The International Bargaining Power of the European Union in ‘Mixed’ Competence Negotiations The Case of the 2000 Cartagena Protocol on Biosafety WORD COUNT: 8,969 Paper presented at the European Union Studies Association (EUSA) Conference March 31-April 2, 2005, Austin, Texas Mark Rhinard Michael Kaeding Postdoctoral Fellow PhD Candidate University of Leiden University of Leiden PO Box 9555 PO Box 9555 2300RB Leiden, The Netherlands 2300RB Leiden, The Netherlands +31 6 1236 6555 +31 71 527 3888 rhinard@fsw.leidenuniv.nl kaeding@fsw.leidenuniv.nl ABSTRACT: Studies of the European Union’s role in global policy negotiations typically focus either on international trade agreements, where the Community enjoys exclusive competence in negotiations, or on military and security accords, where member states take the diplomatic lead in international talks. Between these extremes are a number of issue negotiations in which Community officials and member state representatives speak jointly on behalf of the EU. We analyse one such ‘mixed competence’ case, the negotiations for a global Biosafety Protocol, to understand more about the EU’s behaviour and bargaining power in such settings. Rather than construct a new model for mixed competence cases, however, we aim for explanatory parsimony by applying an existing model designed to predict EU bargaining power in international trade negotiations (Meunier 2000). The analysis reveals that the trade model has considerable explanatory power for some of the outcomes found in the Biosafety Protocol negotiations, but incorrectly predicts that the EU will have no bargaining power. We then identify the sources of discrepancy between the model’s predictions and our empirical results to pinpoint the key differences between exclusive and mixed competence negotiations. We conclude the article by making suggestions for future model building and by reflecting on changes in the new Constitutional Treaty that may affect the EU’s bargaining power on the global stage. ACKNOWLEDGEMENTS: The authors would like to thank Paul ‘t Hart, Markus Haverland, and Martijn Groenleer along with Kenneth Oye and the other members of the panel ‘The Politics Created by the Nesting of International Regimes’ (American Political Science Association Annual Conference, Chicago, IL, September 2, 2005), for helpful comments on previous drafts of this article. The International Bargaining Power of the European Union in ‘Mixed’ Competence Negotiations The Case of the 2000 Cartagena Protocol on Biosafety INTRODUCTION The European Union (EU) has emerged as an influential global political actor over the decades since its inception. On trade-related matters, its twenty-five members negotiate with a ‘single voice’ through Commission negotiators, having entrusted the Community with exclusive competence to complete international agreements. Using such leverage, the EU holds considerable sway on the international stage, as indicated by its influential role in the General Agreement on Tariffs and Trade (GATT) and World Trade Organization (WTO) over the years (Meunier, 2000; Meunier and Nicolaïdis, 1999; Bretherton and Vogler, 1999). On defence and security issues, the EU has been slower to develop a unified front, although recent progress on the Common Foreign and Security Policy (CFSP) suggests the EU is moving beyond simple declarations towards operational forces and a heightened role in global crisis management. CFSP remains essentially an intergovernmental policy area, though, with the Commission and EU officials playing a subordinate role (Hix, 2005). The EU is represented in global settings primarily by its member states, acting through the Presidency and ‘troika’. Yet for most of the EU’s external policy activity, the institutional balance of power falls somewhere in between these high-profile extremes. Policy competence for international negotiations does not lie solely with the Community, nor is it the exclusive preserve of member states. Instead, competence is shared or, in legal terminology, ‘mixed’ (Leal-Arcas, 2001; Sheridan, 2001, p. 15). International negotiations over such issues require the direct participation of both Community and member state officials, and lead to mixed agreements. Recent negotiations over international financial rules, gender discrimination conventions, and environmental accords are three examples in which the EU’s negotiation competences were shared between supranational and national officials. With the increased global activism of the EU in these and similarly ‘mixed’ areas, there is a growing need to understand its negotiating behaviour in such fora. Scholars have indeed modelled the EU’s bargaining capabilities in international negotiations. One of the more coherent and sophisticated approaches is that of Meunier (2000). Meunier draws upon bargaining and principal-agent theory, together with new institutionalism and studies of multilevel games, to construct a model of the EU’s bargaining power in international negotiations. Three variables comprise the model: (1) the nature of the delegation mandate to supranational actors, (2) the EU’s internal voting rules used to determine the mandate, and (3) the characteristics of the international negotiating context. Predictions emerge from different value combinations of each variable. Meunier tests the explanatory power of this parsimonious model with three cases, and finds that if performs rather well. Yet the Meunier model is designed to predict EU bargaining power within international trade negotiations. In those cases, the Community enjoys exclusive negotiation competence (albeit with specific oversight mechanisms). How can we explain the bargaining power of the EU in global negotiations involving mixed competence issues? The EU’s role in mixed competence negotiations is bifurcated between supranational and national officials, yet whether this affects the bargaining power of the EU remains an empirical question. Rather than construct a separate model specific to mixed competence negotiations, the aim of this paper is to replicate the findings of Meunier by applying her model to a different case: the EU’s involvement in the Cartagena Protocol on Biosafety (hereafter, ‘Biosafety Protocol’). The adoption of this protocol in January 2000 established the first binding international guidelines for the transboundary movement of genetically modified organisms (GMOs). By applying this case to an existing model, we aim to provide insights in a cumulative fashion along the lines of a positivist approach to scientific research. The paper is structured as follows. First, we briefly review the literature on the EU in global negotiations before presenting the key parameters and predictions of the Meunier model. Second, we demonstrate the differences between ‘exclusive’ and ‘mixed’ negotiation competences in the EU, and highlight environmental policy negotiations as an example of the latter. Third, we test the predictions of the Meunier model against the case of the Biosafety Protocol negotiations. While the model predicts no bargaining power for the EU in this type of case, we show, against expectations, the EU was a tough bargainer and achieved virtually all of its aims in the final agreement. Fourth, we discuss the potential sources of this discrepancy in the Meunier model and offer suggestions for future model building. The conclusion summarizes the paper and highlights several trends that may affect the EU’s behavior in forthcoming international negotiations. 1. THE EUROPEAN UNION IN GLOBAL NEGOTIATIONS While the European Union is taking part in international negotiations on a growing basis, such a pooling of political influence and national sovereignty raises complex issues. To cut through the complexity, scholars turn to several bodies of literature relevant to international negotiations in order to model the EU’s role and help explain outcomes. Two types of studies stand out: those associated with multilevel games and those concerning delegation. Putnam (1988) initiated a lively debate on the strategic interaction between different bargaining levels, at which negotiators interact simultaneously with both their foreign counterparts and their domestic principals. By demonstrating how developments at one level ‘reverberate’ at other levels, Putnam focused attention onto constraints and opportunities that arise from domestic politics, national and international institutions, and negotiator strategies (1988, p. 450; see also Evans et al., 1993). Multilevel games approaches find rich empirical terrain in the case of the EU, where studies typically expand the model to include the three levels that interact during international bargaining: the domestic, supranational, and international (see, for example, Patterson, 1997). Delegation theories have also been applied with increasing sophistication to the EU. Originally developed to study legislative-executive relations in the US government (Epstein and O’Halloran, 1999), delegation theories help to untangle the variety of political and institutional relationships that comprise the EU political system (Pollack, 1997, 2003; Moravcsik, 1998; Tallberg, 2000, 2002; Majone, 2001; Franchino, 2002, 2004). Scholars have coupled the study of inter-institutional delegation at the EU level with the additional delegation aspects that occur in the context of international negotiations (Meunier and Nicolaïdis, 1999; Jupille, 1999). Such studies reveal the extent to which the EU’s external bargaining capacities are directly related to its institutional structure. Meunier and Nicolaïdis (1999) argue in particular that voting rules at the EU level and the amount of ‘competence’ exercised by EU negotiators help to explain the shape of the final international agreement. Meunier (2000) draws upon these bodies of literature to highlight the institutional factors that provide the EU with varying levels of bargaining capability within global policy negotiations. More specifically, she points to three variables that combine to determine the EU’s external bargaining power: the nature of the delegation mandate, the internal EU voting rules used to decide on the mandate and ratify the agreement, and the characteristics of different negotiating contexts. Determining the Supranational Competence: At the outset of any international negotiation in which the EU takes part, a mandate must be designed and approved (Meunier, 2000; Meunier and Nicolaïdis, 1999). The member states principals use the mandate to control Community negotiators, serving as agents, in the process of negotiations. The mandate is defined by three features: flexibility, autonomy, and authority. Depending on the nature of these three attributes, there are two types of mandates: restrictive and extensive. Flexibility relates to the degree of latitude afforded to Community negotiators by the Council at the outset of negotiations. Member states in Council may, for example, specify precisely which issues are ‘on the table’ and which are not; member state principals may also detail what concessions they will view as acceptable. Alternatively, the latitude provided to the Community by a Council mandate may be more vague and commodious, thus allowing the Commission considerable leeway to take different positions during a negotiation (Meunier 2000, p. 111). Autonomy refers to the degree of oversight that the Council may apply to Community negotiators during actual negotiations. The principal may require regular reporting by the agent or may even sit at the negotiation table alongside the agent to monitor behaviour (Nicolaïdis, 1999, p. 11). The use of a committee of member state representatives to oversee Community negotiators is well-established for both trade and environmental negotiations through Article 300 TEC. Yet for environmental policy and other ‘mixed’ competence issues, the question of autonomy is more nuanced considering the legal right of Council principals to negotiate side-by-side with Community agents. Authority refers to the ability of the agent to deliver on promises made to opponents during negotiations. This attribute is used especially within multi-level game analysis to show how different procedures used for domestic ratification shape whether negotiators can convince opponents that they can keep negotiation promises ‘back home’. As Nicolaïdis points out, this is the ‘most visible and quantifiable constraint’ of the three attributes (1999, p. 11, cited in Meunier 2000, p. 111). The values of each of these attributes combine to create a negotiation mandate that is either restrictive or extensive. In a second step to her model, Meunier argues that supranational competence subsequently is defined by the combination of the nature of the negotiator’s mandate and the voting rule—unanimity or qualified majority voting—that aggregates actors’ preferences. Table 1 identifies two types of supranational competence as described by Meunier. Table 1: Supranational competence Unanimity voting Qualified majority voting Restricted mandate low Extensive mandate high Unanimity voting means that each member state possesses a veto over the mandate at the outset of the negotiations and over ratification at the conclusion of the negotiation. This implies that outcomes in both cases will most likely fall to the lowest common denominator (Garrett and Tsebelis, 1996, p. 281). Under unanimity voting, the common position will be dictated by the most conservative member state, whereas the member with the most extreme preferences will be more flexible, preferring any proposal that represents an improvement over the status quo. Combined with a restricted mandate, unanimity voting leads to a low degree of supranational competence. Under majority voting, member states must gather about one-third of the total EU votes in order to block a procedure. This tends to mitigate extreme positions, and benefit the preferred position of a majority of member states. Under majority voting, the common position falls closely to the proposal originally made by the Commission (Jupille, 1999). An extensive autonomy, under this voting rule, therefore results in a high level of supranational competence. Conducting the Negotiation To mediate the impact of the above internal institutional variables on the external bargaining capabilities of the EU, Meunier identifies two characteristic bargaining contexts (2000, pp. 118-21). The distribution of policy preferences of the EU and its negotiating opponent relative to the status quo leads to either a conservative negotiation context or a more reform-oriented negotiation context. In the conservative case, the preferences of the EU member states and of the negotiating opponent are distributed so that the opponent’s preferences are furthest away from the status quo. Hence the EU is the recalcitrant, or conservative, partner in the negotiations. In the reformist case, the distribution of preferences is such that the EU’s preferences are furthest away from the status quo. Here, the EU is the one making demands for change while the opponent prefers to stay close to the status quo. How does the negotiation situation condition the external impact of the EU’s decision rules? Table 2 summarizes the predictions. Table 2: EU supranational competence and external bargaining power: NEGOTIATION ENVIRONMENT SUPRANATIONAL COMPETENCE CONSERVATIVE CASE REFORMIST CASE LOW (UNANIMITY VOTING AND RESTRICTED MANDATE) HIGH EU BARGAINING POWER NO EU BARGAINING POWER HIGH (QUALIFIED MAJORITY VOTING AND EXTENSIVE MANDATE) SOME LIMITED EU BARGAINING POWER NO EU BARGAINING POWER Source: Meunier, 2000, p. 120. Differentiating between conservative and reformist negotiating contexts, Meunier argues that the EU’s strength as an international trade negotiator is determined partly by two distinct institutional mechanisms: the de facto voting rules used by the EU Council of Ministers and the negotiating competence delegated to Commission representatives. When the collective position of the EU is closer to the status quo than that of its opponent (conservative case), supranational competence relaxes the extremes and renders the conclusion of an international agreement more likely but dispossesses the EU of some bargaining leverage. In this case, the opponent cannot obtain departures form the status quo greater than what the EU is willing to offer. Nevertheless, it can potentially win more or fewer concessions depending on the voting rules governing the EU’s internal decision-making process. Unanimity voting reinforces the potential bargaining power of the Community. Qualified majority voting mitigates the extremes. It increases the likelihood that an agreement will be reached, since there is no uncertainty that the final deal will be approved by the Council. However, it does not enhance the EU external bargaining power as much as unanimity voting does. By contrast, a high degree of national control over the bargaining process renders the conclusion of an international agreement less likely, but can successfully tilt the final result of the international negotiation toward the position of the most recalcitrant member state. Internal EU institutional procedures are of less importance to negotiation outcomes when the opponent wants a smaller departure from the status quo than does the EU (reformist case), but these procedures still influence the process and likelihood of a negotiation. Unanimity makes a Community-led offensive less likely. It is easy to imagine that the negotiating opponent can try to ‘divide and rule’ the member states by introducing its own ‘Trojan Horse’ into the EU’s ranks. The opponent can encourage an EU member state to threaten a veto in order to stop the EU from pressing for policy change. If the negotiation was to proceed, the outcome would stand at the opponent’s ideal point. Qualified majority voting has the effect of making the negotiations more likely. The absence of veto power deprives the negotiating opponent of the option of driving a wedge between member states to derail the EU’s offensive. Meunier tests her model by analyzing three case studies of EU-US trade negotiations on agriculture, public procurement, and open skies (2000, pp. 121-31). Her results confirm that, given exogenous member states’ preferences, the institutional mechanisms through which member states transfer their sovereignty affect international trade agreements. Yet this model was developed to address trade issues. Is it appropriate for predicting EU bargaining in multilateral policy negotiations that address other issues, such as the environment? Environmental policy is a shared competence, between member states and the Community. International negotiations are thus ‘mixed’: the EU delegation is composed of both member state representatives and Community negotiators. The latter do not carry the full authority of the EU behind them, as in trade cases. The next section identifies the general features of mixed competence negotiations and gauges how far Meunier’s model is applicable. 2. ‘MIXED’ COMPETENCES IN THE EUROPEAN UNION The European Union possesses external competences across a broad range of areas, but the extent and nature of this competence differs significantly depending on the policy field. In the area of international commerce, Member States participate only at ‘arms-length’ during international trade negotiations. Here, the Commission enjoys extensive competences. At the other extreme, member states firmly control the pace of policy developments and provide the face behind the CFSP, for instance, in international affairs. In global talks on these latter issues, the Commission plays a subordinate and largely administrative role. The majority of policy areas in which the EU has some external presence, however, do not fit with either the exclusive Community or member state competence models. International negotiations addressing most policy areas fall somewhere between these two extremes. They are ‘mixed’ negotiations, in which both member states and the Community have ‘concurrent powers’ to conclude such agreements. Global environmental negotiations: Global environmental agreements represent an excellent case of a ‘mixed’ competence for the European Union. In contrast to provisions on trade, environmental issues were absent from the 1957 Treaty of Rome. Not until the early 1970s did the European Community begin to develop its environmental policies (Jupille and Caporaso, 1998, p. 220). The rationale for such policies soon expanded, from addressing trade distortions to the need for sustainable development, and by the 1990s the EU had an array of environmental instruments in place (Bretherton and Vogler, 1999, p. 81; Sbragia, 1998). Internal developments preceded the Community’s capacity to act internationally, but during the 1980s four factors contributed to a growing EU role in international environmental politics. First, the cross-cutting nature of environmental policy touched upon many areas that were already part of the Community’s competence. Second, a series of activist judgments by the European Court of Justice established the doctrine of ‘parallelism’, ruling that when the Community acquires internal competence over a policy it also acquires external competence. Third, the adoption of the Single European Act created a solid legal basis for environmental policy both inside and outside the Community. Finally, policy entrepreneurship by the European Commission, particularly its Directorate General for the Environment, has helped to stimulate the EU’s global environmental role (Jupille and Caporaso, 1998, pp. 220-1). Today, European environmental policy has a substantial international dimension. The Community is now a signatory to, and continues to participate in, no less than 31 major multilateral environmental agreements (Bretherton and Vogler, 1999, p. 85), as well as a host of protocols that can be understood as separate accords within each agreement. Yet, in contrast to trade policy, the Community does not enjoy ‘exclusive’ authority to speak on behalf of the EU member states, nor is the Commission the sole bearer of the Community mandate. The uneven and ad hoc development of environmental policy internally, along with a Treaty provision stipulating shared competence externally, means that the Community and member states operate concurrently in most environmental negotiations. This leads to ‘mixed’ agreements. Each new mixed negotiation raises basic questions of competence, division of labour, and ratification procedures: all contentious issues between member states and the Commission. Essentially, the Commission, as representative of the Community during negotiations, has earned the right to participate in those negotiations alongside the member states—rather than in lieu of member states. Analysing Mixed Competence Negotiations: There are two ways to view the role of the EU in ‘mixed competence’ negotiations. One approach is to argue that such negotiations are fundamentally different from exclusive competence negotiations, such as those associated with global trade agreements. Because member states and the EU share competences and since the allocation of competences is often ‘not firmly established ahead of time’ resulting in a ‘on the spot…negotiation-within-a-negotiation’ (Jupille and Caporaso, 1998, p. 218), any model based on a formal principal-agent model might be considered inappropriate. A second interpretation suggests treating a ‘mixed competence’ issue negotiation simply as a variant of the types categorized by Meunier. Our evidence suggests that we can operationalise the same central independent variables as those in the Meunier model. The nature of the negotiating mandate, either restrictive or extensive, is one variable, for instance, that varies in the same fashion as in trade cases. Depending on the three attributes of flexibility, autonomy, and authority, the Community enjoys different levels of competence in international bargaining settings. Voting rules, and the nature of the negotiation context, are two variables from the Meunier model that also apply to most cases. We follow the second interpretation and seek to gauge the applicability the Meunier model, unaltered, to a new category of cases. Not only is this interpretation consistent with our effort to preserve parsimony, it also allows us to pinpoint key areas of similarity and difference with regard to mixed versus exclusive competence negotiations. The following section introduces the 2000 Biosafety Protocol negotiations. Falling under qualified majority voting, and with the Commission enjoying rather high bargaining competence in a reformist negotiating context, the Biosafety Protocol negotiations offer an interesting case to test the explanatory power of the Meunier model. 3. NEGOTIATING THE 2000 CARTAGENA PROTOCOL ON BIOSAFETY The 2000 Biosafety Protocol is the first international agreement regulating the transboundary movement of GMOs. An offshoot of the 1992 Convention on Biological Diversity (CBD), the agreement applies to genetically modified food and feed, allows importing nations to reject GMOs based on the ‘precautionary principle’, and implicitly challenges the subordination of environment regulation to trade law in the global legal order. Discussions on the protocol began as early as 1994, but did not start in earnest until 1998, with the EU taking a forward position in advocating new global rules for biosafety. The US led the opposition in resisting such efforts, arguing through most of the negotiations that a protocol was not only unnecessary but also in possible conflict with global trade rules. Early negotiations saw a largely ineffective EU negotiating team, hampered by a restrictive mandate and operating under the internal decision rule of qualified majority voting. Yet as negotiations proceeded and EU member state support converged, the situation changed. A unified EU negotiating team presented a strong front and forced the hand of a recalcitrant US in final negotiations. The outcome in 2000 was a Biosafety Protocol very much in line with the EU’s policy preferences and victory for the EU’s activist policy intentions. As shown below, outcomes in this case only partly match the predictions of the Meunier model. The Biosafety Protocol negotiations were characterized by EU internal decision making by qualified majority, along with extensive delegation to Union negotiators—two features that Meunier argues combine to create a ‘high’ supranational competence for EU negotiators. The predictions of the model in this regard hold well: the EU coaxed its opponent into negotiations and preserved a united front throughout the talks. However, the ‘reformist’ context of the protocol negotiations, in which the EU’s policy preferences were further away from the status quo than its main opponent, leads to problems with the Meunier model’s predictions for our case. The model predicts that a ‘high’ supranational competence, combined with a ‘reformist’ negotiating context, will lead to no bargaining power for the EU. Yet the outcome of the Biosafety Protocol negotiations does not bear out this prediction. To illustrate the point, the following discussion presents the empirics of the case in light of the key variables of the Meunier model. It first highlights the construction of the supranational competence and the nature of the negotiating context. In a second step, we analyze the EU’s influence in the negotiation process and final outcomes. Constructing the supranational competence The Biosafety Protocol negotiations were characterized by a long gestation period during which the Community negotiation mandate was revised several times. The idea of international rules regulating the transboundary movement of GMOs first emerged in the early 1990s. The Convention on Biological Diversity (CBD), which entered into force in 1993, contained a single provision allowing for parties to the Convention to ‘consider the need for and modalities of a protocol’ on transboundary GMOs (Article 19.3). Discussions slowly took shape within CBD meetings, to which the EU was already sending Community representatives. A report, commissioned by the CBD secretariat in 1995, set the background against which the need for an international agreement on biosafety would be considered. The report highlighted the lack of effective regulatory structures in many developing countries and how a protocol could help protect such countries in the face of powerful, developed nations seeking to export and field test GMOs. From this perspective, an international protocol would create an ‘even playing field’. Yet a different perspective held that a global regulatory framework was unnecessary and perhaps even reckless. With no clear evidence of danger to illuminate a clear regulatory course, and given the potential for conflict with global trade agreements, regulating GMOs at the international level might cause more problems than it solves. The Community’s first mandate was approved even before talks had moved beyond these general parameters. The Commission needed some indication from member states as to their position on such issues. After submitting a proposal for a mandate to the Council in 1995, the Commission received a negotiation mandate that was fairly typical in cases of mixed competences. Member States agreed that Commission negotiators could take part in the negotiations ‘with regard to matters within the Community’s field of competence’ (Council 1995) and authorized the addition of the Community as a contracting party. It did not set out the precise allocation of competences but sounded a general note of caution. Commission negotiators were urged to explore the ‘need’ for a separate protocol and to continue working on biosafety issues in other international fora. The Commission was required to conduct its negotiations ‘in dialogue with’ a committee of member state representatives. This first negotiation mandate reflected some degree of ambivalence by member states. It was not at all clear whether a protocol on GMOs would (a) be necessary in light of other international developments, and (b) stand a chance of approval in the face of counter-pressure from powerful third countries such as the US. Moreover, the EU had recently emerged from a bruising, internal battle regarding its own biotechnology regulatory framework. That outcome had raised the ire of biotechnology proponents and opponents alike and led to a policy fatigue that militated against further activism (Patterson, 2000). Member states preferences at this early stage were largely weak and undefined, therefore, with a default position near the policy status quo. The vagueness of the mandate should not be interpreted as a carte blanche to begin negotiations, but rather as a cautious ‘wait-and-see’ attitude by member states. As we might expect, a vague mandate and unsettled member state preferences led to problems in subsequent talks. A meeting in Jakarta in November 1995 ‘saw the member states challenge the Commission’s exercise of its negotiation mandate’ (Bail et al., 2002, p. 170). Discussions over a protocol at these early stages remained exploratory; nevertheless, the Dutch presidency of the Council at the time argued forcefully that the Commission was too far out ahead, on its own. Member States concurred that either the Presidency alone should lead negotiations, or that the Commission and a small number of member states should work jointly together. As talks progressed, the Commission sought to clarify the EU negotiation position. In June 1996, it submitted a proposal for a second negotiation mandate to the Council. Member states now took the opportunity to outline a clearer division of competences between Community and member state negotiators: the Commission would handle all trade-related matters and those issues related to the acquis, and the presidency would negotiate all other matters. This elaborated, but still restrictive, mandate worked ‘with varying success’ in four meetings through 1998. Negotiations proceeded slowly, with the EU as a whole suffering ‘a loss of outward credibility and uniformity of position in the early stages that resulted from the refusal of member states to allow the Commission to negotiate on its own’ (Bail et al., 2002, p. 171). A third and final mandate was approved by the Council in December 1999, as negotiations were heating up during a critical period. This time, however, the negotiation mandate provided considerable more latitude and support for Community negotiators (Council 1999). The source of this shift can be found within the EU’s own ranks, as member state preferences congealed in support of an ambitious protocol. For one, three new states were installed as EU members by this time, lending weight to a pro-protocol position. The governments of Austria, Sweden, and Denmark were strong supporters of an international instrument for the regulation of biotechnology. France began to move its position from reluctant to supportive around the same time, pushed by public opinion but also spurred by the arrival of a new Green minister for environmental policy in 1998 (Tiberghien and Staars, 2004, pp. 24-8). Similarly, the German government changed to a Red/Green Coalition and took over the rotating Presidency of the Council of Ministers that same year. The German government became, in the words of the Commission negotiators, ‘much happier with our collective stand’ from that point on (Bail et al., 2002, p. 174). The combined effect of these preference shifts was stronger cohesion amongst the EU team and a greater willingness to allow the Commission latitude in the negotiations. The final negotiating mandate made clear the member states’ determination to conclude a protocol, provided flexibility on some issues while drawing a line under others, and signalled their support for the Community effort (Bail et al., 2002, p. 180). The practical effect of this new mandate was to hand the reigns of the EU negotiation team to Commission officials. They assumed the role of sole EU negotiator on most issues, albeit with input from member states during daily preparatory meetings. Using Meunier’s terms, the nature of the negotiation mandate became more extensive over time. The flexibility of the mandates evolved to give the Community agents more room for manoeuvre, with the Council principals placing less emphasis on the division of competences and more on general support for key aims in the protocol. Indeed, the Commission’s last mandate proposal was approved, largely unaltered, by member states in Council. The autonomy prescribed in the mandates was, as in all mixed competence cases, not particularly high. Council principals negotiate alongside the Commission agents during a negotiation, decreasing any chance of agency slack. At the same time, the convergence of member state and Commission preferences from 1998 until the end of negotiations meant a less antagonistic and more unified relationship. From that perspective, the low level of autonomy may not have had a significant influence on the delegation of competence variable. Finally, the authority enjoyed by the Commission was very high. With member states participating alongside Union negotiators, the ability to make promises to opponents and deliver on those promises was quite high. In short, the Commission enjoyed a fairly extensive mandate in the key stages of the Biosafety Protocol negotiations. Another important determinate of negotiation strength, according to the Meunier model, is the internal voting rule within the EU. According to Article 175 TEC, the treaty article related to environmental policymaking inside the EU, decisions are to be made using the qualified majority rule and not unanimity. Meunier argues that qualified majority voting, in conjunction with an extensive negotiation mandate, results in a ‘high’ supranational competence in international negotiations. The variables analyzed thus far confirm that the Biosafety Protocol negotiations represent just such a case. To fully assess the conditions for EU’s bargaining power, however, we must also examine how this supranational competence level relates to the nature of the negotiating context. Identifying the Negotiating Context The Meunier model predicts that a high supranational competence will have a different effect on the EU’s negotiating power, depending on the nature of the negotiating context. The Biosafety Protocol negotiations were clearly ‘reformist’, in that the position of the EU stood further from the status quo than its main opponent, the United States, throughout the negotiation. The EU, spurred on by an activist environmental policy unit in the Commission, close connections to NGOs, and under pressure from the public and press, ‘strove for a strong Protocol’ that covered an expansive set of issues (Cosbey and Burgiel, 2000, p. 6) in contrast to the US. Given the combination of a high supranational competence in a reformist negotiating environment, the Meunier model predicts the EU will have no bargaining power (see p. 8 above). Yet she concedes that qualified majority decision making by the EU in a reformist negotiation context has the effect of making the start of negotiations more likely, since only a majority of member states are needed to challenge the status quo. Opponents cannot drive a wedge through the EU delegation, owing to the lack of individual veto power, and thus the EU delegation remains cohesive (pp. 117-8). Meunier also suggests that the EU’s extensive competence in a reformist context will allow the EU more latitude to make demands, offers, and concessions without requiring constant member state approval. This will increase the likelihood of finishing a negotiation (p. 119). As a general rule, however, the Meunier model predicts that when the opponent wants a smaller departure from the status quo than the EU, outcomes will most likely equate with the opponent’s position and the EU will have little bargaining power. Are the Meunier model predictions born out in the case of the 2000 Biosafety Protocol negotiations? The next section probes the explanatory power of the model. Analyzing Process and Outcomes Initiating Negotiations Despite early discussions regarding a protocol, it was not immediately clear if and when actual negotiations would commence. Countries with advanced industrial resources in biotechnology (such as the US) opposed a formal biosafety agreement. In Europe, member state countries were largely divided over the subject in the mid- 1990s. Germany and France had clear reservations about the need for a protocol, having just finished the battle over biotechnology policy inside the EU. The creation of a stringent regulatory framework had aroused the concerns of biotechnology companies to which both governments were trying to make amends. The UK, Netherlands, Spain, along with most other EU members, were largely ‘centrists’. They were ambivalent on a protocol and had no strong feelings for or against. The last group of Sweden, Denmark, and Austria, were new EU members at the time but held preferences firmly in favour of initiating negotiations. Under unanimity voting rules, conservative outliers such as Germany and France probably would have curtailed any Commission proposals for a mandate to begin negotiations. Yet qualified majority voting was the established rule for environmental policy, making the consent of a majority of EU member states sufficient to approve a negotiation mandate. In October 1995, the Council agreed to the Commission’s first mandate proposal and thus signalled to the world the EU’s intention to actively participate (even if the full extent of their policy ambitions was still evolving). With the EU onboard, negotiations could begin in earnest even with the US maintaining its obstructionist position. The EU’s participation thus had the effect of kick-starting negotiations, drawing its recalcitrant opponent into talks just as the Meunier model predicts. Negotiating Process The three main negotiation issues and delegation positions formed early in the Biosafety Protocol talks. The first main issue concerned the scope of the agreement. Should a protocol apply to all GMOs, including those in processed commodities like food, animal feed, and pharmaceuticals, or only to ‘living’ GMOs such as seeds and plants destined for environmental release? The second issue involved the right of importing nations to reject GMO imports based on environmental or health grounds. Should such nations be allowed to use the ‘precautionary principle’ as a standard for risk assessment, a standard that some countries view as overly-precautious and a potential trade barrier? A third issue concerned the relationship between the Protocol and other international agreements. Would the Protocol be subordinated to the global trade order, or would it take precedence in future conflicts? The group of states with the most conservative answers to these questions staked out an early position. The ‘Miami Group’, led by the US, was a group of major grain exporting nations that objected to any protocol at the early stages of negotiation. Argentina, Australia, Canada, Chile and Uruguay united within the Miami Group to oppose an agreement, and even when the outlines of a protocol began to form, took very conservative positions close to the status quo. A second coalition of countries comprised the ‘Like-Minded Group’. Including many G-77 developing countries, this grouping took a more expansive view of the need for a protocol and advocated a protocol with sweeping scope. Other groups with less clear preferences included the Compromise Group, a collection of countries without obvious territorial or interest- based commonalities (Japan, South Korea, Mexico, New Zealand, Norway and Switzerland), and the CEE group of Central and Eastern European Countries. The latter was ridden by division, caused by a conflict between members with agricultural exporting interests and others with EU membership ambitions. As we might expect in today’s global political economy, the US and EU emerged as the main contenders in the Biosafety Protocol negotiation. Over the course of the negotiations, the relative positions of these two major trading partners bore a strong influence on outcomes. The US, occupying a position close to the status quo, rebuffed attempts at a broad and expansive protocol. The result was little movement in the first four negotiation sessions from 1996-1998 (Falkner, 2000). Lack of progress was evident even though consensus was growing amongst EU member states for a strong protocol. Subsequent revisions of the EU bargaining mandate in 1996 and 1999 revealed support for a strong position on the most important elements of a future protocol: a wide reaching instrument that allowed for ‘precautionary’ risk assessment and that took precedence over WTO rules. This growing consensus emerged in large part due to changing political circumstances in France and Germany. As stated above, the new Green minister in France since 1998 reversed the French government position on GMOs and made public statements in support of a strong Biosafety Protocol (Tiberghien and Starrs, 2004, pp. 27-8). In Germany, the Red/Green coalition came out in support of EU efforts as well and used its position as Council President to lend weight to the proceedings. With preferences converging amongst EU member states and Commission negotiators, the EU team rallied to take a more coherent and forceful role. Commission negotiators recall ‘engaging in bridge- and trust-building activities amongst other delegations’ and pushing a pro-protocol line more firmly through 1998 (Bail et al., 2002, p. 172, 174). Yet the Miami Group refused to budge. A text was produced for a 1998 meeting in Montreal, but the positions of the two key delegations remained far apart. In a bold move to force an agreement, the Danish chair of the proceedings, Veit Koster, convened a meeting in Cartagena, Colombia in February 1999 with the express intention of hammering out an agreement. Koster, along with the chair of the CBD signatory meetings, Juan Mayr, used a new negotiating structure to coax agreement between the parties. Small, informal groupings plus a private ‘Friends of the Chair’ session amongst key negotiators hoped to bridge the gulf between the parties. A new draft was circulated by Koster that contained solutions to some of the smaller issues associated with the protocol, but the three main issues – scope, informed consent, and compatibility – were left unresolved. In hopes of compromise, EU negotiators convinced developing country representatives to drop their extreme demands and back the EU position. Leading a united front, the EU put pressure on the Miami Group to compromise. But the Miami Group refused to continue, arguing that the ‘no deal is better than a bad one’. As the Meunier model predicts, the EU’s extensive delegation, and the use of qualified majority decision rules, resulted in cohesion amongst the EU ranks. The recalcitrant opponent was deprived of the option of driving a wedge through the European delegation. Yet, as bargaining theory implies, the US held the keys to the preservation of the status quo. Negotiations broke down in 1999, consistent with the Meunier model’s predictions. However, circumstances began to change in 2000 with a result that does not conform as well to the model. Negotiating Outcome A final attempt to complete the Biosafety Protocol negotiations commenced soon after the Cartagena failure. The EU sponsored a preparatory meeting in Vienna in July 1999 to discuss how to resume talks and to reach out to other negotiating groups. With the Like-Minded Group supporting the EU position since Cartagena, efforts were also made to bring the Compromise Group into a coalition. While the Compromise Group remained on its own, it did back the EU’s efforts to clarify the protocol’s relationship with the WTO (Earth Negotiations Bulletin, January 2000). Momentum grew for a resumption of talks in January 2000, in Montreal. Also becoming clear by this time was the shifting context in which the negotiations were taking place. In terms of policy developments, the EU was pushing its ‘precautionary principle’ and ‘savings clause’ in other venues and gaining support in such bodies as the Codex Alimentarius Commission. In terms of policy salience, public and press attention to the Biosafety Protocol negotiations rose dramatically in late 1999. This related closely to the failure of the December 1999 WTO talks. The ‘Battle in Seattle’ helped to politicize the protocol negotiations in two ways. First, within the WTO talks, the Miami Group attempted to shift responsibilities for transboundary GMO regulation from the CBD to the WTO. The moved failed as the EU and developing countries reacted strongly to what looked like an attempt to sabotage the Biosafety Protocol negotiations. This event ‘strengthened the resolve of EU ministers to provide full backing to the conclusion of the biosafety negotiations’ (Bail et al., 2002, p. 180). Second, the overall failure of the WTO talks in Seattle underscored the linkages between trade, environment, and social policy issues and shifted public attention toward the Biosafety Protocol negotiations. Hundreds of journalists attended the Montreal negotiation session in January 2000, while ministers resolved to participate personally in the negotiation process (Cosbey and Burgiel, 2000, p. 6). What had been a relatively low profile affair primarily amongst environmental experts was now recognized for its wide ranging implications across a number of different policy areas. In preparation for the final negotiation session in Montreal in January 2000, the EU’s negotiation mandate was updated for the final time (described above). Environment ministers threw their political weight behind the Commission negotiators, provided flexibility on the questions of scope, but encouraged a hard-line on the precautionary principle and the protocol’s relationship with trade rules. When talks commenced on 24 January 2004, 750 participants from 133 countries were present, but the key sticking points still related to the US and EU positions (Falkner, 2002, p. 20). Top EU officials, along with almost all EU environmental ministers, worked hard to put public pressure on the Miami Group. Moreover, they met daily with Community negotiators and pressed colleagues in the Miami Group for concessions. This joint effort added to the cohesion and influence of the EU’s position, a benefit that Commission negotiators refer to as the ‘added value’ of mixed competence issue negotiations (Bail et al., 2002, 171). In small negotiation groups, representatives narrowed discussion to the three core components of the protocol (Earth Negotiations Bulletin, January 2000). In terms of scope, the Like-Minded Group was persuaded to accept the EU’s ‘mid-range’ approach, which included all GMOs except those used for pharmaceuticals and for contained use in laboratories. The US continued to oppose the inclusion of non-living GMOs (commodities), but its position was weakening in the face of nearly unanimous opposition. In terms of the ‘precautionary principle’, day-and-night negotiations failed to narrow the gap. The US remained firm in its opposition to provisions that allowed nations to reject imports based on non-scientific grounds. The EU toned down its proposals, but rallied support for preserving the ‘precautionary principle’ as a standard of risk assessment. Finally, the relationship between the Biosafety Protocol and the global trade order proved one of the final sticking points. All parties except the Miami Group supported removing text from the draft protocol that subverted the regulations to WTO agreements. To avoid deadlock, opponents of this ‘savings clause’ proposed replacing it with more ambiguous language. Leading the reformist charge, Community negotiators had the US ‘isolated on nearly all the unresolved issues’ (Bail et al., 2002, p. 184). Negotiations dragged on until the early hours of the morning of 29 January 2000 as each side struggled to advance its positions. In the face of growing opposition and EU activism, however, the US and Miami Group found itself in an untenable position. By 4:40 am, their negotiators acquiesced, forsaking key provisions and agreeing to significant compromises. At 6:00 am, an international agreement on biosafety was in place (Earth Negotiations Bulletin, January 2000). In the final text of the Cartagena Biosafety Protocol, ratified by all member states by 2003, the EU achieved virtually all of its aims. Perhaps the most significant victory was the adoption of an international, binding protocol in the first instance, considering the initial unlikelihood of such agreement (Cosbey and Burgiel, 2000). On the key issues, the protocol encompassed a broad scope of GMOs, requiring advanced informed agreement between exporters and importers on both living seeds and plants, with a similar but less stringent regime for commodities. Moreover, the insertion of the ‘precautionary principle’, which the EU had helped to develop in the early 1990s, into the protocol was a major victory. Never before had the precautionary principle ‘been defined in the operational part of an international environmental agreement’ (Bail et al., 2002, p. 184), representing its successful ‘export’ as a new principle of international law. Finally, the status of the Biosafety Protocol in relation to WTO rules was left unresolved. However, most observers argue that with language in the Protocol indicating its provisions are ‘not subordinate’ to WTO rules, this result is ‘widely seen as a defeat for the Miami Group’ (Winfield, 2000). The 2000 Biosafety Protocol negotiations ended with a text that looked, in the eyes of one observer, ‘basically like an EU Protocol’ . Several members of the US Congress immediately lashed out at this result, with one US Senator complaining that ‘We stood up to [the EU and developing countries] in Seattle, but apparently not in Montreal’ . Nevertheless, as a major GM exporter, the US agreed to abide by the provisions of the agreement. 4. DISCUSSION What explains the discrepancy between the predictions the follow from the Meunier model and the case of the Biosafety Protocol negotiations? We identify three areas which require further exploration. The first concerns the characteristics of the model itself. Bargaining theory features prominently in Meunier’s approach, particularly the argument that the player with preferences closest to the status quo controls the bargaining outcome. Accordingly, a ‘conservative’ context, where the EU’s preferences lie close to the status quo, gives it a high degree of bargaining power. A ‘reformist’ context, where the opponent’s preferences lie close to the status quo, has the opposite effect. In the former context, the EU’s intransigence heightens the importance of different rules. Unanimity reinforces the intransigence of the most conservative EU state and makes the EU a stubborn—and powerful—opponent. Qualified majority voting mitigates the extremes within the EU member bloc, increasing the likelihood of an international agreement while still preserving most of the EU’s bargaining strength. Yet, in the latter context, rules play a less important role in explaining outcomes: no matter if the EU makes internal strategy decisions by unanimity or qualified majority voting, the fact that the opponent holds the ‘keys’ is the main explanatory element. We are left with fewer predictions as a result. We know only how negotiations might start, and what strategies are available to the opponent for dividing the EU vote. In effect, the model under specifies the predictions related to the EU in ‘reformist’ cases, unable to capture the nuance of EU bargaining power in such situations. A second explanation for these results might relate to the nature of the negotiation mandate. Meunier bases her explanation for ‘restricted’ versus ‘extensive’ mandates on three different attributes of the mandate, each closely related to the tenets of principal-agent theory. Member states can give Commission negotiators more or less instructions (flexibility), more or less oversight during negotiations (autonomy), and more or less ability to keep ratification promises (authority). However, the key features of a mixed competence negotiation may militate against a strict principal- agent approach to understanding delegation. For instance, during negotiations over issues of mixed competence, EU member state officials sit alongside Commission negotiators. Both national and supranational officials have an allocated set of responsibilities, although clearly they work together on almost all key issues. In light of the Meunier model, such circumstances beg the question: does the Commission have more or less flexibility, autonomy, and authority in cases of ‘mixed’ negotiations? Through careful analysis and a close reading of the Meunier approach, we determined that the Commission enjoyed an extensive mandate in the Biosafety Protocol negotiation. Yet future approaches might gain additional explanatory power when the strict principal-agent assumption is relaxed slightly, taking on board insights regarding negotiator skill, the participation of powerful politicians, and the preference alignment of EU member states. Do shirking and slippage (McCubbins and Page, 1987, pp. 410-11) in principal-agent models develop new meanings under mixed competences? Finally, we note a key distinction between exclusive and mixed competence issue negotiation that might bear on outcomes: ratification procedures. For trade issues, the ratification of agreements is limited to the Council of Ministers, whereas under mixed competences, national parliaments must also ratify the international bargaining outcomes. To predict international negotiations of ‘mixed competence’, negotiation models should account for the domestic ratification component more explicitly. During the Biosafety Protocol negotiations, ‘EU environmental ministers were under intense pressure and scrutiny from the public and the press, who had so recently flexed their muscles to great effect at the WTO’s failed Seattle ministerial’ (Cosbey and Burgiel, 2000). The need to ratify agreements at home, when domestic publics might play a stronger hand, makes the EU’s role in mixed agreement negotiations more sensitive to domestic influence. Returning to Putnam (1988), we note the importance of the strategic interaction between domestic and international factors in international negotiations. Since each actor possesses an effective veto over any agreement that changes the status quo, Putnam’s model implies that an agreement always lies at the intersection of each of the actors’ win set – the set of alternatives that each actor prefers to no agreement. Scholars modelling the EU in international negotiations, however, have paid little attention to national constraints in international negotiations. Clark et al. (2000) argue that the frequent success of the EU in its negotiations with the US is actually the result of the bargaining power that institutional arrangements grant its negotiators. The central hypothesis in that study maintains that understanding the US-EU trade relationship requires an appreciation of the ways in which ‘domestic’ political institutions shape the bargaining behaviour of international actors (see also Frieden, 2004). CONCLUSION This paper set out to improve our understanding of the EU’s behaviour in international negotiations. While most analysts focus on the EU’s role in global trade talks, we turned our attention to the EU’s bargaining power in ‘mixed’ competence negotiations such as those concerning global environmental policy. Using a model of EU trade bargaining power devised by Meunier (2000), we tested the empirical case of the 2000 Biosafety Protocol negotiations. Our analysis revealed that the Meunier model has considerable explanatory power for some of the outcomes described here, but incorrectly predicts that the EU will have no bargaining power in such negotiations. The Biosafety Protocol negotiations took place in a ‘reformist’ negotiating context, with the EU using qualified majority as an internal voting rule and Community negotiators enjoying extensive competence. As the Meunier model predicts, the EU succeeded in a drive to initiate talks. The evolution of negotiations, however, revealed the EU to be a tough bargainer, in contrast to the model’s prediction that the EU would lack negotiating power. As EU member state preferences, especially those of France and Germany, aligned with those of the Commission over time, the EU’s influence over the negotiations picked up speed. By 1999, the EU’s was dictating the terms of the negotiations, even in the face of several attempts by the US to scuttle negotiations entirely. The end result was a binding set of GMO provisions that enshrined most of the EU’s policy goals. In seeking to explain the discrepancy between Meunier’s predictions and our empirical results, we highlighted several aspects of the model that require further elaboration. At the same time, we pointed out key differences between negotiations where the Community has exclusive competence and where it has mixed competence. In the case of the latter, national public opinion might play a strong role in outcomes as the result of domestic ratification requirements. Moreover, the more nuanced relationship between member state and Community negotiators that exists in a mixed competence negotiation may also affect outcomes. Relaxing the strict principle-agent assumption, analysts may find space for more individual factors like bargaining skill and the role of the EU’s Presidency that might influence outcomes. Adjusting models to fit broader empirical parameters and more nuanced contexts usually comes at the expense of parsimony. This paper has offered up only one case of a mixed competence negotiation: clearly more are needed. If further testing reveals ongoing problems with the model, however, researchers might explore whether a completely new model of the EU’s negotiating power in such negotiations is necessary. Recent events demonstrate the importance of addressing such questions swiftly. Globally, attention is growing to the types of policy issues that fall within the EU’s shared competence category. The WTO debacle in Seattle (1999), and indeed the results of the Biosafety Protocol (2000), demonstrated the interconnectedness of social, environmental, and trade agreements and heightened their salience in the eyes of the public. Within the EU, the draft text of the Constitutional Treaty suggests that additional constraints (or opportunities) regarding the EU’s external bargaining power might be emerging. The draft Treaty preserves the rights of member states and Community to both engage in mixed negotiations, although the joining of the various EU pillars will mean that ‘Union’ will supplant ‘Community’ in all current usages. More important than legalities or semantics, however, is the proposed role for the European Parliament. Legal experts suggest that the Constitutional Treaty will strengthen the role of the Parliament in all external areas except the Common Foreign and Security Policy (De Witte 2003, Cremona 2003). Parliamentary consent will be required before Council approval of international agreements (Art III-325, indent 6), injecting a new dynamic into the EU’s institutional setting that may likely affects its external behaviour (van Schaik and Egenhofer 2005). Future efforts to model the EU’s negotiation power in international politics should pay close attention to these developments. References Bail, Christoph, Jean Paul Decaestecker, and Matthias Jorgensen. 2002. European Union. In The Cartagena Protocol on Biosafety: Reconciling Trade in Biotechnology with Environment and Development?, edited by C. Bail, R. Falkner and H. Marquard. London: Royal Institute of International Affairs. Bretherton, Charlotte, and John Vogler. 1999. The European Union as a Global Actor. London: Routledge. Cosbey, Aaron, and Stas Burgiel. 2000. The Cartagena Protocol on Biosafety: An analysis of results, IISD Briefing Note. Winnipeg, Canada: International Institute for Sustainable Development. Council of Ministers. 1995. Council Conclusions: Preparation of the Second Conference of the Parties to the Convention on Biological Diversity (Doc. No.10560/95, 11 October 1995). Brussels: The Council of the European Union. Council of Ministers. 1999. Council Conclusions: Preparation of the 'Enviroment' Council Meeting on 13, 14 December 1999. Negotiations on the Biosafety Protocol y (Doc. No.13344/99, 26 November 1999). Brussels: The Council of the European Union. Cremona, M. 2003. The Draft Constitutional Treaty: External Relations and External Action. Common Market Law Review 40:1347-1366. De Witte, Bruno. 2003. The Constitutional Law of External Relations. In A Constitution for the European Union: The First Comments on the 2003-Draft of the European Convention, edited by I. Pernice and M. P. Maduro. Nomos: http://www.ecln.net/elements/conferences/contributions_lisbon.htm. Earth Negotiations Bulletin. Various Issues. A Reporting Service for Environment and Development Negotiations. Winnipeg, Canada: International Institute for Sustainable Development. Epstein, David, and Sharyn O'Halloran. 1999. Delegating Powers. A Transaction Cost Politics Approach to Policy Making Under Separate Powers. Cambridge: Cambridge University Press. Evans, Peter B., Harold K. Jacobson, and Robert D. Putnam, eds. 1993. Double-Edged Diplomacy: International Bargaining and Domestic Politics. Berkeley, CA: University of California Press. Falkner, Robert. 2000. Regulating Biotech Trade: the Cartagena Protocol on Biosafety. International Affairs 76 (2):299-313. Falkner, Robert. 2002. Negotiating The Biosafety Protocol: the international process. In The Cartagena Protocol on Biosafety: Reconciling Trade in Biotechnology with Environment and Development?, edited by C. Bail, R. Falkner and H. Marquard. London: Royal Institute of International Affairs. Franchino, Fabio. 2002. Efficiency or Credibility? Testing the two logics of delegation to the European Commission. Journal of European Public Policy 9 (5):677-694. Franchino, Fabio. 2004. Delegating Powers in the European Community. British Journal of Political Science 34:276-293. Frieden, Jeffry A. 2004. One Europe, One Vote? The Political Economy of European Union Representation in International Organizations. European Union Politics 5 (2):261-276. Garrett, Geoffrey, and George Tsebelis. 1996. An Institutional Critique of Intergovernmentalism. International Organization 50 (2):269-99. Gupta, Aarti. 1999. Framing 'Biosafety' in an International Context: The Biosafety Protocol Negotiations. In Global Environmental Assessment Project, Belfer Center for Science and International Affairs. Harvard University, Cambridge, MA. Hix, Simon. 2005. The Political System of the European Union. London: Palgrave. Jupille, Joseph. 1999. The European Union and International Outcomes. International Organization 53 (2):409-425. Jupille, Joseph, and James A. Caporaso. 1998. States, Agency, and Rules: The European Union in Global Environmental Politics. In The European Union in the World Community, edited by C. Rhodes. Boulder, CO: Lynne Rienner. La Vina, Antonio G. M. 2002. A Mandate for a Biosafety Protocol: The Jakarta Negotiations. In The Cartagena Protocol on Biosafety: Reconciling Trade in Biotechnology with Environment and Development?, edited by C. Bail, R. Falkner and H. Marquard. London: Royal Institute of International Affairs. Leal-Arcas, Rafael. 2001. The European Community and Mixed Agreements. European Foreign Affairs Review 6:483-513. Majone, Giandomenico. 2001. Two Logics of Delegation: agency and fiduciary relations in EU governance. European Union Politics 2 (1):103-122. McCubbins, Mathew, and Talbot Page. 1987. A Theory of Congressional Delegation. In Congress: Structure and Policy, edited by M. McCubbins and T. Sullivan. New York: Cambridge University Press. Meunier, Sophie. 2000. What Single Voice? European Institutions and EU-US Trade Negotiations. International Organization 54 (1):103-135. Meunier, Sophie, and Kalypso Nicolaïdis. 1999. Who Speaks for Europe? The Delegation of Trade Authority in the EU. Journal of Common Market Studies 37 (3):477-501. Moravcsik, Andrew. 1998. The Choice for Europe. London: UCL Press. Nicolaïdis, Kalypso. 1999. Minimizing Agency Costs in Two-Level Games: The Controversies over Trade Authority in the United States and the European Union. In Negotiating on Behalf of Others, edited by R. Mnookin and L. Susskind. Beverly Hills: Sage. Patterson, Lee Ann. 1997. Agricultural policy reform in the European Community: a three- level game analysis. International Organization 51 (1):135-166. Patterson, Lee Ann. 2000. Biotechnology Policy: Regulating Risks and Risking Regulation. In Policy-making in the European Union, edited by H. Wallace and W. Wallace. Oxford: Oxford University Press. Pollack, Mark A. 1997. Delegation, Agency, and Agenda-Setting in the European Community. International Organization 51 (1):99-134. Pollack, Mark A. 2003. The Engines of European Integration: Delegation, Agency, and Agenda Setting in the EU. Oxford: Oxford University Press. Putnam, Robert D. 1988. Diplomacy and Domestic Politics: The Logic of Two-Level Games. International Organization 42 (3):427-60. Sbragia, Alberta M. 1998. Institution-Building from Below and from Above: The European Union in Global Environmental Politics. In European Integration and Supranational Governance, edited by W. Sandholtz and A. Stone Sweet. Oxford: Oxford University Press. Sheridan, Brian. 2001. EU biotechnology law and practice: regulating genetically modified and novel food products. Bembridge: Palladian Law Publishing. Tallberg, Jonas. 2000. Supranational influence in EU enforcement: the ECJ and the principle of state liability. Journal of European Public Policy 7 (1):104-121. Tallberg, Jonas. 2002. Delegation to supranational institutions: Why, how, and with what consequences? West European Politics 25 (1):23-46. Tiberghien, Yves, and Sean Starrs. 2004. The EU as Global Trouble-Maker in Chief: a Political Analysis of EU Regulations and EU Global Leadership in the Field of Genetically Modified Organisms. Paper read at Conference of Europeanists, Council of European Studies (CES), 13 March 2004, at Chicago, IL. United Nations Environmental Program (UNEP). 1995. Report of the Open-ended Ad Hoc Group of Experts on Biosafety, July 1995. UNEP/CBD/COP/2/7: Full text available at http://www.biodiv.org/biosafety/. United Nations Environmental Program (UNEP). 2000. Cartagena Protocol on Biosafety. Full text available at http://www.biodiv.org/biosafety/. van Schaik, Louise, and Christian Egenhofer. 2005. Improving the climate. Will the new Constitution strengthen the EU's performance in international climate negotiations? Brussels: CEPS Policy Brief 63/February. Winfield, Mark S. 2000. Reflections on the Biosafety Protocol Negotiations in Montreal. Full text available at http://www.biotech-info.net/biosafety.html. We use the term ‘European Union’ (EU) to describe the broader European institution, following linguistic convention after the entry into force of the Maastricht Treaty in 1993. However, we use the term ‘European Community’ (EC) with regards to the case study, as EC is still the official designation of the EU’s ‘first pillar’ where both trade and environmental policymaking takes place. As of 1 November 2004, 232 votes of at least 13 member states are required for adoption. In addition member states constituting the qualified majority should represent at least 62% of the total population (284,331,400). The report of the ‘Open-ended Ad Hoc Group of Experts on Biosafety’, held in Madrid from 24-28 July 1995, can be found in document UNEP/CBD/COP/2/7. The group’s task was to debate the possible components of a future biosafety. Unable to reach agreement, they produced a list of consensual and non-consensual elements of a protocol (Falkner, 2002, p. 8). Personal interview by the authors with an official from the Directorate-General for Environment, August 2004. See also La Vina, 2002. The exact division of competences was as follows: liability (Presidency), social issues (Presidency), compliance mechanism (Presidency), advanced informed agreement (Community), commodities (Community). Personal interview by the authors with an official from the Commission’s Directorate General for Environment, August 2004. This is not to suggest that the other groupings did not play an important role. On the contrary, the Like-Minded Group was another source of momentum for the negotiations. Two non-state coalitions also had a strong presence in the negotiations, namely the Global Industry Coalition (GIC) consisting of agricultural, food, and pharmaceutical companies, and an international coalition of consumer safety and ‘green’ groups (Gupta, 1999, p. 9). Yet, from a negotiation perspective, the US and EU demand our attention. Their relative positions and negotiating strength, based on their structural position in the global economy, has a more decisive impact in determining outcomes. Tim Galvin, US delegation member, quoted in ‘Montreal GML talks likely to be difficult – USDA’, Reuters News Service, 10 January 2000. See also ‘Talks on Biotech Food Today in Montreal Will See US Isolated’, The New York Times, 24 January 2000. ‘In the end, their participation was near perfect, as the commissioner [Wallstrom] and the 10 EU ministers played a decisive role in preparing the ground for the final agreement while staying away from the formal negotiating process’ (Bail et al., 2002, p. 182). ‘Talks on Biosafety Food Turn on a Safety Principle’, The New York Times, 28 January 2000. The Cartagena prefix was attached to the Protocol’s title in deference to the original plans to conclude the negotiations in Cartagena, Colombia in February 1999. ‘Brussels’ bad science will cost the world dear’, opinion-editorial by Henry Miller and Gregory Conko, Financial Times, 14 August 2003. See also Cosbey and Burgiel, 2000, p. 12. Personal interview by the authors with an official from the Directorate-General for Research, July 2004. This official was not part of the official negotiation team and disagreed with the policy goals of other members of the Commission. US Senator John Ashcroft, quoted in ‘Ashcroft Assails GE Accord’, St. Louis Times-Dispatch, 9 February 2000. Ashcroft and other Senators were upset that since the US is not a signatory to the CBD, the parent agreement of the Biosafety Protocol, the Protocol did not come before the Senate for ratification. US industry nevertheless had to comply with the rules when exporting to countries that are parties to the Protocol. ‘130 Nations Agree on Safety Rules for Biotech Food’, The New York Times, 30 January 2000. 20 30