosure would not jeopardise the bargaining outcomes of the on-going negotiations.Another factor that may explain this unexpected alliance is the expectation that, due to a high degree of politicisation, the risks associated with the publication of the mandate were lower than the risk of nondisclosure.Accusations of secrecy and the public's mobilisation around it could jeopardise the entire negotiations.
A second instance where an unexpected alliance emerged was in the context of the public release of negotiating documents.Similarly to the Commission and the Council, the EP supported keeping documents secret in the case of ACTA.The EP's support becomes even more evident in its response to the EO's own-initiative inquiry into the transparency of the trilogue meetings pertaining to internal negotiations.The former President of the EP, Martin Schultz, indicated the challenge to "find the right balance between ensuring transparency to the public, while at the same time ensuring that all political groups can fully follow and influence the negotiations" (European Parliament, 2015a).It is here that one of the key differences between the EP and the EO can be identified in their role as guardians of transparency.The legislative role of the EP puts it in an internal negotiating context where it also faces a "limit position", in contrast to the EO who remains largely outside of the legislative process.Furthermore, an MEP can be held to account by an electorate that is affected by their actions.This implies the MEP may find themselves in a position where they wish to be shielded from public scrutiny.

Institutional Politics of Transparency
Transparency is not merely an instrument of democratic trust and participation.Indeed, in an institutionalised setting it also becomes a tool of power dynamics in evolving constitutional structures.In our analysis three aspects of the politicised side of transparency emerge: the consolidation of the EP's legislative powers, the difficult balance between the Council's legislative and executive roles, and the evolving interpretation of the exception to public access to documents with regards to international relations.
Throughout the EU's history, the EP has always pushed for greater powers, and institutional access to information was an important pre-condition to perform these legislative functions.With Lisbon, much has been rectified and the powers granted to the EP represent a significant change.Institutional access to information featured both as an objective (cfr.SWIFT, ACTA or TTIP) but also enabled the EP to gain a similar status to the Council even in those areas where it has no legislative prerogatives, as shown in the Mauritius case.Yet, with the empowerment of the EP we have also seen a decline in its support for public access to documents.For example, following its early rejection of the TFTP agreement, the EP increasingly recognised that as colegislator, it shared the responsibility to consider member states' security concerns.The result was a softened stance on data protection (Ripoll Servent & MacKenzie, 2011).Being exposed to the (external) negotiating context generates a greater sense of responsibility for or complicity in the agreed outcome and thus-much like the Council or the Commission-the EP must assess whether disclosure of the requested documents could undermine the protected (public) interest (Abazi, 2016).When the EP still found itself on the side-lines of EU decision-making, it was easy to take more ambitious positions on transparency (or policy-objectives) without much ramifications.As a formal co-legislator, this has clearly changed.
A second observation pertains to the increasing pressures exerted on the Council to become more transparent.Edgar Grande's paradox of weakness explained how member states' delegation of powers to the EU strengthened their autonomy vis-à-vis organised interests within their constituencies (Grande, 1996).The multi-level decision-making context led by a bureaucratic Commission provided an ideal scapegoat to advocate national policy reforms that were too difficult to sell publicly at home.A key condition for this mechanism to work was a certain degree of secrecy in Council decisionmaking.In short, from its inception, the Council has used secrecy both in domestic as well as international negotiations to its own advantage.Therefore, it is evident that this institution is more conservative when it comes to transparency.This secrecy paradigm is currently pressured from multiple angles.Political pressures arise from both the EP as well as the Commission, with each-for their own reasons-trying to make the Council acknowledge its legislative role and calling for a logic of (greater) transparency (Hillebrand, 2017, in this issue).
In asserting its powers as co-legislator and demanding institutional access to documents, the EP's engagement with the Council is often predicated by a desire to be treated on equal footing.For example, the EP has increasingly pushed for its own access to information when it concerns the release of the negotiating mandate, but also calls on the Council to play a larger public role (European Parliament, 2015b).Reluctance from the Council to comply with such demands is not limited to the insurance of autonomy from their constituents or the national parliaments, but also the potential shift this can create in the balance of power to their detriment.Similar to the EP, the Commission is also concerned that the silence and secrecy within the Council is becoming detrimental to the European project.The EO also corroborated such views in a recent interview: What I've been trying to say to the Council, and to EU institution leaders as well: If you want to break through the myths, if you want to break through the caricature, then you have to allow people to see how laws are actually made, and how power is actually dis-tributed…between the EU institutions and the member states. 34 final reflection concerns the evolving interpretation of what is considered an acceptable level of openness.Once it has been shown that the public release of specific documents has not undermined negotiations, the bar is raised as to the evidence required to keep a document secret.The EO and CJEU recognise and uphold the exception to public access to documents when it may jeopardise international negotiations.With the EO defending an interest that stands outside the internal (and external) negotiations and the Court mostly seeking to assert its jurisdiction, their interests do not collide.On the contrary, the initiatives of the EO can lower the threshold for legal contestation.The case on the disclosure of the Council's TTIP negotiating mandate may be a case in point, as opposed to the release of the mandate in the TFTP that came about as a result of a Court judgment.When documents are actually already public, such as through leaks, both the EO and the Court maintain that there are no firm grounds to defend nondisclosure of documents.

Conclusions: Towards a Steady State of Oversight Interplays?
This paper questioned the general assumption in the literature that EU oversight institutions are in favour of transparency in international negotiations and that traditionally EU executive institutions prefer some level of secrecy in the international arena.In doing so, the paper sought to analyse the positions of the EU oversight institutions towards transparency in a more holistic manner and examine their interplays in delivering transparency.Through this joint analysis, we identified whether parliamentary, judicial and administrative branches of oversight are allies in pursuing the objectives of transparency but also when their positions diverge.
The role of each oversight institutions is of course different in the EU context.It may even be questioned why they should be analysed together.Indeed, the position of the Court as the legal authority to rule on issues of transparency-and thereby create conditionality of how transparency is practiced-may not be fully comparable with the institutional preferences of the EP that rather creates political dynamics on how far transparency expands in international negotiations.But the paper took these functional differences into account and questioned institutional preferences in light of such institutional variations.One such difference is the position between the Court and the EO.Namely, whereas the Court's position on transparency is evident and bound in its role to interpret transparency rules, the EO has often moved beyond its statutory roles in interpreting the scope for public transparency.
This paper showed that while there is general support from oversight and executive institutions to the idea that international negotiations warrant secrecy, variations across the institutions emerge for the degree of openness.An alliance was found among the legislative, administrative and judicial institutions as far as institu-tional access to documents is concerned.Yet, the image is less clear with regard to public access and the degree of openness.Moreover, our analysis reveals that some 'unexpected' alliances have emerged particularly when the negotiations were politicised.
The analysis provided in this paper contributes towards an understanding of the dynamics among oversight institutions in the field of transparency.Linking this knowledge with the institutional preferences we have derived, it becomes possible to provide some foresights on how debates on transparency will develop in the future.Foreseeable alliances most likely would feature the EO and the CJEU furthering or changing the contours of confidentiality in international negotiations.However, these changes do not impact the steadily created core to institutional and public access.Numerous cases make it clear that access to documents for the EP in international negotiations is mandatory throughout the process of negotiations.It is only a question of changes in institutional habits of the Council and the Commission to meet this legal requirement.Yet public access may prove more contentious as there is an increased public demand for transparency but it is met with a solid preference by both oversight and executive institutions for confidentiality, especially during the negotiating process.Looking forward, it seems more likely that "the degree of openness" will become the main bone of contention in the EU's international negotiations.

Table 1 .
Mapping institutional preferences on transparency of the negotiating mandate.Hence, we do not have data from case law on what precisely the position of the EP would be with regard to the mandate; yet, it can be added that the EP does not get actively involved in furthering public access to information and does not draw sufficient attention to this issue in its Resolutions.The latter instrument is a particularly significant tool for the EP to influence the margins of disclosure for the mandate, although resolutions are not legally binding to the Council.

Table 2 .
Mapping institutional preferences on transparency during negotiations.