1Mixed FTAs covering issues of shared competence between the EU and Member States have been a widely debated issue in the context of the Common Commercial Policy (CCP). When it comes to enforcement, these agreements are characterised by a two-fold layer of complexity. First, they require lengthy processes of national-level ratification, which are currently proving a source of considerable stalemate in the case of the CETA (Craig and De Búrca 2015: 83; Roederer-Rynning and Kallestrup 2017). Second, issue-areas falling under shared competence are amongst the ones proving most salient and politically intricate, accounting for the increased complexity of deep and comprehensive trade (Miklin 2013). This is particularly the case with investment protection and dispute settlement, which the ECJ Opinion 2/15 qualified as elements of shared competence (European Court of Justice 2017). In the face of similar hurdles, the need to preserve an efficient and credible trade policy has led the Commission to negotiate investment-protection and portfolio-investment chapters separately from main FTA texts, rescuing the latter from otherwise lengthy national ratification processes (Van der Loo 2017). This has been the case ever since EU-Singapore negotiations, when Opinion 2/15 was issued.

2Tested against the benchmark of credible enforcement, mixed agreements emerge as a pillar obstacle to an efficient CCP. Their ratification, in principle, offers a major window of opportunity for runaway EU Member States to simply reject a provisionally enforced FTA they deem suboptimal. At the time of writing, for instance, the CETA has been provisionally in force for nearly three years, but only about half of EU Member States have completed ratification (Council of the European Union 2020a). In light of this, the exit options offered by mixed FTAs justify the progressive exclusion of this category of agreements from the Commission agenda. Yet, we argue that the fact of considering mixed agreements as solely belonging to pre-Lisbon intergovernmentalism shows its analytical limitations. Mixed FTAs are a vivid sign that European integration is still an ongoing process, and can be expected to remain part of EU trade policy at least in the foreseeable future. Most recently, for instance, Member States have challenged the Commission’s strategic reticence towards negotiating mixed FTAs. They have reiterated that “[i]t is for the Council to decide […], on a case-by-case basis, on the splitting of trade agreements” (Council of the European Union 2018: 3).

3Against this background, this chapter focuses on a final theoretical puzzle in our research. It looks at the policy mechanisms sustaining the Commission in its attempt to ensure the survival of mixed FTAs in the face of national ratification processes that are difficult to complete. To that end, we make empirical reference to the CETA, provisionally in force since September 2017. On that ground, we preliminarily assess how the provisional application of a mixed agreement can trigger lock-in dynamics ensuring the de facto survival of a text even in the absence of full de jure ratification. We posit that a provisionally enforced mixed FTA can find its source of survival in lock-in dynamics arising in the context of EU implementation. The analysis is grounded on a notion of “lock-ins” understood as constraining forms of path dependency (Bardach 2009: 348). By “lock-in,” we mean an array of policy choices and processes through which a mixed FTA sanctions its de facto survival in the absence of de jure ratification. Before unpacking this notion further, we acknowledge that – at the time of writing – the number of empirical cases of mixed FTAs already provisionally in force is very limited, amounting essentially to the CETA (European Commission 2019c). Nevertheless, the CETA is a comprehensive case likely to inform upcoming implementation processes – notably, the EU-Mercosur FTA. Therefore, running multiple observations on the CETA can still shed analytical light on the policy reality of mixed agreements. We distinguish lock-in dynamics as operating at two main levels, involving i) official-level veto players within Member States and ii) advocacy players operating in Brussels.

5.1 Member States and official-level veto players

4The first lock-in mechanism pertains to official-level veto players, and more specifically to national governments and parliamentary assemblies ultimately in charge of completing mixed-type ratification. As previously discussed through this research, we conceptualise the aforementioned actors as shapers – rather than mere “sinkers” – of EU trade policy (Allee and Elsig 2017). This notion helps us recast the role national governments and parliaments can play in the ratification of mixed FTAs. Specifically, we expect national ratifiers to leverage their ultimate veto powers to request changes and safeguards, including after provisional EU enforcement. Yet, it is revealing to observe that even national-level ratifiers tend to act as first-movers when it comes to shaping ex post the content of a negotiated text. A relevant example is that of Wallonia, which acted as a first-mover in demanding a number of safeguards on the CETA investment court system (ICS) ahead of Council and EP ratification (Van der Loo 2016). This pushes us to consider how provisional enforcement means more than the simple “first step” in a lengthy and by no means guaranteed ratification process, but rather a watershed after which it is increasingly hard for domestic veto players to shape the terms of an FTA, and above all to sanction its suspension.

5Against this background, we assume that national veto players, albeit finding the terms of a mixed FTA suboptimal, will tend to be locked into an agreement once provisional enforcement is enacted. We ground our argument on the theoretical insights provided by Fritz Scharpf (1988) and his notion of “joint decision trap.” This framework was formulated to account for the survival of albeit suboptimal European policy regimes like the Common Agricultural Policy (CAP).1 Scharpf saw this stickiness as a function of the CAP being the result of unanimous decisions by EC Member States, and of the fact that the policy qualified as an “ongoing decision system” in which policies are implemented continuously. To him, in ongoing decision systems there will be a progressive deterioration of the “goodness of fit” of a policy to at least some decisionmakers taking part in the process (EU Member States) (257). Yet, Scharpf concludes, the fact that a policy rests on unanimity voting makes the costs of one veto higher, which would sanction the end of a policy regime supported by all, and therefore puts greater pressure on would-be veto players to agree on ongoing decisions, or at least not to reject them in the long run (265). Overall, this mechanism sanctions the path dependency of albeit suboptimal public policy regimes.

6Marking a novelty in EU trade studies, this research applies Scharpf’s (1988) framework to the provisional enforcement of mixed FTAs.2 We deem this framework valuable to the study of a policy field with highly redistributive implications like trade. To a greater extent than in regulatory policies, EU “winners” from trade will face costs in exiting a favourable FTA, despite the latter being potentially suboptimal to some fellow Member States. Conversely, negative impacts will in principle be more pronounced than in regulatory policies, making lock-ins of unfavourable actors particularly revealing of the mechanism of interest. In this light, EU trade policy meets the two fundamental conditions highlighted by Scharpf (1988) for the “joint-decision trap” to occur. First, provisional FTA enforcement can be considered as an instance of i) an ongoing decision system, in which, unless a formal veto is cast by a Member State, “non-agreement is likely to assure the continuation of existing common policies” – i.e. provisional enforcement (257, emphasis in original). In that regard, EU law foresees no specific term after which national ratifications should be completed. Further, the domestic impacts of trade policies often become visible in the long term, which implies that national political leaderships may come to realise that an FTA is suboptimal only at a later stage. Second, the provisional enforcement of FTAs is the result of ii) a unanimous decision by the Council, alongside EP consent. As a customary practice, although for issues directly pertaining to the CCP qualified-majority voting (QMV) would suffice, the Council approves FTAs by unanimity. Beyond formal legal requirements, “there is significant pressure to adopt agreements by consensus, or else endanger national ratification processes” (Keukeleire and Delreux 2014: 99). In light of this, if even one out of twenty-seven Member States turns against an FTA at a later stage, its decision will nullify a unanimous Council deliberation.

7Within EU Member States, policy mechanisms accounting for the “joint decision trap” are observable against the backdrop of national-level ratification. In that regard, no linear transition exists between the unanimous provisional enforcement of an agreement by the Council and national-level approval. As the case of the CETA attests, three years following unanimous consent by the Council, fourteen Member States are yet to complete their own ratification (Council of the European Union 2020a). Similar inconsistencies depend on two main dynamics in national trade politics. First, the nature and outcomes of national ratifications can be hard to predict. Domestic debates are often highly politicised, going beyond the specific terms and provisions of a treaty (Meissner and Rosen 2019). As a DG Trade official observed: “[within ratifications at the national level] questions […] sometimes are more fundamental. Even about you know: Why do we need trade? Is trade good? Is trade useful? It is really basic questions even about the economic model we live in.”3 Second, within democratic systems like those of Member States, national parliamentary majorities can change in ways not always favouring smooth policy processes at the EU level. In that regard, we expect the effects of the “joint decision trap” to be particularly evident when national majorities have changed from the day a Member State backs an FTA in the Council. This is especially the case if the new leadership has been critical of an agreement in the past, and yet prefers not to express a formal veto on it. For instance, populist and Eurosceptic majorities are particularly likely to have short time horizons, seeking easy political gains (Smith 1989). Yet, we assume they will be less prone to do so if vetoing an FTA means going against the fait accompli of a unanimous deliberation, putting an end to potential gains from trade by other EU Member States.4

8Overall, national reticence towards expressing a final word on an FTA can serve Commission strategic needs in at least two respects. First, it maintains a negotiated text de facto implemented through provisional enforcement. In the case of the CETA, the agreement is fully operative with the exception of investment protection and portfolio investment chapters (Council of the European Union 2016a). Second, reticence to put an FTA to a national vote can prevent the revival of contestations amid national constituencies, by not creating critical junctures at which intense debate may re-spark. That is particularly relevant considering that national contestation is especially intricate for a supranational bureaucracy like the Commission to tackle. Parliamentary debates often go beyond arguments on specific FTA provisions, posing more generalised threats to the legitimacy of EU trade policy tout court.

5.2 Civil-society veto players

9Lock-in dynamics also operate at the level of private veto players taking part in the implementation and monitoring of an FTA once this has provisionally entered into force. We focus in particular on civil-society advocacy organisations, for these have been pivotal in mobilising against FTAs recently concluded by the EU. As argued by De Ville and Siles-Brügge (2017), “normative, values-based concerns (instead of purely economic or material motivations) and the role of NGOs will have to feature more prominently in future studies of EU trade politics” (1496). This is not to deny the importance of other business interest groups, which can still prominently affect the content of FTAs. Yet, besides specific divergences, business groups tend to operate within the same paradigm of an economic-minded bureaucracy like DG Trade. Conversely, civil-society organisations often confront EU trade policy with more fundamental social, environmental, and ethical concerns that can clash with the overwhelmingly economic rationale of Global Europe.

10Against this background, it is analytically relevant to explore how advocacy groups in principle critical of an FTA can become involved in its implementation, by playing a monitoring and advisory role. Domestic Advisory Groups (DAGs) are instances of similar processes of civil-society inclusion into enforcement by the Commission. Crucially, the visibility and leverage advocacy can gain as part of implementation is essentially conditional upon FTAs being in force, which can be seen as a lock-in mechanism making these actors paradoxically dependent on agreements they would in principle oppose. Scholarship has traditionally conceived lock-ins as a function of increasing returns and positive feedbacks (Bardach 2009; Pierson 2000). In the specific case of EU trade advocacy, we suggest that lock-ins are better conceived in terms of the comparatively high transaction costs advocacy groups would incur if they were to lobby beyond the scope of the “coalition opportunity structures” provided by the Commission during implementation (Sabatier and Weible 2007: 199-201). This is particularly the case for advocates whose policy interests are rather homogeneous, which allows them to obviate collective-action problems when auditing Commission policies (Olson 1965). For instance, advocacy actors within the DAGs consist of trade unions and NGOs pointing at the potential erosion of EU environmental, labour, and social standards as a result of market liberalisation (European Economic and Social Committee 2020b). Providing that common goals exist, advocates are thus arguably better off being part of an FTA implementation mechanism, rather than keeping themselves outside of it, by acting as fragmented constellations. To further our understanding of lock-in dynamics, we analyse empirical data related to the CETA DAG. The latter is a particularly relevant case to look at, since, besides its general monitoring functions, it is also supposed to take part in the review of the CETA’s TSD chapter, as provided for by the Joint Interpretative Instrument (JII) adopted by the EU and Canada5 (Council of the European Union 2016c).

11We distinguish three types of transaction costs: i) institutional, ii) epistemic, and iii) resource-based. These are discussed in turn. First, in terms of institutional membership, “the transaction costs of participating in a coalition are relatively low compared with other forms of collective behaviour because of shared belief systems, high trust, and willingness to distribute costs fairly” (Sabatier and Weible 2007: 197, emphasis added). In this regard, implementation platforms like the DAGs represent key institutional windows of opportunity for advocacy veto players to gain visibility on trade-related issues. A CETA DAG member observed: “[the DAG] really is a test for us, because we want to see how much leverage we can have by being in there.”6 Advocates benefit from the proximity of their mandates, often running preparatory meetings ahead of plenary sessions.7 Requests and opinions can be directly voiced to the Commission, and result in officially circulated reports. Conversely, it would be reasonably harder for citizen groups operating outside of these fora to enjoy equal influence and visibility. Forum-shopping in other Brussels venues like the EP Committee on International Trade (INTA) makes less sense when an FTA has already been ratified by MEPs. Second, trade policy, owing to its epistemic complexity and technicality, requires a considerable degree of expertise in view of effectively monitoring treaty enforcement. In this respect, EU implementation fora are an opportunity for advocates to share information and better keep track of highly technical discussions. Third, we expect Brussels-based advocates to face significant resource constraints if they were to exercise their lobbying beyond the scope of these implementation institutions. That would require additional financial disbursements and staff in order to, inter alia, carry out autonomous FTA monitoring and lobbying, likely exceeding what is normally allocated by single organisations to trade policy. Further, engaging individually in auditing activities would greatly diminish the leverage otherwise exercised by an advocacy coalition vis-à-vis the Commission.

12Overall, empirically assessing the ways advocacy actors can become locked into the implementation of mixed FTAs contributes to this research on two main grounds. First, it can shed light on the policy mechanism aiding the Commission in its strategic endeavour of getting a mixed FTA de facto enforced. Second, making sense of advocacy lock-ins contributes to recasting traditional views of veto players in political science, emphasising their contributions, rather than merely their challenges, to FTA provisions (Allee and Elsig 2017). Finally, from a Commission point of view, we suggest that the inclusion of civil society into implementation contributes to enhancing the transparency of EU policymaking (Majone 1998). In that regard, a Commission trade official observed how “mak[ing] the implementation inclusive and transparent […] shows that there is no sort of hidden agenda, no secret meetings and things being decided to circumvent the rules or the rights of governments to regulate, or to impinge on regulatory sovereignty.”8 With the aim of empirically exploring dynamics discussed in sections 5.1 and 5.2, we formulate the following hypothesis (H3):

13H3(a): If a mixed FTA is provisionally enforced by unanimity, lock-in dynamics will prevent national parliaments from formally rejecting the agreement.

14H3(b): If the Commission involves private veto players in the provisional implementation of a mixed FTA, lock-in dynamics will disincentivise these actors from lobbying for the rejection of the agreement.