Multilevel Trade Policy in the Joint‐Decision Trap? The Case of CETA

Wallonia’s refusal to ratify CETA in October 2016 suggests that multilevel trade politics may increasingly be subject to the pitfalls of joint decision‐making, or even a joint‐decision trap. This article, however, presents a more nuanced per‐ spective that builds on a comparative analysis of intergovernmental configurations that underpinned constituent units’ participation in CETA in the four formal federations Canada, Belgium, Germany


Introduction
Trade policy has emerged as a new domain for the comparative study of intergovernmental relations (Broschek, 2023a;Broschek & Goff, 2020a;Egan & Guimarães, 2022;Freudlsperger, 2020;Paquin, 2022).At first glance, this is surprising, considering that the power to initiate and conduct trade agreement negotiations has been a prerogative of higher-level governmental tiers.In most federations, authority over trade was formally vested at the federal level (Watts, 2008, p. 195).The EU as a quasifederal system is no exception.From Rome to the Treaty of Lisbon, the community method was further extended, consolidating the supranational level's formal jurisdiction (Garcia, 2020;Woolcock, 2015).
Although allocating powers within a federal system is notoriously conflict-laden, furnishing the federal level with exclusive trade policy jurisdiction was rather uncontroversial.Creating internal markets and promoting economic welfare was a key goal of modern state-building and political unification (Bartolini, 2005;Egan, 2015;Hueglin & Fenna, 2015).This has changed, however, over the past decades.Trade policy has not only become more contested by civil society actors but also by regional and constituent units.First, the scope of trade policy agreements has expanded significantly since the 1970s and 1980s (Baccini, 2019;Baccini et al., 2015;Young, 2016).With the inclusion of non-tariff trade barriers, trade policy began to affect jurisdictions of lower-level tiers, directly or indirectly (Kukucha, 2008).Second, in Europe, trade policy has become increasingly politicized in recent years (De Bièvre & Poletti, 2020;Duina, 2019;Leblond & Viju-Miljusevic, 2019).Both factors variously mobilize governments from lower-level tiers to shape trade politics and policy.
CETA offers a fascinating glimpse into the potential implications of this trend: That constituent unit would play a role in this agreement became evident before negotiations started.It was the provincial government of Quebec-not the federal government-who took the initiative to relaunch negotiations of a bilateral trade agreement in 2007.However, the EU responded only reluctantly to the Canadian initiative.One important reason was that the European Trade Commissioner Peter Mandelson perceived the Canadian provinces as a potential obstacle to successful trade negotiations (Schram, 2019, p. 97).Ironically, it was regional units within the EU, rather than the Canadian provinces, that impeded the ratification process.
This raises the question: Does the rise of regional units as stakeholders in trade policy-making create a joint decision trap with the potential to jeopardize future trade agreement negotiations?Regional opposition to recent trade agreements suggests that this may be the case.One expectation derived from the Belgian case is that regional units will continue to participate more actively in trade policy-making.Their new self-assertive role has at least the potential to create a joint decision trap, especially under heightened politicization (Bollen et al., 2020).Indeed, Belgium still has not ratified CETA.At the same time, however, Germany-the prototype of joint decision-making-has.Although several Land governments articulated concerns about current trade policy agreements or were even opposed to CETA, they eventually ratified the agreement in the Bundesrat, the second chamber, in December 2022.
I use CETA as a case study to advance the following argument.First, the emergence of joint decision-making as one particular mode of intergovernmental policy coordination, among others, is contingent upon federalism's historically established institutional configuration.The analysis reveals that while joint decision-making does not apply to the Canadian case, it can capture European multilevel trade policy-making under certain conditions.More specifically, only if trade agreement provisions transcend the exclusive jurisdiction of the EU does joint decision-making surface as a mode of intergovernmental coordination between the supranational and member state level and, potentially (but not necessarily), between the member state level and regional units.
Second, I focus in particular on the implications of joint decision-making for trade policy.In line with existing research, my findings confirm that formal veto power alone cannot explain ratification failure (Benz, 2016(Benz, , 2020;;Benz & Broschek, 2013a;Benz et al., 2016;Scharpf, 2011;Scharpf et al., 1976).What matters are meaningful opportunities for regional units to be included in the trade policy cycle prior to the ratification itself (for trade policy, see in particular Freudlsperger, 2020).I argue that despite the importance of joint decisionmaking as a mode of intergovernmental coordination, especially in the EU, it is rather unlikely that multilevel trade policy-making will increasingly be subject to the joint decision trap.

Joint Decision-Making in Multilevel Trade Policy: Theoretical Expectations
Joint decision-making represents a distinct mode of intergovernmental coordination.Although research has identified variations of joint decision-making (Benz et al., 2016;Heinz, 2012;Scharpf, 2011), on a general level, this mode encapsulates a configuration where governmental tiers are institutionally required to collaborate in decision-making processes.Scharpf et al. (1976) originally identified joint decision-making as the main mode of intergovernmental coordination in Germany, concluding that it entails significant-yet not insurmountable-institutional constraints for effective policy-making.The notion of a joint decision trap, in particular, suggests that these constraints have the potential to endogenously paralyze the political system (Scharpf et al., 1976, p. 54).Although this may cause frustration among policy-makers over time, this institutional configuration is difficult to disentangle as the short-term benefits, particularly the power to block (or threaten to block) political change, don't incentivize long-term reform (Scharpf, 2006).In the 1980s, Scharpf then demonstrated the potential for extending the concept beyond the German case, and its applicability to European politics (Scharpf, 1988).But is joint decision-making also the dominant intergovernmental mode for trade policy coordination, and is trade policy-making subject to a joint decision trap?
Comparative federalism research offers different models for analyzing intergovernmental relations comparatively (Behnke & Mueller, 2017;Benz & Broschek, 2013b;Bolleyer, 2009;Schertzer et al., 2018;Schnabel, 2020;Simmons, 2017).To analyze multilevel trade policymaking, I use a simple distinction of four modes of intergovernmental policy-making that can be mapped on a continuum between the two institutional principles that underpin, in various ways, every federal system (see Table 1): The separation of powers (or self-rule) on the one hand; power-sharing (or shared rule) on the other (Broschek & Goff, 2020b, p. 15;Skogstad & Bakvis, 2012, p. 7).
Federal systems that emphasize a separation of powers offer opportunities for unilateral action.This can be harmful if the decisions taken by one jurisdiction transgress the boundaries of others, generating negative externalities (Bednar, 2009).Unilateralism, however, can also take other, less antagonistic forms like competition or mutual adjustment (Benz, 2012;Scharpf, 1997).In the case of trade policy-making, the question Table 1.Modes of intergovernmental relations.

Unilateralism
ConsultaƟon CooperaƟon Joint decision-making is whether higher-level governmental tiers use their formal authority to act without considering constituent units' preferences.
Consultation is one step closer on the continuum toward power-sharing.Here, higher-level tiers with authority to make decisions invite constituent units' feedback on trade policy agreements in different stages of the policy-making process.Although they are free to ignore concerns, a consultation process can generate soft pressure on higher-level tiers, especially in times of polarization when trade policy is more salient.Cooperation, then, represents a stronger form of collaboration.The federal level and constituent units work together to shape trade policy provisions that affect the latter.However, intergovernmental collaboration remains voluntary.If no agreement can be reached, the federal level still has the discretion to withdraw and decide on its own terms.Finally, joint decision-making only applies to intergovernmental configurations where the higher level has no such exit option.In this case, the final decision to ratify a trade agreement is contingent upon either unanimous or significant support from constituent units (Scharpf, 1988(Scharpf, , 2011)).
Constituent units seek to shape trade policy for different reasons (Broschek, 2023a(Broschek, , 2023b;;Broschek & Goff, 2022;Freudlsperger, 2018Freudlsperger, , 2020)).As trade policy agreements began to impinge upon their jurisdictions through behind-the-border measures, constituent units in Europe became increasingly concerned about a creeping loss of authority and their ability to provide key social services.The politicization of trade policy since about 2013 has further amplified this trend.While concerns over authority are also a driving force in other federations outside Europe (for the US, in particular, see Freudlsperger, 2023;Jaursch, 2023), regional economic preferences are often more dominant (Broschek & Goff, 2022;Kukucha, 2008).Regardless of what type(s) of preference motivate(s) their engagement, the institutions of federalism position both constituent units and the federal level in different ways to address the need for trade policy coordination.
Accordingly, I formulate two different theoretical expectations regarding the emerging mode of intergovernmental trade policy-making: First, the more a federal system aligns with the separation of powers in the field of trade policy, the broader the corridor for different modes of intergovernmental trade policy coordination, which can switch between unilateralism, consultation, or cooperation, depending on the preferences of the federal level.
Second, the more a federal system aligns with powersharing in the field of trade policy, the narrower the corridor for different modes of intergovernmental coordination: Trade policy coordination is more likely to take the form of joint decision-making.
The mode of intergovernmental policy coordination has important implications for the power of constituent units to shape trade policy content.Federal sys-tems that lean more towards a separation of powers tend to favor higher-level governmental tiers since they ultimately have formal jurisdiction over trade policy.Therefore, they enjoy considerable freedom regarding their responsiveness to constituent units' preferences and how they seek to engage these actors through consultation or even cooperation.Ratification failure, therefore, is unlikely as long as the federal level supports a trade agreement.
By contrast, joint decision-making furnishes constituent units with veto power.The federal government is not able to ratify an agreement without constituent units' consent.Accordingly, ratification failure is a potential scenario.Two conditions are crucial in this respect.
First, if a trade policy agreement is politicized, joint decision-making will likely result in a stalemate.However, research has shown that deadlock is less common than one would expect as political actors, in anticipation of this problem, seek escape routes (Benz et al., 2016;Falkner, 2011;for trade, in particular, Gheyle, 2022;Scharpf, 2011).Second, the system of intergovernmental relations is of particular importance here.A dense, highly institutionalized system of intergovernmental relations offers channels for policy coordination between and among governments to address concerns early on and to negotiate positions based on common ground.It also strengthens executives and their institutional selfinterest, shielding them to a certain degree from the pressures of party politics.Both can contribute to mitigating politicization and help avoid ratification failure in a configuration of joint decision-making.

Case Study Design
When does joint decision-making emerge as a mode of trade policy coordination (as opposed to other intergovernmental modes), and what conditions promote or mitigate stalemate and, eventually, the joint decision trap?The following study uses CETA as a case to examine the configuration of intergovernmental trade policy coordination.As one of the most encompassing trade policy agreements of our times (Fafard & Leblond, 2013;Kukucha, 2013;Schram, 2019, p. 70), CETA represents the general trend of increased intergovernmental trade policy coordination, despite the fact that the federal (or supranational) level enjoys formal exclusive jurisdiction.First, CETA entailed provisions that affected, directly or indirectly, jurisdictions of the provinces and member states, including those of regional or constituent units in decentralized or federal states.Second, in several European member states, CETA, like other post-Lisbon trade agreements, was highly politicized, especially between 2014 and 2017 (De Bièvre & Poletti, 2020;Gheyle, 2019;Leblond & Viju-Miljusevic, 2019;Meunier & Czesana, 2019).Increased issue salience and contestation manifested themselves not only in mass demonstrations in major European cities like Berlin, Brussels, or Vienna but also in sustained lobbying efforts of civil society organizations and political party organizations calling upon regional units-such as the German Länder-to block the ratification of CETA and other pertinent trade policy agreements (Broschek et al., 2020;Gistelinck, 2020;Siles-Brügge & Strange, 2020).
CETA thus offers an excellent test case as it allows for examining the fundamental challenge for federal systems-even if jurisdictions are formally exclusive, policy interdependencies create a need for intergovernmental coordination (Benz, 2020;Benz & Broschek, 2013b;Bolleyer, 2009;Bolleyer & Thorlakson, 2012).The EU-Canada comparison entails two perspectives: a comparison across federal systems (Canada and the EU) and a comparison within the EU (the EU and its federal member states; see Fossum & Jachtenfuchs, 2017).Accordingly, I conceptualize not only Canada but also the EU as a federal system or, more specifically, as a multilevel federation (Keating, 2017).Rather than understanding the EU as a distinct form of an international organization established by its member states, the comparative federal perspective suggests that it can be analyzed as an instance of federal state-building (Fossum & Jachtenfuchs, 2017, p. 471).
Conceived this way, the EU comprises at least three orders of government: supranational, member-state, and regional levels.While the relationship between the EU and its member states has always been a constitutive element of the EU polity, the regional level has emerged because of territorial rescaling.Its role is highly contingent upon the member state under consideration (Hooghe et al., 2010;Jeffery & Wincott, 2010;Keating, 2017;Tatham, 2018).For the purpose of this study, I focus exclusively on EU multilevel trade policymaking in the three member states that are formal federations: Austria, Belgium, and Germany.I expect that only in federal systems do constituent units have a potentially significant, constitutionally entrenched role in foreign and European affairs.By contrast, Canada represents a federation comprising two orders of government (10 provinces and three territories).
One point of entry into analyzing the EU through a comparative federalism lens is intergovernmental relations (Fabbrini, 2017;Fossum & Jachtenfuchs, 2017).At this empirically observable intersection of political arenas, governments representing different constituencies interact to address the challenge of policy coordination.From a comparative perspective, Canada's federal architecture is almost exceptional in epitomizing a separation of power (or self-rule), an institutional characteristic further accentuated through the combination with Westminster democracy (Broschek, 2020(Broschek, , 2021)).Accordingly, modes of intergovernmental relations are expected to be flexible, with a comparatively strong role of the federal government.Joint decision-making is a rare intergovernmental mode in Canada, and it does not capture federal-provincial coordination in trade policy.The federal government controls the entire trade agreement formation process, including the ratification of agreements.From a formal procedural point of view, the provinces would only enter the trade policy-making process in the implementation phase when agreement provisions require the adjustment of domestic law within their jurisdictions (Paquin, 2020).
By contrast, the EU's three-tiered federation combines a separation of powers and power-sharing.Efforts to strengthen supranational self-rule in trade policy-making through the Lisbon Treaty have been constrained-again-through the scope and depth of recent trade policy agreements.The EU Commission reluctantly declared that it would consider CETA a mixed agreement.As a result, the agreement requires ratification in all member states.The Commission insisted that, from a legal point of view, it considers the agreement as falling within exclusive EU competence (European Commission, 2016).This assertion proved to be wrong.In May 2017, the European Court of Justice ruled in its "Opinion 2/15" on the European Union-Singapore Free Trade Agreement that provisions covering non-direct foreign investment as well as dispute settlement mechanisms are not within the EU's jurisdiction (European Court of Justice, 2017).
Consequently, joint decision-making unambiguously captures the institutional configuration of trade policy configuration in the EU when we look at the relationship between the EU and its member states.Although the Lisbon Treaty provided a space for member state involvement through the Council (Garcia, 2020), member states now have a veto through the ratification process in the case of mixed agreements.What is less clear, however, is if the inclusion of regional units prompts a twofold multiplication of this configuration in EU multilevel politics (Hrbek, 1986) and whether this additional intergovernmental layer contributes to a joint decisiontrap in trade policy.Formally, at least, only the Belgian Regions and Communities and the German Länder have a constitutional right to approve the ratification of mixed agreements.
CETA has been in effect provisionally since September 2017.The Council of the EU eventually authorized and signed CETA on a provisional basis, and subject to several exemptions that are included in the two formal decisions (Council decision 2017(Council decision /37, 2017;;Council decision 2017Council decision /38, 2017)), the Joint Interpretative Instrument as well as 38 statements submitted by the Commission, the Council, and several member states.As of April 2023, 18 member states (including the UK as a former member state) have ratified the agreement, while the ratification of 10 member states is still pending.Constituent units in all four federations were variously involved in the trade policy-making process, contingent upon the institutional configuration of federalism that positioned them in different ways.As the summary agreement.In the following sections, I will examine the institutional configuration that produced different modes of intergovernmental coordination to include constituent units and their implications for CETA.

Austria: Between Consultation and Unilateralism
The institutional foundations of Land government participation in trade policy were laid in the early 1990s before Austria joined the EU.The Länder anticipated further constraints through EU accession on their already limited ability to regulate and provide key social services.Eventually, they accepted a transfer of legislative authority to the supranational level in exchange for constitutionally entrenched participation rights in European affairs within the domestic institutional framework.The Austrian Länder reached an agreement with the federal level to create a new procedure for Länder participation, the so-called Länderbeteiligungsverfahren, constitutionalized through the new Article 23 of the Federal Constitution.This new provision guaranteed the Länder not only timely access to information regarding future and ongoing negotiations on the supra-and international level but also the right to submit two types of resolutions in the consultation process.
The first is the so-called uniform opinion.The Länder can invoke this provision whenever European negotiations affect their jurisdictions.It is binding insofar as the federal government is only allowed to deviate from the Länder position if it can make a compelling argument related to supranational or international constraints.The federal government is also obliged to inform the Länder in writing about these circumstances.The second type entails a more general opinion pertaining to questions not directly affecting Land competencies.These general resolutions are not binding but often have a polit-ical effect, especially if supported unanimously (Broschek et al., 2020;Bußjäger, 2006).
The Länder articulated their concerns through resolutions in accordance with Article 23.It is noteworthy that CETA was not the first agreement that mobilized the Länder.They had already engaged in other trade-related matters, most notably the General Agreement on Trade in Services (GATS) in 2003.However, increased trade policy salience and politicization resulted in a heightened activity level of Land governments (and parliaments) since 2013 (Broschek, 2023b;Broschek et al., 2020).Between 2014 and 2017, the Austrian Länder adopted three uniform opinions and one general opinion on trade agreement negotiations, which did not exclusively focus on CETA.Rather, Land governments initially began to address Transatlantic Trade and Investment Partnership (TTIP) but extended their positions within a very short time frame to also cover CETA and the Trade in Services Agreement (TiSA).
Issue salience was a catalyst for rather than a cause of these activities.The main concerns articulated in the four opinions were very similar to those raised in the context of GATS in 2003.Although the Länder repeatedly acknowledged their support for trade liberalization, they were highly critical of the depth of certain provisions entailed in recent trade agreements.In essence, the Länder consistently rejected provisions that they expected would limit their constitutional authority, most notably investor-state dispute settlement mechanisms (ISDS), the negative list approach, and the creation of committees to promote regulatory regulation.In this respect, they emphasized their constitutional authority to provide key public services such as water supply, waste management, infrastructure, education, and health.Moreover, Land governments expressed concerns that provisions facilitating deep economic integration would contribute to the erosion of established regulatory standards in consumer protection, animal welfare, environmental standards, and social services (see also Broschek & Goff, 2022).
Accordingly, the Länder took advantage of their constitutional right to launch intergovernmental policy coordination through consultation.It is remarkable that all nine Länder were able to formulate unanimous resolutions across party lines through their horizontal peak organization, the Landeshauptleutekonferenz (Council of Land Governors, LHK), and opposed CETA in its current form.The Austrian Bundesrat echoed these concerns.
In May 2016, the Standing Committee for European Affairs recommended considering Land governments' concerns and called upon the federal government to refrain from a provisional implementation of the agreement (Österreichischer Bundesrat, 2016).However, in the case of trade policy, the separation of powers prevailed.The Austrian Bundesrat not only lacks the veto power of the German Bundesrat, but it is also not composed of Land governments.Its members are elected by the Land legislatures.On 14 October 2016, the federal Vice Chancellor Mitterlehner sent a letter to the secretariat of the LHK, informing Land governments that, in the federal government's view, all concerns had been addressed (Bundesministerium für Bildung, Forschung und Wissenschaft, 2016).The federal government eventually used its prerogative and ratified the agreement unilaterally in 2018 with majority support in both chambers, the Nationalrat and the Bundesrat.

Germany: Joint Decision-Making Without the Joint-Decision Trap
At first glance, intergovernmental trade policy-making seems to reveal important similarities to the Austrian case.First, recent trade agreements did not trigger the German Länders' trade policy involvement, as they have participated in this field since at least the late 1990s.Second, like the Austrian Länder, they adopted traderelated policy resolutions much more frequently due to heightened politicization and the negotiation of highprofile trade agreements such as TTIP, CETA, or TiSA since 2013.Third, the main concerns expressed by the Austrian Länder essentially mirrored those identified by the German Länder: Although they support trade liberalization in principle, the Länder argue that provisions such as ISDS or regulatory cooperation represent a potential threat to their capacity to regulate, to provide key public services, and for social and environmental standards in general.
Beneath the surface, however, the German case differs profoundly from Austria's.While the federal level is formally obliged to consult with Land governments, their participation is de facto more powerful as they must eventually ratify mixed trade agreements directly through the Bundesrat.The intergovernmental mode, therefore, is joint decision-making: the Länder have the authority to block the ratification of trade agreements, which was a realistic scenario in the case of CETA (Broschek et al., 2020).The institutional foundations for this strong form of power sharing were laid in the late 1980s and early 1990s in the context of the Single European Act and the Maastricht Treaty, both of which required approval by the Bundesrat.The Länder were able to gain major concessions from the federal government, entrenching new participation rights in European affairs through the Gesetz zur Einheitlichen Europäischen Akte in 1986 and the new Article 23 Basic Law in 1992 (Kropp, 2010).
Like in Austria, resolutions on CETA emerged from the Länders' engagement with TTIP.Overall, the Länder adopted five resolutions on trade policy between 2013 and 2017.The first two resolutions focused directly on TTIP (Deutscher Bundesrat, 2013Bundesrat, , 2014)), while the others addressed trade policy and trade policy agreements such as CETA more generally.The most encompassing and detailed resolution was issued in 2015 (Trade Policy for All: Towards a Responsible Trade and Investment Policy) and entails 39 points (Deutscher Bundesrat, 2015), which were updated in 2017 (Deutscher Bundesrat, 2017).
Although these resolutions indicate general support for EU trade policy, Land governments were far less unified in terms of whether they would support CETA ratification than their Austrian counterparts.Land coalition governments led by Christian democrats were generally in favor, while social democratic and Green party-led governments were either opposed to CETA in its current form or undecided.Since abstention counts de fact as an opposed vote in the Bundesrat, the ratification remained in limbo (Broschek et al., 2020).
Three factors facilitated the agreement and, eventually, the successful ratification in the Bundestag and Bundesrat in December 2022.First, the federal government was able to buy time, thanks to a pending ruling of the German Federal Constitutional Court on the constitutional conformity of the agreement.The case was resolved with a decision in February 2022, which opened the door to the ratification process.Second, by that time, the period of increased politicization and issue salience of trade policy had already passed.Other issues, most notably the Covid-19 pandemic and the Russian invasion of Ukraine, dominated the political agenda.Third, and perhaps most important in this context, German federalism features a highly differentiated and institutionalized system of intergovernmental relations.The resolutions on trade, in particular, were not simply an outcome of Bundesrat debates.Rather, they emerged from intense horizontal coordination within several sectoral peak organizations over the years, most notably the Conference of Ministers for the Environment, Economic Affairs or Consumer Protection (Broschek et al., 2020, p. 223), which were then also discussed vertically in close collaboration with the federal level.As Freudlsperger (2018Freudlsperger ( , 2020) ) has shown, these opportunities for close, ongoing intergovernmental coordination are crucial to facilitate a less conflictual and more problem-oriented mode of interaction in trade policy.While party politics and politicization influence these negotiations, the strong institutionalization of intergovernmental policy coordination also shields Land governments to a significant extent from the immediate effects.Basic concerns over ISDS, moreover, were shared across party lines.The opportunity to participate in trade policy coordination throughout much of the entire trade policy cycle, combined with easing politicization dynamics after 2018, eventually paved the way out of the potential joint decision trap.

Belgium: A Joint Decision-Trap Despite Power Separation
Belgium made headlines in October 2016 when the minister-president of the Walloon Region, Paul Magnette, declared the region would not ratify CETA.In fact, two of the three regions, Wallonia, and the Brussels-Capital Region, along with the French-language community, opposed the agreement for the same reasons as the Austrian Länder (Egan & Guimarães, 2022; Parlement de la Région de Bruxelles-Capitale, 2016;Parlement Wallon, 2016).By contrast, the Flemish Region supported the agreement (Bursens & De Bièvre, 2023), as did the German-language community.
While some concerns had been addressed in the aftermath of the announcement through high-level negotiations, Belgium's signature is still pending.Thus, it is the only country in this study that has not yet ratified the agreement.
Belgium and Germany are similar insofar as constituent units are very powerful.They participate through joint decision-making, which affords them a veto in the ratification process.In the case of Belgium, however, this appears like a paradox, contrary to the theoretical expectations, as it is a federation based on the separation of powers.Why?
Belgian federalism emerged from a gradual process of constitutional reform that started in the 1970s.The fourth state reform eventually marked the transition from a decentralized unitary state to a formal federation in 1993 (Swenden & Jans, 2006).Unlike German federalism, which epitomizes the principle of powersharing, the federalization of Belgium was inspired almost entirely by the principle of separating powers between and among governmental tiers.Yet, in the case of foreign affairs, including trade policy, joint decisionmaking results from one particular constitutional provision: the so-called in foro interno, in foro externo principle.Already established in the third state reform of 1988, the principle was fully entrenched at the level of the regions in 1993.It stipulates that the communities and regions enjoy full competencies over the external aspects of a policy that they own domestically (Bursens & Massart-Piérard, 2009).Since encompassing trade agreements such as CETA inevitably affect regional jurisdictions, this rather unique constitutional provision creates-de facto and de jure-a requirement to coordinate and find consensus.In other words: The federal level is dependent on the approval of constituent units.
While in the German case, the highly differentiated and institutionalized system of intergovernmental relations, combined with a "buying time" approach pursued by the federal government, enabled governmental actors to avoid stalemate and ratification failure, both factors were absent in Belgium.Although Europeanization has intensified pressure on both governmental tiers to institutionalize intergovernmental policy coordination (Beyers & Bursens, 2006), these nascent structures are still weak (Bursens & De Bièvre, 2023;Egan & Guimarães, 2022).The institutional configuration of Belgian federalism, with its strong emphasis on separating powers, has-reinforced through the linguistic division-perpetuated the existence of rather isolated political arenas where party politics dominate.This represents a significant impediment to effective policy coordination across these separate domains within a pan-Belgian framework.Accordingly, the effect of partisan politics and the politicization of trade policy was felt much more directly than in the dense web of intergovernmental relations in Germany.

Canada: Effective Cooperation
Canada is the only case in this study that represents cooperative federalism, strictly speaking.The institutional configuration of federalism is similar to Belgium, as it strongly emphasizes the separation of powers.Formal institutional elements that would incentivize power sharing are weak and limited.However, Canadian federalism has no equivalent to the in foro interno, in foro externo principle.Accordingly, and in line with the theoretical expectation, joint decision-making is not an option.On the more informal level of intergovernmental relations, unilateralism, consultation, and cooperation variously characterize policy coordination between and among governments, depending on the policy sector (Schertzer, 2020;Simmons, 2017).
All three modes of intergovernmental coordination surface in the field of trade policy, depending on the trade agreement and approach of the federal government.Whenever a trade agreement potentially affects provincial jurisdictions, the federal government is incentivized to coordinate with the provinces as it cannot enforce compliance with provisions in this area (Fafard & Leblond, 2013;Kukucha, 2008;Skogstad, 2012).Since the late 1970s, the federal government and the provinces have established a rudimentary framework for consultation, the so-called C-Trade.However, it lies within the federal government's discretion to decide if it prefers to act unilaterally, consult, or cooperate more closely with the provinces.Moreover, most relevant exchanges rarely happen within C-Trade, but are ad hoc.For this reason, the Canadian provinces and territories have repeatedly demanded that the federal government enter into negotiations for a more formal intergovernmental framework for trade policy coordination (Council of the Federation, 2010, 2011).
Arguably, the cooperative approach underlying CETA was not all down to the EU's demand that the provinces be included.Considering the depth and scope of the agreement, it is difficult to imagine a unilateral approach, nor one that would exclusively rely on consultation.Indeed, CETA negotiations have been described as an exceptionally effective form of cooperation (Kukucha, 2015(Kukucha, , 2020;;Skogstad, 2012).This degree of close coordination has never been reached again, not even in the renegotiation of the North American Free Trade Agreement in 2018, which marked a serious threat to Canada (Paquin, 2020(Paquin, , 2022)).Unlike constituent units in Europe, participation of the Canadian provinces materialized in more than just resolutions.In addition to resolutions adopted through the Council of the Federation, the peak intergovernmental organization that serves as a forum for the provinces and territories to coordinate their positions horizontally, the provinces were also able to participate more directly.Provincial and territorial governments were part of the Canadian delegation during several negotiation rounds, received information from the chief federal negotiator in real-time, and were able to directly shape Canada's negotiation position in those areas that affected their jurisdiction (Hederer & Leblond, 2020;Kukucha, 2020;Paquin, 2020).
Important differences compared to their European counterparts facilitated this cooperative approach.The Canadian provinces neither operated in a highly politicized environment nor did they pursue any political goals like, for example, environmental or social standards.Moreover, the provinces had no concerns about a potential loss of institutional authority due to ISDS, regulatory cooperation, or a negative list approach.Regardless of what political party formed the provincial government, all provinces (and territories) perceived CETA primarily as an opportunity to strengthen economic ties with Europe.This was further reinforced when TTIP negotiations began to stall: The provinces considered CETA an opportunity to gain a first-mover advantage in the European market over the US.Only the province of Newfoundland and Labrador decided to withdraw from the CETA process until a conflict over fishery policy was settled in 2016 through the creation of a Fisheries Innovation Fund in support of the regional industry.

Conclusion
Although joint decision-making was originally identified as the dominant mode of intergovernmental coordination in German federalism, research has shown that it is not a German idiosyncrasy.Its broader applicability has proven particularly useful in multilevel policy coordination in the EU.Considering the comparatively high degree of institutional rigidity entailed in joint decisionmaking, existing research has been interested in determ-ining the conditions that promote or prevent stalemate and the joint decision-trap.The potential deadlock inherent in joint decision-making has animated much of this research.
The field of trade policy lends itself particularly well to testing the applicability of joint decision-making and its implications for intergovernmental policy coordination in more depth.First, trade policy today is a domain increasingly populated by governments representing different territorial scales.In particular, constituent units in federal states have emerged as actors with the authority to variously participate in trade policy-making.Second, at least in Europe, trade policy has become increasingly politicized between about 2013 and 2018.Both conditions are conducive to the joint decision trap.
Against this backdrop, this study examined how, and with what effects, constituent units in the four formal federations that have been part of CETA participated in trade policy coordination.Although it confirms that joint decision-making can play a role, it is not a ubiquitous mode.Depending on the historically established institutional configuration of federalism, constituent units, and the federal level can rely on different modes of intergovernmental coordination.Only in two out of the four federations-Germany and Belgium-did constituent units participate through joint decision-making.In Austria and Canada, unilateralism, consultation, and cooperation prevailed.It is also noteworthy that only in one case-Belgium-did constituent units prevent the ratification of CETA.
The case of CETA provides an opportunity to extend the analytical focus of comparative intergovernmental policy-making.While existing research tends to be centered on the EU, the inclusion of Canada contributes to a better understanding of the variety of modes through which governments cope with coordination problems across jurisdictions.As a next step, the ongoing implementation of CETA offers one potential avenue for analyzing in more detail how federal systems shape the role of constituent units in trade policy coordination after trade agreement negotiations have been concluded (see also Hederer & Leblond, 2020).Another avenue is the extension of units of observation that transcends formal federations by including regionalized unitary member states such as Italy and Spain.

Table 2 .
Table 2 presents, only Belgium has not yet ratified the Case studies overview.