The role of the European Parliament in international negotiations after Lisbon

ABSTRACT

After the entry into force of the Treaty of Lisbon, the European Parliament (EP) can give or withdraw its consent to most international agreements. This new scenario calls for theoretical models that help us to better understand the roles and strategies of European Union (EU) institutions in international negotiations. Departing from ‘two-level-games’ and principal–agent models, this article examines three interrelated elements (levels of negotiations; decision-making stages; and strategies) to explain the first international negotiations under the consent procedure (SWIFT Agreement). This case shows how the EP made use of day-to-day decision-making to informally expand its formal veto powers. The EP is now capable of controlling the EU negotiator during both the agenda-setting and the negotiation stages. Its informal involvement is set to transform its relationship with the Council and their collective capacity to influence and control the Commission.

KEY WORDS:

• Consent procedure
• European Parliament
• international agreements
• principal–agent
• SWIFT
• two-level games

INTRODUCTION

When, on 11 February 2010, the European Parliament (EP) decided to vote against the ratification of an agreement (SWIFT Agreement¹) between the European Union (EU) and the United States (US) on bank data transfers (European Union 2010), the reaction was one of surprise. Previous EU–US agreements on internal security needed only the signature and ratification of the Council. The long-term co-operation between the two sides and the absence of rejections under the assent procedure (renamed ‘consent’ in the Treaty of Lisbon) had led member states to consider the EP's ratification as a formality.

The rejection of the agreement by the EP is of particular importance, since the entry into force of the new treaty in December 2009 enlarged the use of the consent procedure to all areas in which the EP has the power to co-decide under the ordinary legislative procedure (Article 218 Treaty on the Functioning of the European Union [TFEU]). In practice, it means that the EP gained the capacity to ratify international agreements in key areas such as trade, agriculture and internal security matters – areas from which it had previously been excluded. In view of this new institutional setting, this article examines the role of EU institutions in international negotiations using principal–agent models and Putnam's ‘two-level game’ metaphor. Although these theoretical frameworks have been previously adapted to the EU's sui-generis system of governance in international negotiations, the role of the EP has been absent from such studies (Collinson 1999; Delreux 2008; Dür and Zimmermann 2007; Kerremans 2004; Larsén 2007a; Meunier 2000; Patterson 1997). This article proposes a model adjusted to the way the EU negotiates international agreements after the Treaty of Lisbon – focusing on the role of the EP and its relationship with both domestic (EU) and international actors. Looking at actors' formal and informal roles and strategies, the model aims to strengthen the study of EU international agreements and Putnam's two-level game metaphor by exploring how domestic (in this case, EU) political institutions can ‘provide different actors arenas and power resources and with rules’ (da Conceição-Heldt 2013: 590). Therefore, it focuses on the shifting roles of EU institutions and their strategies to influence the different stages of international negotiations (agenda-setting, negotiation, and ratification) both at the EU and international level.

Despite its particularities, the SWIFT Agreement allows us to explore these shifting dynamics in a context of institutional uncertainty. The objective of the case study is, therefore to contribute to the theoretical understanding of EU international negotiations and open new research questions. This article uses qualitative data² to formulate a structured comparison of the major players' strategies. Twelve semi-structured élite interviews were held in Brussels between March 2010 and April 2011 with EU officials, Members of the European Parliament (MEPs), diplomatic sources, and other participants such as political advisors who had been direct participants in or witness to the negotiations. In order to capture the complexity of the institutional setting after Lisbon, the first section reframes previous models based on principal–agent and ‘two-level game’ metaphors. It discusses the different levels and decision-making stages of negotiations before and after the new treaty rules, as well as the strategies adopted by EU and international actors during negotiations. The second section concentrates on the negotiations and ratifications of the SWIFT Agreement. It divides the empirical analysis into two stages, the first one looking at the negotiations for an interim agreement – which ended with the failed ratification of February 2010; the second stage analyses the subsequent negotiations for a permanent agreement, successfully ratified in July 2010. The two rounds of negotiations allow for a structured comparison, which allow us to draw wider conclusions on the shifting relationships between EU institutions and with the US. The findings show a clear episode of interstitial change (Farrell and Héritier 2003, 2007; Héritier 2007; Hix 2002), where the EP made use of a shift in the day-to-day decision-making (cf. Pollack 1999) to expand its participation beyond its formal power to ratify international agreements.

EU INTERNATIONAL AGREEMENTS PRE- AND POST-LISBON: A FRAMEWORK OF ANALYSIS

In order to understand the practices of EU institutions in the framework of EU international agreements, it is necessary to specify the level of negotiations; the decision-making stages in which they operate, as well as the strategies used by the different institutional actors to maximize their influence. These different elements are not stable over time and, therefore, need to be mapped before and after the entry into force of the Treaty of Lisbon.

Levels of negotiation: a two- or three-level game?

Given the EU's complex system of governance, most studies used principal–agent models to examine the act of delegation in international negotiations (e.g., Hawkins et al. 2006; Pollack 1997); they attempted to explain why member states allowed the Commission to negotiate in their name and to evaluate how much freedom the Commission enjoyed while acting as an agent (e.g., Damro 2007; Dür 2006; Elsig 2007; Kerremans 2004). Before Lisbon, most international agreements fell outside the scope of the consent procedure (introduced in the Single European Act); consequently, the Commission would generally represent the EU at the international level and the Council would be in charge of drafting the mandate, as well as signing and ratifying the agreement. The Commission was conceptualized as a single agent acting as a unitary actor (except for Larsén [2007a]), while the Council was ultimately depicted as a collective principal regrouping multiple interests (Elsig 2007; Kerremans 2004).

This constellation only held for those international agreements based on first-pillar policies. In the case of third-pillar matters (i.e., Justice and Home Affairs), the principal was still a collective actor formed of member states, but it was the Presidency that acted as the EU agent. As Delreux (2008: 1074) noted, the appointment of the Presidency as EU negotiator created distinct dynamics, since the Presidency acted both as one of the principals inside the Council and as an agent of the collective group of principals. This particular constellation affected the flow of information and the capacity of the agent to act independently.

Despite the differences between the principal–agent constellations in these two pillars, the negotiation levels were modelled in a similar manner. The presence of member states as a sole collective principal led to the development of ‘three-level game’ models, in which level I was the international arena, level II the EU arena, and level III the domestic arena (e.g., Collinson 1999; Meunier 2000; Patterson 1997). However, the extension of the consent procedure requires a re-adjustment of the levels; the formal steps of the consent procedure fit better into a ‘two-level game’ model. Here, level I (international level) represents the international arena in which the EU is represented by its negotiator (either the Commission or the High Representative for Common Foreign and Security Policy matters). Level II (domestic level) is formally composed of the Council and the EP, since they are both in charge of ratifying the international agreement. Therefore, three-level games would be only appropriate in cases of mixed agreements, since these, too, have to be ratified by member states (Delreux 2008; Meunier and Nicolaïdis 1999).

Decision-making stages

Decision-making can be divided into three stages: agenda-setting; negotiation; and ratification. These three stages act as heuristic tools to analyse the arenas in which actors exert their influence. Before Lisbon, the Council set the agenda by drafting a negotiation mandate (agenda-setting stage), which established the path for the EU negotiator at level I (negotiation stage). The Council was also in charge of signing and ratifying the agreement (ratification stage) – except in cases of mixed agreements (see above).

The control over all decision-making stages that member states enjoyed before Lisbon led most studies to concentrate on the shadow that ratification cast on the agenda-setting and negotiation stages (Bièvre and Dür 2005; Dür 2006: 366). Larsén (2007b: 9–10), for instance, noted that ratification was usually ‘a matter of just “voting up or down” … and does not expose any tensions in the relationship between the principals and the agent’. Therefore, since member states could use the threat of non-ratification to influence outcomes in earlier stages, the conclusion of international agreements in itself was not a focus of previous analyses. At the same time, ratification was also not an issue in those cases where the consent procedure applied (e.g., enlargement or major constitutional reforms), since they were generally valence issues with a high symbolical impact for which the EP was reluctant to withdraw its consent (Judge and Earnshaw 2008: 235–6; Wessels and Diedrichs 1997: 8). However, with the changes introduced by the Treaty of Lisbon, it is necessary to give the ratification stage its due attention.

The extension of the consent procedure in the Treaty of Lisbon to all those international agreements for which the legal basis foresees the use of codecision (Article 218 TFEU) widens the chances of the EP to have a say in international agreements. In this new scenario, the Council is still involved in each stage of the procedure by formulating the mandate (agenda-setting stage), overseeing the Commission's activities by use of a special committee foreseen in Article 218.4 (negotiation stage), and signing and concluding the international agreement (ratification stage). In comparison, despite the right to be informed at all stages of the procedure (Article 218.10), the EP is only involved in the ratification stage. However, the capacity of the EP to ratify international agreements cast a shadow over the entire procedure, which makes it all the more important to understand the strategies of the EP if one wishes to explain the success or failure of international agreements.

Actors’ strategies

The levels and decision-making stages of the procedure are clearly important to determine the potential strategies of actors in ‘two-level games’, which have been summarized in Figure 1.

Figure 1 Strategies of a ‘two-level’ game

Source: Adapted from Moravcsik (1993: 32).

As Figure 1 shows, Putnam puts more emphasis on the strategies used by negotiators (agents) at level I. Therefore, principal–agent explanations are used to better understand the strategies of actors at level II – particularly in their attempts to influence and control developments at the international level.

At level I, negotiators choose different strategies depending on how their preferences fit with those of the domestic level and the opponent (Moravcsik 1993: 31; Putnam 1988: 460). If level I and level II actors share the same preferences, the negotiator can concentrate its efforts on obtaining more concessions from the opponent. This may be achieved either by tying its own hands – claiming it has little room for manoeuvre at the domestic level (Moravcsik 1993: 28) – or by restructuring the opponent's preferences – offering side-payments, raising the cost of no agreement, or persuading domestic actors to change their preferences (Putnam 1988). In contrast, when the negotiator is closer to the interests of the international opponent (‘dove’) or when it departs from both the domestic and the opponent's interests (‘hawk’), negotiators have high incentives to escape their mandate. One option is to use a strategy of mutual support (collusion) at level I in order to reach the negotiators' preferred outcome, even if it enhances the risk of failed ratification (Moravcsik 1993: 26). The second option is to ‘cut slack’ in order to widen the preferences of domestic actors – offering side-payments, creating synergistic issue-linkages (using the international agreement to create a policy option that was beyond domestic reach), or presenting more opportunities for domestic actors to participate in and control negotiations (Moravcsik 1993; Putnam 1988).

At level II, domestic actors can also attempt to tie the negotiator's hands (i.e., influence the developments of international negotiations) by using various control mechanisms.

Table 1 Formal control mechanisms before and after Lisbon

	Ex ante (agenda-setting stage)	Ad locum (negotiation stage)	Ex post (ratification stage)
Pre-Lisbon(first pillar)	CouncilMandate	CouncilCouncil directives/special committee	CouncilSignature and ratification
Pre-Lisbon	Council	Council	Council
(third pillar)	Mandate	Presidency's double role	Signature and ratification
Post-Lisbon	CouncilMandate	CouncilCouncil directives/special committee	CouncilSignature and ratification	EP Ratification

During the agenda-setting stage, the Council has the capacity to control the agent ex ante by confirming the authorization of the EU negotiator (Delreux and Kerremans 2010: 361) and approving the mandate. Ad locum controls previously varied depending on the legal basis of international agreements. In first-pillar matters (now extended to all negotiations under Article 218 TFEU), the EU negotiator often received new negotiating directives or had to consult a Council special committee; in some cases, member states even sat alongside the Commission at level I, which allowed for direct oversight (Collinson 1999; Damro 2007; Delreux 2008; Kerremans 2006). In third-pillar matters, negotiations were led by the Presidency, which afforded even closer control by the Council, since the EU negotiator was one of its members (Delreux 2008). However, it is the ratification stage that has changed the most after Lisbon. There, the ex post controls are relatively straightforward; both the EP and the Council may (threaten to) withdraw their consent from signing or ratifying the agreement, facing the EU negotiator with an ‘involuntary defection’ (Putnam 1988: 438).

Finally, it is also worth taking into account the level I opponent, which might set out to restructure the EU's domestic level using various mechanisms such as side-payments, raising the cost of no agreement, or reverberation (Putnam 1988).

Therefore, it is clear that three interrelated elements: levels of negotiations, decision-making stages, and strategies, are necessary to understand how the EU institutions behave in international negotiations and when they have more chances to exert influence over the outcome of the agreements. The following section analyses these three elements in relation to the first international agreement passed after the change of procedure in the Treaty of Lisbon.

EU–US SWIFT AGREEMENT: PRE- AND POST-LISBON NEGOTIATIONS

The EU–US SWIFT Agreement was the first international agreement to be subjected to ratification by the EP. The objective of the Agreement is to manage the access of US authorities to personal bank data contained in the servers of the Belgian SWIFT company. After years of controversy around the use of SWIFT data by US counter-terrorist authorities (European Parliament 2007; González Fuster et al. 2008), the company decided to move the server containing European data to servers in Switzerland and the Netherlands (Kaunert and Leonard 2011). In practice, this meant that US authorities could not have access to data considered essential for their security, which triggered talks on an agreement that would allow the US to access European data without having to resort to bilateral agreements, which were seen as too cumbersome (Occhipinti 2010: 137).

The particularity of the SWIFT Agreement is that it was negotiated in two different rounds that fell in the midst of a procedural shift elicited by the Treaty of Lisbon. A first interim agreement, planned to last nine months, was initiated under the old decision-making rules of Amsterdam (third-pillar rules). Negotiations during the summer and autumn of 2009 were led by the Swedish Presidency and finalized and agreed upon by the Council on 30 November 2009, one day before the entry into force of the Treaty of Lisbon. This decision created clear tensions with the EP, which eventually refused to ratify the agreement on 10 February 2010. After a few weeks of stalemate, the US decided to press on with a new permanent agreement, ratified by the EP on 8 July 2010 (see also Ripoll Servent and MacKenzie 2011, 2012). Therefore, these two rounds of negotiations allow us to compare the strategies of a same set of actors leading to opposite results (failure of ratification, followed by success).

Interim SWIFT Agreement: explaining the ‘involuntary defection’

In order to understand the events of the first round of negotiations, it is important to underline the two dimensions that structured the debates, namely a set of substantive interests that considered the necessity and proportionality of the agreement – particularly in view of data protection concerns – and a set of procedural interests that reflected different views on how the game ought to be played.

On the one hand, the EP was characterized for its liberal substantive positions and a loose interpretation of the rules of the game.³ Therefore, the EP made it clear that it would only accept an agreement that upheld high levels of data protection. It preferred to ratify a text that included judicial review in the US (instead of a mere administrative review), as well as a system that would avoid transferring data in bulk to the US. Bulk data refers to a ‘pull system’ where the US can access information without any previous filters. In contrast, a ‘push system’ (preferred by the EP) requires a previous request for specific data that is then transmitted to the US on a case-by-case approach. These misgivings actually led the EP to question the necessity of this instrument, considering that having no agreement might be preferable to infringing upon data protection standards (European Parliament 2009). At the same time, the EP wished to participate in determining the outcome of negotiations, not simply being informed about the proceedings.⁴ In the EP's understanding of the rules, the ability to ratify the agreement translated into a genuine capacity to accept or reject the outcome reached at level I and was not intended to be a mere rubber-stamping exercise.

In comparison, the Council and the US shared a very different understanding of the issue at hand. In the case of the US, its definition of the substantive problem was in almost complete opposition to that of the EP. They wanted a pull system, which gave them more freedom to analyse bulk data and find useful leads in their fight against terrorism (Argomaniz 2009: 129). In general, the US negotiator was reluctant to introduce any clauses that would force the US Congress to change its domestic legislation. Therefore, it would refuse any requests for the inclusion of judicial reviews or those aiming to strengthen the US's (low) data protection standards (González Fuster et al. 2008) on grounds of having its hands tied by its domestic level. In procedural terms, the US was in favour of sticking to the formal rules of the game, so as to have only one interlocutor – especially one that was close to its own interests.

Indeed, the Presidency essentially reflected the Council's position on these issues, which was very close to the US's substantive and procedural concerns. In relation to a Terrorist Finance Tracking Programme (TFTP) instrument, most member states had an interest in having an agreement because they were convinced that offering the US access to bulk data would be an efficient way to outsource their own security (Occhipinti 2010: 137). The Presidency and the Council diverged only when countries such as Germany, Austria or Hungary raised concerns around the lack of parity between EU and US citizens in cases of review of their personal data, and the lack of oversight over the transfer of bulk data (Monar 2010). In procedural terms, both the Council and the Presidency interpreted the rules in their most literal form and expected the consent procedure to replicate previous experiences, where the EP's ratification had been a simple formality.

Therefore, the wide gap between the position of the two domestic actors (EP and Council) led to a particular scenario. During the agenda-setting and negotiation stages, the Council used its proximity to the Presidency (acting as its agent) to control the developments at level I, which led to a strategy of collusion between the EU and US negotiators. This strategy, however, was differently perceived by the EP, which saw the Presidency as a ‘dove’ – acting in the interest of the US and ignoring MEPs' concerns.

Figure 2 Strategies used during negotiations and ratification of the interim SWIFT Agreement

Note: Dashed line indicates strategies used only during the ratification stage. Dotted line indicates secondary US strategy.

This strategy proved hazardous during the ratification stage. After the signature of the Agreement, all actors involved realized that there was a high risk of ‘involuntary defection’. As a result, they engaged in a frenetic strategy of restructuration with some members of the Council and the (now Spanish) Presidency attempting to put pressure on the EP⁵ and the US partaking in various lobbying activities (Monar 2010: 145). However, these last-minute attempts failed to prevent the rejection of the interim agreement in February 2010.

Permanent agreement: re-negotiations and successful ratification

This ‘involuntary defection’ eventually led to the renegotiation of the interim agreement, this time aiming to achieve a permanent agreement. From the start of the re-negotiations, it was clear that the gap at the domestic level would have to be closed in order to ensure the EP's consent. As a result, the different sides put several strategies into place aiming to satisfy both the EP's substantive and procedural interests. It is important to recall here that this second round of negotiations was the first to occur fully under the Treaty of Lisbon. Consequently, while the Council was still responsible for approving the negotiation mandate and signing the agreement, it was the responsibility of the Commission to lead talks with the US.

Figure 3 Strategies used during negotiations on the permanent SWIFT Agreement

Note: Dotted line indicates secondary US strategy.

Both the Commission and the Council wished to reach an agreement with the US that could satisfy the EP without stepping on the US's red lines.⁶ As a result, level I continued to be characterized by an important degree of collusion; similarly, the US negotiator resumed its tying-hands strategy – pointing at the potential veto of the US Congress to avoid changes in its domestic legislation.⁷ In order to convince the EP, the Commission cut slack on substantive terms and offered some side-payments in the new negotiating mandate, such as increased data protection safeguards (European Commission 2010c). It also created synergistic issue-linkages, promising to use the expertise and support of the US to develop a similar system inside the EU (European Commission [2010b]; see references to an EU-TFTP in European Union 2010: Article 11). In addition, the Presidency decided to bypass the Commission and add a supplementary side-payment in the form of a ‘scrutineer’ – an EU official posted in the US for the purpose of overseeing the use of European SWIFT data (European Union 2010: Article 12).

These efforts on the part of the Commission and the Council were reinforced by the US's use of a restructuration strategy. First, the cost of non-agreement was raised by threatening to opt for bilateral agreements instead of a single EU–US deal (Monar 2010: 145). This meant that the international agreements would have to be ratified by each member state and that the EP would lose any leverage on the content of negotiations. Second, the US also offered side-payments, such as making more explicit the parity between EU and US citizens (European Union 2010: Article 5). Third, the US used reverberation to convince EP actors to change their understanding of the issues at stake. To this effect, they invited key MEPs, led by the rapporteur and the LIBE committee's chairman,⁸ to visit the US (Euractiv 2010). Additionally, US Vice-President Joe Biden addressed the EP on 6 May 2010 (European Parliament 2010b). These combined efforts made the level I deal more acceptable and legitimate for the EP.

Their efforts also targeted the procedural dimension – where the main difficulties rested in the different understandings of the new procedural rules introduced by the Treaty of Lisbon (to be informed vs full participation). The EP's willingness to effectively exert its new power of consent led the Commission to include it in early talks – even before the draft mandate was put in front of the Council,⁹ which allowed the MEPs to participate in the agenda-setting stage. The negotiating directives openly stated that ‘the Commission shall, in the course of negotiations, aim to achieve the specific objectives set out in detail below [while reflecting, to the extent practicable, the recommendations made by the European Parliament in its Resolution (P7_TA(2010)0143)]’ (European Commission 2010a). Therefore, the Commission's actions were not only guided by the Council's mandate – as foreseen in the Lisbon Treaty – but also by the EP resolution of 5 May 2010 (European Parliament 2010a).¹⁰

However, the EP increased its capacity to control the outcomes of negotiations not only by exerting influence during the agenda-setting stage – which might be seen as a measure of the Commission to anticipate the preferences of a veto player¹¹ – but also by developing new mechanisms of ad locum control. On the one hand, it shifted the procedural understandings of the Council, which became more open to EP initiatives and gave the MEPs a chance to control the pace of negotiations.¹². The inclusion of the EP was especially apparent during the last stages of negotiations, when the Council Presidency dealt directly with the EP and forced the Commission to renegotiate the inclusion of an EU ‘scrutineer’.¹³ On the other hand, the EP's failure to ratify the interim agreement also led to a change in the position of the US on the procedural dimension. US negotiators helped the EP to exert more control over the agenda-setting and negotiation stages by accepting its inclusion in level I talks. As mentioned above, the US invited key MEPs to visit the US on several occasions in order to discuss the main issues with senior US representatives (Euractiv 2010), notably with Janet Napolitano (United States Secretary of Homeland Security).¹⁴ The US efforts in finding informal ways to enter into contact with key MEPs¹⁵ served to tip the balance, since it by-passed the Commission in its exclusive right to conduct talks at the international level. This strategy gave the EP a better chance to steer the pace and content of negotiations and, thus, to exert more control over the Commission. The combination of this multi-front strategy served to close the gap in the interpretation of the formal treaty rules.

Table 2 Control mechanisms used during negotiations for a SWIFT Agreement

	Ex ante (agenda-setting stage)		Ad locum (negotiation stage)		Ex post (ratification stage)
Post-Lisbon (formal rules)	Council	–	Council	–	Council	EP
Mandate	–	Council directives/special committee	–	Signature and ratification	Ratification
Pre-Lisbon (interim)	Council	–	Council	–	Council	EP
Mandate	–	Presidency's double role	–	Signature and ratification	Ratification
Post-Lisbon (permanent)	Council	EP	Council	EP	Council	EP
Mandate	Informal consultations	Council directives/special committee	Observers	Signature and ratification	Ratification

The outcome of negotiations offers a different picture of the EP's capacity to influence and control EU international negotiations in a two-level game. They show how the EP was not satisfied with the new formal ex post control mechanism (ratification) and, therefore, made use of day-to-day decision-making to informally expand its treaty powers. By re-interpreting the formal rules of the Treaty, the EP shifted its role from that of a veto player (i.e., able to block international agreements during the ratification stage) to that of an agenda-setter able to influence both level I and level II negotiations. At level II, the EP gained an effective agenda-setting power, which gave it the chance to control the content of the Commission's proposal for a new mandate. It also kept a close collaboration with the Council Presidency, which helped it to shape the mandates and negotiating directives issued at later stages of negotiations. At level I, the EP managed to set a direct dialogue with the US by entering into early talks with key US authorities, both in Brussels and Washington. In addition, the EP managed to control the movements of the Commission and actively steer the outcomes of the agreement – even to the point of intervening directly (with the support of the Presidency) during level I negotiations to find a more favourable outcome.

CONCLUSION

The analysis of a failed ratification followed by a successful one has allowed us to determine changes in the strategies used by institutional actors at the EU and international levels. The change in strategies shows also a shift in their formal and informal power to exert influence and control over the negotiations during the various decision-making stages (agenda-setting, negotiation, and ratification). Therefore, an in-depth examination of the SWIFT Agreement shows how looking into domestic arrangements helps us to understand how formal rules may be informally challenged and expanded. The ability of specific EU institutions to increase their autonomy and influence resonates with Pollack's (1999) findings in the framework of inter-governmental conferences. The EP used the institutional uncertainty that followed the entry into force of the Treaty of Lisbon to successfully reinterpret its right to be ‘informed’ in international negotiations and transformed it into a right to be ‘involved’. As a result, the EP was able to do more than merely rubber-stamp international agreements; it asserted its position as an agenda-setter capable of steering the EU negotiators' agenda and controlling the proceedings of negotiations both at the domestic and international level. Therefore, the shift in the formal and informal role of the EP emphasizes the importance of domestic political institutions to understand the outcomes of level I negotiations. In future, informal institutions and processes of change should be given greater consideration when constructing models to explain EU international negotiations after the Treaty of Lisbon.

Certainly, institutional roles are still in flux. Interstitial changes need time to become accepted and institutionalized. However, it is interesting to observe that the EP's efforts in the course of the SWIFT negotiations have borne fruit in later developments. First and foremost, the new Framework Agreement between the EP and the Commission (European Parliament & European Commission 2010) dedicates Section II and Annex III to spelling out the rights and duties of both institutions in international negotiations. The agreement makes explicit the involvement of the EP by stating that the Commission shall provide Parliament with all the information on the negotiation and conclusion of international agreements ‘in sufficient time for [the EP] to be able to express its point of view if appropriate, and for the Commission to be able to take Parliament's views as far as possible into account’ (Points 23, 24). The agreement also foresees the possibility of including a ‘delegation of Members of the European Parliament as observers’ in negotiations (Point 25). The annex specifies that ‘the Commission shall take due account of Parliament's comments throughout the negotiations’ (Annex III, Point 3). Therefore, the Commission has accepted and even formalized the EP's attempts to participate and be fully involved at all stages of negotiations.

In addition, the EP has used the opportunity in later international negotiations to expand its influence and show its willingness to become more than a mere veto player. The negative vote on the Anti-Counterfeiting Trade Agreement (ACTA) proved the willingness of the EP to assert its ambitions vis-à-vis the Commission and the Council. Matthews (2012) noted that one of the reasons for the EP to withdraw its consent was the need to exercise its new veto power; it also was concerned with the transparency of negotiations. Indeed, after the rejection, Hannes Swoboda, President of the Socialist & Democrats group, declared that ‘[f]or the first time, the European Parliament has used the powers conferred upon it by the Treaty of Lisbon to reject an international trade agreement. The Commission and Council are now aware that they have to reckon with the Parliament, which represents and defends the citizens’ (Agence Europe 2012). The EP's Recommendation also underlined that:

in line with Article 218(10) of the Treaty on the Functioning of the European Union (TFEU) Parliament must be immediately and fully informed at all stages of the procedure; [the EP] takes the view that adequate transparency has not been ensured throughout the negotiations on ACTA; recognises that efforts to inform Parliament have been undertaken by the Commission, but regrets that the requirement of transparency has been construed very narrowly and only as a result of pressure by Parliament and civil society. (European Parliament 2012: Point 3)

The growing influence of the EP, as well as its capacity to exert control over all stages of decision-making, has the potential to develop into a more stringent relationship where the EP would be able to (informally) delegate tasks to the Commission – converting it into a (informal) principal. The phrasing of the negotiating directives for a permanent SWIFT Agreement shows an incipient form of delegation, since the EP Resolution did not only shape but also informally bind the Commission during negotiations. A similar development was observed by Elsig and Dupont (2012) in their examination of the EU–South Korea trade agreement of 2011, where the EP used its newly acquired ratification power to force the Commission to renegotiate certain clauses of the agreement and set a clear path for implementation. Therefore, they concluded that ‘a newly emerging principal (competing with other principals), the EP, ascended as a key player during the ratification game … For the EP, the treaty with South Korea provided the possibility for it to flex its muscles and set a precedent’ (Elsig and Dupont 2012: 502).

The ability of the EP to transform itself into an agenda-setter, or even an emerging informal principal, opens new avenues for future research. The SWIFT case calls for further comparative studies that allow for more light to be shed on the institutional dynamics at the domestic level (i.e., between EP and Council) and at the international level (namely, their capacity to delegate and control the Commission or High Representative). It is important to understand whether this new constellation leads domestic institutions and negotiators to adopt new strategies – both in cases where the EP and Council share the same interests or where they face potential cases of inter-institutional conflict. The case study gives us some clues about future strategies; for instance, the willingness of the Council and the EP to take over the role of negotiator from the Commission in the last stages of the SWIFT negotiations might be an example of an extreme form of ad locum monitoring, and it would be necessary to see whether it becomes a new practice in future negotiations. At the same time, it is equally important to explore further cases were the EP and the Council present conflicting interests. This configuration might be particularly challenging for the Commission or the High Representative as agent, since they would need to develop new strategies to please actors at the domestic and international level without being seen as a ‘dove’ or a ‘hawk’ by either the EP or the Council. It will be interesting, therefore, to observe how the EP makes use of its newfound position and how the relationships between multiple domestic institutions (and potential principals) and a single agent evolve in the coming years.

ACKNOWLEDGEMENTS

I wish to thank Florian Trauner, Tom Delreux, Guido Schwellnus, Sebastian Krapohl, Christian Dorsch and Kai Oppermann for their invaluable observations. I am also grateful for the comments I received at the 19th CES Conference (Boston) and at an internal workshop of the Sussex European Institute. Also for Alex MacKenzie's co-operation in the early stages of this project and the JEPP referees for their helpful comments.

Additional information

Biographical note: Ariadna Ripoll Servent is junior professor at the University of Bamberg, Germany.

Notes

1 Note that the official name of the agreement is the TFTP (Terrorist Finance Tracking Programme) Agreement. It was, however, known by the name of the SWIFT company (Society for Worldwide Interbank Financial Telecommunication), which dominates the bank transfer market.

2 The study integrated as much hard data as it was possible to gather in the form of official documents, press releases from EU institutions or from individual MEPs/political groups, EP plenary votes, questions and debates in the EP, and press articles from newspapers or advocacy groups.

3 There were no roll-call votes for the EP resolution of 17 September 2009 or the vote on the interim agreement. For more information on the position of the different political groups on this issue, see (Ripoll Servent 2011: 165 ff.)

4 Diplomatic source A, interview; EP official A, communication, March 2010.

5 EP official A, communication; diplomatic source A, interview, March 2010

6 Diplomatic source A, interview, March 2010.

7 EU official, interview, March 2010.

8 Hennis-Plasschaert, MEP, interview, March 2010; MEP, interview, July 2010.

9 Diplomatic source A; Hennis-Plasschaert, MEP, interviews, March 2010; Commission official G; diplomatic source B; EU official, interviews, July 2010.

10 Commission official F, interview, July 2010; European United Left - Nordic Green Left (GUE/NGL) political advisor A, interview, March 2011.

11 I want to thank an anonymous referee for pointing out this alternative explanation.

12 Alvaro, MEP, interview, July 2010.

13 MEP; diplomatic source A; EU official; Commission official G, interviews, July 2010.

14 MEP, interview, July 2010.

15 Commission official G, interview, July 2010.