Principals and transceivers: regional authorities in EU trade negotiations

The EU conducts its external trade policy under close scrutiny by member-states. Yet, what about the role of regional authorities in federal states? To answer, we look at the principal-agent chain of delegation and explore to which extent regions with stakes and competencies in the matter exert control over EU policies. We distinguish three ideal-type roles regions can perform: principals in their own right, members of collective principals, and introduce the new concept of ‘transceivers’. A region can perform the role of principal with formal competencies, as does Flanders, which as one of multiple principals can withhold a Belgian decision. A region can also be one of a collective group of principals, as is the case for the region Hesse. Regions can finally perform the role of a ‘transceiver’ of information from policy-makers to constituencies (and vice versa), akin to the role of endorser, without formal framework for involvement – as does Scotland. Since regions and states alike depend on private sector information, two conclusions prevail. First, despite formal competencies, regions that are members of collective principals struggle becoming the sole interlocutor for the private sector. Second, despite lacking competencies, transceiver-regions are crucial interlocutors complementing member-state control over EU trade policy. ARTICLE HISTORY Received 25 September 2018 Accepted 12 December 2019


Introduction 1
The European Union conducts its external economic policies towards third (non-EU) countries under close scrutiny of its member-states. One the one hand, the latter have delegated rulemaking powers to the EU-level, while on the other hand, they have opted to keep tight control over EU politics and policies (Delreux and Adriaensen 2017). This act of delegation from member states to the European Union has been studied in the Principal-Agent literature on the European Union (see for example Delreux and Adriaensen 2017). But what happens when we add regions to the picture? 2 This is important for two distinct reasons. For one, EU member-states have allowed the EU to expand EU competencies on external economic relations, particularly with regard to trade and investment policy. Second, some regional authorities within EU member-states have succeeded in extending their range of competencies to these areas as wellwhether in the framework of federalization, devolution, separatism, or political autonomyand they often have considerable economic and political stakes in EU decision-making in these areas, as is the case foramong others -Flanders, Scotland, and the German Länder. Therefore, an analysis of European multilevel politics in the field of EU policies in general, and the important policy field of EU external economic policies in particular, has to come to terms with the resulting demands of European regions.
In this paper, we apply the principal-agent (PA) framework to explore how regions deal with their trade and investment competences. Carefully mapping the principal-agent relationships is a necessary first step before the framework can be applied to the study of the politics of delegation and discretion (Delreux and Adriaensen 2017, 14). Furthermore, Delreux and Adriaensen point out 'that differentiating principals from endorsers is not always straightforward in the EU when agents are required to consult with public actors' (Delreux and Adriaensen 2017, 15). Therefore, we raise the following question: What types of roles do regions play in the principal-agent chain of delegation in EU trade negotiations?
Our objective is two-fold. First, we want to fill a conceptual gap in the literature. We do so by developing ideal types of regional authorities and by introducing 'transceiver' as a new concept. This allows us to adapt the P-A framework to include regional actors in the chain of delegation. This is important as we expect that their formal competencies determine their position in the chain and that this position affects their ability to compensate for the information-asymmetry and agency losses that arise when policymaking involves different levels of government, and in turn, different principal-agent links in a delegation chain. We argue that regions can be conceived of as either principalswhen they have competencies over a policy fieldor as 'transceivers'when they find themselves in a position analogous to one of the domestic interest groups who, as outsiders, have to rely on pressure politics. Within the group of regions as principals, we distinguish regions as one of the multiple principals with a capacity to block any decision at the higher level in the chain of delegation from those that are just one of a collective group of principals. Second, we check empirically whether regions perform the potential roles in our conceptualization (of principal or transceiver), arguing that their role influences their capacity to control the agent, i.e. the European Commission 3 and sets different incentives for the private sector to engage with them. We show how these ideal-typical roles play out in the cases of Flanders, Hesse, and Scotland. In the conclusion, we reflect on the roles of regions and the implications for the control they can exert over EU trade policies.
State of play: which role for regional authorities?
The principal-agent (PA) framework provides tools to analyse hierarchical relationships between several actors, and is particularly useful to systematically map the elements that affect control in situations of delegation. As such, the framework also holds the potential to fruitfully investigate how regional authorities try to represent their interests in EU decision-making on foreign trade and investment policy.
PA starts from the idea that when a principal delegates tasks to an agent who acts on behalf of the principal, or when multiple principals agree to delegate to one agent, problems of 'agency losses' may arise (Mitnick 1973;Ross 1973). Such losses arise when principals are unable to act against an agent that is not (fully) pursuing their interests, or that even undermine these (Hawkins et al. 2006;Kerremans 2006;Pollack 1997aPollack , 1997b. The principal, however, can adopt various procedures to limit the scope of agency activity, or to reduce the risk of losses 4 (McCubbins, Noll, and Weingast 1987;McCubbins and Schwartz 1984;Pollack 2003). One key reason for agency losses is an asymmetrical distribution of information. Indeed, due to the tasks delegated to it, the agent is likely to have an informational advantage vis-à-vis the principals. When member-states grant the European Commission a mandate to negotiate a trade agreement, the Commission as agent is directly informed about the positions of the third country partner with whom it negotiates whereas the member-states (and by extension, regions) as principals are not. This is so because the Commission negotiators conduct the international negotiations, while the member-state principals do not. This may make evaluation and control of the agent's work difficult for the member states. The agent's involvement in such negotiations, where the other party aims to convince her to make concessions, also increases the need for control, as it can require the agent to push the limits of what is going to be acceptable for its principals. Principals are thus motivated to know about the agent's (i.e. the Commission's) activities, in order to prevent or remedy possibly harmful concessions that the latter may make (De Bièvre and Dür 2005).
Whereas empirical research has paid considerable attention to the question of delegation and control by EU member-states (Da Conceicao 2010;Damro 2007;De Bièvre and Dür 2005;Delreux 2011;Delreux and Kerremans 2010;Dür and Elsig 2011;Elsig 2007;Franchino 2007;Kerremans 2006;Pollack 1997aPollack , 1997bPollack , 2003Tallberg 2002), to agency discretion (Gastinger and Adriaensen 2019), and to dynamics within (or in the context of) principals and agents (Conceição-Heldt 2017; Delreux and Laloux 2018), regions and their possibly peculiar positions in the principal-agent relationships have been neglected, particularly in trade negotiations conducted by the Commission on behalf of the EU. That is remarkable, as at least some regions in the EU have become vocal players on these issues. Indeed, regions within the EU member states may have and have thought to have good reasons to engage in monitoring and control too, specifically in EU trade policy. Regions may be adamant at following trade negotiations for the very same reasons as member-state governments do, and on top of that, at finding out what their own member-state governments are defending, opposing, or approving in the EU Council of Ministers. However, regions may suffer even more from informationasymmetries than their member-state governments do, and face greater difficulties to cope with these. They may have to rely on their member-state government's efforts to control the Commission, and on the willingness of that government to channel information to them. Moreover, they have to face the risk that such information may be biased against them, or incomplete. As such, if regions want to control the Commission, they will have to be able to control their member-state and/or find other ways to reduce the information-asymmetries. Regions vary, however, in their capacity to do so. For the understanding of this variation, it is useful to make use of ideal types of regions based on their formal competenciesor the lack thereofand their resulting location in the chain of delegation.
Analytical framework: the chain of delegation When applying PA analysis to EU trade policy, regional authorities in certain member states can be conceptualized as principals that delegate authority, e.g. through a negotiation mandate, to the member state as their direct agent, and through that member state to the European Commission as the ultimate agent. The resulting chain of delegation 'comprises several hierarchically organized PA relationships' (Figure 1). All actors at intermediate levels (i.e. the member states) in the chain fulfil the roles of principal and agent simultaneously. Societal interests and firms with a potential stake in this policy field have no formal decision-making authority in democratic political systems, meaning they do not act as principals, but can be best qualified as 'endorsers' as they can 'alert political actors to the consequences of various policies' (Milner 1997, 60;Delreux and Adriaensen 2017) and 'declare one's (public) approval or support of' (Oxford Dictionaries 2012). They are thus 'signalers' that provide interest-driven knowledge by which they can exert pressure on political actors and shape the latter's choices.
As mentioned in our literature overview above, regions are generally excluded from PA models to analyse EU external relations. We propose a framework to include them in a systematic fashion. In doing so, we locate them in the chain of delegation on the basis of the role that theyfollowing the formal competencies to which they are entitledcan play. Their position can thus vary according to region and to member-state. We distinguish two key types of roles: the one of principal or the one of transceiver (Figure 1 5 ). In case of a principal, there is a direct contractual link between the region and the ultimate agent in the delegation chain. In case of a transceiver, the region plays a role that is narrowed to both receiving and transmitting politically relevant information (among which support or opposition) in the delegation chain. It receives information from endorsers (interest groups, firms, NGOs). It transmits that information upward in the delegation chain through the link that it has with the immediate next player in the chain. That player may be interested in that information for substantive reasons or because it wants to be able to anticipate opposition or support in its own dealings with players even higher in the chain (that is, closer to the ultimate agent). Transceivers may also try to leapfrog the different links in the chain, by directly targeting the ultimate agent or its closest principals.
Whether a region can be principal or transceiver is necessarily conditioned by the formal competencies to which it is entitled in its own member state. A region can only be a principal in the chain when according to the powers to which it is entitled in its member state require that its formal approval of the trade agreement is necessary for the member state to which it belongs to be able to formally approve that agreement as well.
Even if such a region may be conceived as principal there is an important difference with a member state government as principal. Before the European Commission starts negotiating an FTA with a non-European country, it needs an official authorization from the Council, an institution in which regions are not directly represented. In some cases, howeverlike Belgiuma region participates directly in the Council of Ministers, but as representative of the EU member-state as a whole, not just as a representative of its own regional government's position.
Indeed, regions always co-exist with others in their country. This has an impact on the relations that they can establish with the EU-agent. Tierney's notions of 'collective' and 'multiple principals' are particularly helpful here (Tierney 2008) (Figure 2). In the case of the former, the ability of a region to exert pressure on the ultimate agent is conditioned by the willingness of the others within the same member stateand eventually also of the national governmentto act likewise. The pressures that can be exerted are then, the result of a compromise among the regions concerned, and may be diluted. In the case of the multiple principals, each region can exert direct pressure on the ultimate agent as each one's consent is required for the trade agreement to enter into force.
The concepts of multiple and collective principals help to differentiate among regions with respect to their relations with the EU-agent. Some may depend on others for the control that they can possibly exert on the agent. Others may have the tools to do so directly and independently from others.
A region does not necessarily need constitutional competencies, however, in order to have a place in the chain of delegation. It may also perform the role of 'transceiver', as it can both receive and transmit information. Regional authorities may receive information from endorsers, check this information's pertinence against the whole region's interests and legislation, decide on a position, and transmit this position and information to its member-state principal and/or agent. In this way, a region may have a more informal rolesomewhat akin to that of an endorseras it may have an interest and ability to exert control over policy fields in which they do not have formal competencies. A transceiver can try and use a broader array of channels compared to endorsers, by 'coloring outside the lines of its formal competencies' to broaden its capacity to exert control. Such a region can thus utter concerns during intra-state coordination on issues that do not strictly fall under those competencies, or can use the information in direct communication with the ultimate agent.
In short, when a region has no formal competencies over a topic, there is no act of delegation from that region to the agent and it can only act as a transceiver. When a region holds such competencies, however, it does so as part of a collective principal together with the other regions of its country. When an individual region's ratification of an agreement is required, it canat least formallyveto it, and thus perform the role of one of multiple principals.

Delegation and controlinformation
Principals can deal with the problems of agency loss by engaging in activities to control the agent. Control can take several forms. It can refer to the authorization and the instructions granted by the principals to the negotiating agent, i.e. the negotiation mandate for the European Commission. Principals can also use ad locum control mechanisms through which they can monitor the agent's negotiation behaviour by attending negotiations or through the establishment of a committee (Kerremans 2006;Delreux 2008). In addition, principals can influence the agent's behaviour through the application of (positive and negative) sanctions. Among the array of sanctions at the disposal of principals are budgetary control, control over appointments, and power to override agency behaviour through new legislation and to revise the administrative procedures laid down in the agent's mandate. Ultimately, principals can threaten to reject the agreement that the agent has negotiated, and, in case of agent non-compliance, effectively reject it. As the political cost of doing often is very high, principals tend to invest in approaches that reduce the information asymmetries when the agent is still in the process of negotiating with its third country partner(s). Regular agent reporting, double-checking information provided by the agent through alternative channels than the agent's, and eventually passive participation in the negotiations themselves are prominent among those.
Why does this matter? The formal role a region occupies may well determine a region's capacity to control. When regions are immediate principals, they may have access to formalized channels to get information on the agent's actions and to articulate their preferences. There can be formal requirements for instance, for a member-state to inform and debrief its regions on the state of the negotiations, which allows the latter to accurately assess agent behaviour. But regions may also enjoy the formal right to attend the formal sessions in which the Commission informs the member-states about the negotiations. Such right further reduces the information-asymmetry to which the regions may be exposed during the negotiations. They are likely to be able to exert more control than when they are transceivers, who have to obtain information and articulate their preferences by informal means. So, even if a transceiver can try to extend its monitoring, knowledge and capacity to control by trying to fish for information in coordination meetings on other policy fields where it does have competences, the level and quality of information will not be equal as in a case where the region is a principal and has a formally recognized role in and access to formal coordination mechanisms with respect to the specific policy field of trade.
How do regions perform the roles of principals and transceivers in EU trade negotiations?
In the next sections, we check how regions perform the role of principalwhether as one of the multiple principals or as part of a collective principalor transceiver. We thus focus on the way regions act in a context where different levels are involved in a chain of delegation, and engage in activities that reduce information-asymmetry with regard to EU external trade policy.
As indicated above, whether a region is principal or transceiver can be identified by looking at the formal powers of the regions on the policy domains, as determined in the constitution of their member state or in equivalent legal frameworks. If a region has exclusive competences over a policy field and a veto right on its own, it is one of multiple principals. If competences require preliminary coordination and a joint decision by regions, the regional authorities are part of a collective principal. They cannot simply veto a decision on their own.
Regarding the choice of regions, we opted for most-similar regions, namely regions with significant legislative powers (constitutionally 'strong' regions) that were relatively similar to each other on a number of economic indicators, so as to be able to focus on the theoretically relevant variation in types of roles (Deforche 2012;Eckhardt 2011;George and Bennett 2005;Gerring 2007;Rowley 2002;Yin 1994). Rowe argues that '[t]here is a clear preference for constitutional regions to form ad hoc advocacy coalitions, which consciously exclude weaker, non-constitutional regions' (Rowe 2011, 8-9). As constitutional (or regulatory) regions pursue different goals than non-constitutional regions do, they do not make for good comparisons. The website of the Regions with Legislative Powers informs us that this narrows the possibilities down to regions from Austria, Belgium, Germany, Spain, the United Kingdom (Wales, Scotland and Northern Ireland), Italy, Finland (Åland) and Portugal (Azores, Madeira). We note large differences in the degree of autonomy these regions enjoy, especially with regards to competencies over EU trade agreements. The Belgian regions had the strongest competencies over this domain, and among those Flanders is the region with the highest trade dependency numbers.
The 'Regional explorer' of the OECD 6 informed us about the relative economic strength of regions, so as to keep exogenous variation stemming from there roughly constant. We used eight indicators, namely gross domestic product (GDP), GDP per capita, total population, area, gross value added (GVA) for sectors A-B (which comprises agriculture and fishery), sectors C-E (i.e. mining and manufacturing), sectors G-I (i.e. wholesale, retail trade, repair of vehicles and personal household goods, transport storage and communication, …), and GVA for sectors J-K (i.e. financial intermediation, real estate, …). Using the data on the value added for those different sectors allowed us to find regions with a similar composition according to sectors' importance. Eliminating regions with a country's capital, this search for regions with great economic similarity, yet relevant differences in their formal competences, lead us to the regions of Flanders, Catalonia, Hesse and Scotland. Although we did conduct interviews in Catalonia, we decided to leave out this case, as we have serious concerns about the reliability of the interview data.
The empirical data thus consists of interviews, document analysis, literature study and two non-participant observations for the case of Flanders: at the Belgian intra-federal coordination meeting (DGEon 28 March 2013) and at the EU's Trade Policy Committee (on 1 March 2013). The information for the case of Flanders was collected between March 2012 and December 2013, for Hesse between April and June 2014, and for Scotland, one interview was conducted in June 2014, whereas the bulk of interviews were done after the Scottish referendum on independence on September 28, 2014, namely February-March 2015. The interviews thus constitute the most important basis for the analysis below, yet we limit the number of quotations and references, as some respondents preferred not to be quoted or to verify quotes before publication.
In the next section, we look (1) whether these regions have competencies over EU trade policy and whether we find an act of delegation in the federal framework (meaning that they are a principal), and (2) whether they are active on EU trade policy. If a region is active without having competencies, we categorize, describe and elaborate on its behaviour in its role of transceiver.

The region of Flanders as one of multiple Belgian principals
Among the most trade dependent in the EU and the world, the Flanders region of Belgium holds full competence for sales and export policy. Whereas the federal level is competent for 'multilateral trade policy', federal legislation is not hierarchically superior to regional legislation. Moreover, due to the principle of 'in foro interno, in foro externo', Flanders holds competence over trade-related aspects of agriculture, environment, transport, employment, culture and education, science and innovation, and health services (Bursens and Deforche 2010;Deforche 2012). Nevertheless, the representation in the EU Council of Ministers' Foreign Affairs configuration (under which the Trade Configuration) is done by a federal Minister, according to the Cooperation Agreement (2003 amendments) (Vlaamse Vertegenwoording bij de EU -Permanente Vertegenwoordiging van België 2010). The extent of control by Belgian regions over EU trade negotiations was demonstrated by Wallonia's temporary refusal to ratify the CETA-agreement with Canada in 2016a veto that caused much dismay among Flemish policy-makers who held very favourable policy preferences about the prospective preferential trade agreement with Canada. Following internal Belgian negotiations, and others with Canadian and EU representatives, the Walloon government agreed a memorandum which clarified the arbitration procedure. The Walloon Parliament subsequently gave consent for the Belgian Foreign Minister to give approval for the European Commission to sign the agreement (Walker 2017, 15; Van der Loo and Pelkmans 2017).
Within the unit 'International Entrepreneurship of the Policy-division' in the Flemish Department of Foreign Affairs (DiV), one official works (almost) full-time on EU trade policy, while a second acts as back up. The unit monitors the EU Common Commercial Policy and serves as the Flemish voice in the intra-federal coordination meetings. Flanders' stance is determined through consultation with interested services or experts within the Flemish government through its 'Working Group EU Trade', containing representatives of all policy domains, both at the bureaucratic and the political level (Interviews with DiV on August 27, 2013 andApril 15, 2014, Brussels). Finally, the region has a Flemish representation to the EU, located within the Belgian Permanent Representation, that occasionally has direct contact with the Trade Commissioner and Commission officials, yet attends neither formal nor informal meetings with regard to EU trade policy. Although there is a Flemish liaison office to the EU (VLEVA), it is barely involved in trade matters (Interviews with Flemish representation to the EU on April 15, 2014, Brussels and with Flemish Chamber of Commerce and Flemish Liaison office to the EU (VLEVA) on November 7, 2013, Brussels). Flanders is thus a region with full formal competencies that makes it one of multiple principals with strong pro-trade preferences and operating, yet small operational capacity to perform that role.
At the Belgian federal level, the two most important actors are the unit for trade policy EU and WTO in the Foreign Affairs' Directorate-General European Affairs and Coordination (DGE) of the Federal Public Service Foreign Affairs, as well as the Federal Public Service Economy. The first is responsible for multilateral trade policy, organizes the coordination between all relevant authorities, and meets every Thursday before the EU Trade Policy Committee (TPC) to discuss the items on the TPC agenda. Always invited are the FPS Economy, the offices of the ministers of Foreign Affairs (federal and regional), the DiV and its counterparts in the other regions, and depending on the agenda, other relevant regional and federal administrations and cabinets. Four people work on trade policy full-time, plus the trade ambassador who works most of her time on trade policy matters. DiV actively participates in the Belgian coordination meetings. In case of conflicting Flemish and Walloon preferences, the FPS Foreign Affairs mediates, relying on the notion of 'complementary interests', so as not to forego the possibility to co-decide in the Council of Ministers (Interview with FPS Foreign Affairs, Foreign Trade and Development Cooperation, Directorate General European Affairs, September 9, 2013, Brussels).
The Flemish representation does engage in targeted actions towards EU officials of DG Trade and even the Trade Commissioner, but only on a specific request of endorsers and in a joint effort with the FPS Foreign Affairs (Interview with Flemish representation to the EU on April 15, 2014).
Illustrative of the weight Belgian regions can bring to bear, is the following episode of how they forced the FPS Economy and Foreign Affairs to transmit regional preferences to the European Commission directly. The regions wanted strong guarantees that audiovisual services be excluded from the transatlantic negotiating mandate for the European Commission. The Minister for Economy Vande Lanotte and his FPS Economy administration did not think it was necessary to do this, since refusing to negotiate substantially about them could still always be done later (Interview with Office of Minister Vande Lanotte, Minister for Economy, SME, Middle Class and Energy September 27, 2013; Interview with FPS Economy, SME, Middle Class and Energy, General Direction Economic Potential on August 27, 2013, Brussels). The Federal government of Belgium thus represented the position decided by the Belgian regions, which was included in the final mandate for the negotiations, albeit that also other EU member-states shared this position (Inside U.S. Trade 2013).
In matters of exclusive regional competencies like trade, agricultural and cultural matters, and even in matters of mixed competence, Flanders is one of multiple principals, as it can veto a federal decision. Paradoxically, however, this results in no Belgian position through which it cedes decision-making power to fellow EU member-states in the Council of Ministers. However, in its role of one of multiple principals, Flanders receives information from TPC meetings directly (Interviews with DiV on August 27, 2013 and April 15, 2014, Brussels and with FPS Foreign Affairs, Foreign Trade and Development Cooperation, September 09, 2013, Brussels). It can follow negotiating texts even on matters not directly related to its competences, is able to put topics on the Belgian agenda and propose actions for the (federal level) agent. When no other actor in Belgium vetoes it, the federal representatives will present this in the Council.

The region of Hesse as part of the collective principal of the German Bundesrat
In the Federal Republic of Germany, each Land participates in the legislative process through the Bundesrat, the second chamber (Basic Law for the Federal Republic of Germany, Art. 50, 2014). Although trade matters are of exclusive competence of the federal government and regional and local bodies do not individually and directly participate in the ratification process (Rowe n.d., 15;Schuh et al. 2018, 59), so-called mixed trade agreements do have to be approved by the Bundesrat when they are touching on Länder competences. With regard to important recent EU trade agreements such as TTIP and CETA, the Bundesrat has held Investor-State Dispute Settlement to be of concern to them and it has also held that its constitutionally enshrined sovereignty over culture and media policy means that it has to set the terms of the German national position on those treaties (Rowe n.d., 16;Eschbach 2016, 11; see also Walker 2017, 19). In case consent is required, the Bundesrat has a collective veto. In case the Bundesrat can only lodge an objectionand does this, a majority in the Bundestag can reject it (Eschbach 2016, 12; see also Auel 2008, 424-439;Linn and Frank Sobolewski n.d.;Jefferey 1997, 254-255). The Länder themselves thus do not formally and directly participate in the ratification process (Schuh et al. 2018, 59). The region or Land Hessen thus is but one of constituent unit of a collective principal in the chain of delegation in EU trade policy.
In order to leverage this extent of control, the Land has a unit Foreign Trade and Location Policy, situated in the Department for External Economy, Middle Class, Vocational Education and Technology, of Hesse's Ministry of Economic Affairs, Energy, Transport and Regional Development. One to two people work on trade policy, one as trade expert and one on communication with the Bundesrat. They are not working on trade full-time, the expert spends about ten to twenty percent of the time on trade issues (Interview with Hessian Ministry of Economic Affairs, Energy, Transport and Regional Development, May 23, 2014 by telephone, and with the Representation of the State of Hesse to the EU, April 7, 2014, Brussels). The Representation of the State of Hesse to the EU functions an antenna for information. The representatives do not take part in the decision-making, not even informally. The Representation of the State of Hesse in Berlin performs a similar function, but collects information from and coordinates with the Bundesrat and the federal authorities (Interview with the Representation of the State of Hesse to the EU, April 7, 2014, Brussels).
At the federal level, the central actor monitoring EU trade policy is the department for external economic policy (Aussenwirtschaftspolitik) in the Directorate-General V for external economic policy of the Federal Ministry for Economic Affairs and Energy (Bundesministerium für Wirtschaft und Energie; BMWI). Also the Directorate-General European Policy (DG E) and/or the DG Political Coordination (DG L) may play a role. Within the department for external economic policy, two main units work on trade policy: one for general trade policy (EU/WTO), and one for customs, the system of general preferences, policy instruments, and rules of origin. Both units employ eight civil servants. In addition, and especially of use for the monitoring of FTA negotiations, there are desks with a geographic focus, each with someone working part time on trade issues (8 parttimers in total). They coordinate with the other Ministries, Representations and other Länder by way of video conferences held every Thursday before the EU Trade Policy Committee meeting in Brussels. Since the Länder hold certain competences in the area of services (e.g. regulation of professions), they coordinate on market opening for services (Telephone Interview with Federal Ministry for Economic Affairs and Energy, June 17, 2014).
The German Foreign office or Auswärtige Amt engages in trade policy formation as well as trade promotion activities in cooperation with Germany Trade and Invest. It takes part in negotiations on trade and investment promotion and protection, double-taxation conventions and represents the Federal Government in international bodies such as the WTO, the OECD, the UN and the EU. The Foreign Office and its Permanent Representation in Brussels closely follow the discussions in the EU TPC, and the Foreign Office regularly sends someone to the video conferences organized by the Ministry for Economic Affairs. The Permanent Representation of the Federal Republic of Germany to the EU is the link to the EU Commission and other actors. It does not, like in many other countries, represent Germany in the Trade Policy Committee meetings. The Permanent Representation does take part in video conferences organized by the BMWI, communicates to other actors in the EU, engages in social networking, and thus provides input to the BMWI, which formally takes the decisions (Telephone Interview with Federal Ministry for Economic Affairs and Energy, June 17, 2014).
In contrast to Flanders, the state of Hesse, or any German region, thus never is one of multiple principals. The region itself cannot veto a decision agreed upon by other actors, but is rather part of a collective principal. The Federal Ministry for Economic Affairs coordinates regularly with the Länder through the Bundesrat, informing them on its position and trying to convince them to go along, yet not necessarily interacting with them individually.
When such a region wishes to influence EU trade negotiations, it has to try and form a coalition with other German regions. In order to find such partners, they may look at regions with the same party(-ies) in the coalition and/or at regions with similar economic interests and similar interests in trade policy. They may also try and approach the Ministry for Economic Affairs informally and communicate interests, in which case the region acts somewhat like a transceiver. Remarkably, 'Chances at success may be higher than by working through the Bundesrat, as this way, a region can ask for additional information and avoids having to coordinate with the other fifteen regions' (Telephone interview Vereinigung der Hessischen Unternehmerverbände/Hessen Metall).

The region of Scotland as a transceiver
The UK is an asymmetrically decentralized unitary state with three devolved governments: Scotland, Northern Ireland and Wales. Although the UK parliament has absolute sovereignty, Scotland obtained its own legislative 'devolved' powers (Committee of the Region, n.d.). The UK Parliament legislates in reserved areas in Scotland, listed in the Scotland Act, and may also legislate on devolved mattersonly with a consent of the Scottish Parliament (listed in the Sewel Convention). As the UK government is responsible for defense, foreign affairs, economy, social security and the constitution, it holds reserved powers in international relations, including foreign trade except for observing and implementing EU and European Convention on Human Rights matters. Also important reserved powers related to trade policy are competition, monopolies and mergers; law on companies and business associations, insurance, corporate insolvency and intellectual property; regulation of financial institutions and financial services; the regulation of certain professions; employment legislation; most consumer protection; data protection; transport safety and regulation; etc. (Committee of the Region, n.d.). The Scottish government on the other hand, has jurisdiction over health, education, justice and policing, and the local government (Committee of the Region, n.d.). In addition, it has devolved powers in economic development; agriculture, fisheries and forestry; environment, etc. (Committee of the Region, n.d.). 7 However, EU trade agreements, such as CETA, include provisions for the trade of agricultural goods, food and drink, as well as public procurement. 'These areas impact upon devolved policy, within the competence of the Scottish Government and Scottish Parliament'. This has given rise to discussion on the need for common frameworks to be developed in the UK on those areas (Walker 2017, 7).
The most important Scottish government actor is the DG Enterprise, Environment & Innovation. The Business Directorate's EU and Trade Policy team (within the European Structural funds division) is responsible for strengthening economic growth, increase the overall productivity, improve partnerships through engagement with the business, consumer advocacy and trade union communities, local government and the broader public; develop high quality business support particularly through Scottish Enterprise (SE) and Highlands & Islands Enterprise (HIE)the economic, innovation and investment agencies. There are five people working in the EU and Trade Policy Team, which coordinates cross-government groups on topics as procurement, health services, lawyers, IP, meeting three times a year (Interview with Scottish Government Business Directorate, January 15, 2015, Brussels).
Scotland Europa promotes the Scottish interests with the European institutions and the representatives of the regions and member-states. 8 It is a member organization with a diverse partnership of public authorities, businesses, local governments, trade unions and education and voluntary organizations. The Scottish Government as a whole is represented to the EU through the Scottish Government European Union Office. It works on EU policy that has an impact in Scotland, identifies opportunities for engagement and promotes the Scottish interest (Scottish Government 2014).
At the national level, the most important UK government actor is the Department for Business Innovation and Skills (BIS). The Secretary of State for BIS has the lead on trade policy. The Minister of State for Trade and Investment sits in both BIS and the Foreign and Commonwealth Office. The Trade Policy Unit, mainly situated in BIS, employs people from the Department for International Development as well as from the Foreign and Commonwealth Office. It contains three staff members from the Foreign Office, joining the people from different ministries with a focus on trade. BIS fulfils a coordinating role based on input from business and NGOs, on their own economic analysis and information from (other) ministries. BIS thus draws up the government's position on policy dossiers, after which they (try to) get this position cleared by the British government. Second, it is tasked with reflecting on trade policy in the long term. And third, BIS focuses on crosscutting issues relevant to the British government such as the possible transfer of competences to the EU-level, keeping an especially close eye on mode 4 of services (migration) (Interview with UK Department For Business, Innovation and Skills, March 20, 2015, London).
The Trade Policy Unit employs thirty-five up to forty officials, seven of which work for the Bilateral and Regional Trade Negotiations Team with a geographical focus, while other teams specialize in issues (services, …). While the Trade Policy Unit monitors the Council's Trade Policy Committee, the Trade Policy Steering Group is the main coordinating body that joins directors of most key policy departments, the UK Permanent Representation in Brussels and in Geneva. The Trade Policy Unit organizes fortnightly meetings with colleagues from across White Hall on all negotiations, with a separate fortnightly meeting for TTIP (Interview with UK Department for Business, Innovation and Skills, March 20, 2015, London).
The UK Permanent Representation to the EU represents the UK in the EU negotiations. It has an EU trade policy team of 3 people, in cooperation with the Trade Policy Unit. At the moment of interviewing, two of the representatives were from BIS and one from the Foreign Office (Interview with UK Department for Business, Innovation and Skills, March 20, 2015, London).
Within this national level administrative division of labour, the so-called Scotland Office is the key body through which Scotland can represent its interests within the UK government in reserved matters such as foreign trade. The policy division advises the UK government on the impact of UK policy in Scotland (Scotland Office n.d.).
Crucially, Scotland thus does not have any formal powers over trade policy, which is a reserved matter to the Westminster government. Therefore, the devolved authorities cannot veto any type of decision, not even by joint cooperation. This, however, does not mean that the Scottish government is not active on EU trade policy. On the contrary, it monitored trade negotiations, especially in the run up to the 2014 referendum on Scottish independence. The idea was that if Scotland were to be independent, it would also have to take care of this policy area. In addition, TTIP caught the eye of Scottish actors, as the US is the largest non-EU export market for Scotland, is a country with historical ties to Great Britain, and negotiations drew a large amount of media and general public attention. Lastly, the devolved competence of health services, which in Scotland is organized according to a public model (in contrast to the English private model) have dominated this public attention for the TTIP negotiations, while perhaps followed by agriculture as the second item attracting public attention (Telephone interview with Scotland Europa February 5, 2015). The Scottish government has for a large part been trying to clarify how trade agreements apply to member-state's devolved administrations. De facto however, meetings of the Joint Ministerial Committee and the TTIP working group are entirely open to the participation of officials from the devolved administrations (Interview with UK Department For Business, Innovation andSkills, March 20, 2015, London andwith Scottish Government Business Directorate, January 15, 2015, Brussels).
The devolved administrations and Scotland in particular thus engage in discussions on matters of external trade policy that are strictly not of their competence. They get information from Westminster on UK ministers' positions, on what is of concern in other member-states and how the UK Foreign Office contributes. At the same time, Scottish public officials receive relevant information from the private sector and transmit this to UK actorsalong with the private sector transmitting itself that information directly to the UK government, who see it as an additional channel if the Scottish government pushes for their case as well. Hence, even absent formal competencies, Scotland acts as a transceiver, transmitting information on private sector preferences within the chain of delegation of EU trade policy formation.
Of course, as a transceiver, Scotland can merely suggest topics for the agenda. They can propose actions but have no way of knowing whether the agent will act upon them and cannot refrain the agent from taking actions. However, as the UK has an asymmetrical structure, there are few regions with devolved powers and Scotland shows the most interest in trade negotiations. Due to Scotland's focus on a number of key areas, it can have input on those areas. Moreover, the representation office has direct meetings with the EU Commission, which means that it can obtain information directly from Brussels. Scottish actors do and cannot directly steer the European Commission's conduct of trade negotiations, but a 'slip of the tongue' is always possible during such conversations (Interview with Scottish Government EU Office, January 21, 2015, Brussels).

Conclusion
In this article, we have explored the different roles regional actors can play in the monitoring of the European Commission as agent in European external trade negotiations. We have identified regions to differ in their roles as transceiver, collective principal or one of multiple principals. Flanders acts as one of multiple principals, Hesse can act as part of a collective principal, and Scotland acts aswhat we have calleda transceiver.
Two main conclusions stand out. First, despite their formal far reaching competencies, regions that are members of collective principals have a hard time becoming the sole interlocutor for the private sector. And second, despite not having any formal competencies, transceiver regions are still crucial interlocutors complementing member state control over EU external trade policy.
The different ideal typical roles we have thus distinguished are associated with different amounts of control (or influence) a region can exert over the EU negotiating agent. Of all types of regions, members of collective principals have the best access to information, yet somewhat surprisingly also transceivers do hold non-negligible considerable sway over the agent through their informal channels. We have indeed found that although information-asymmetry is considerably larger for Scotland than for Flanders and Hesse, it has gained access to meetings that remain inaccessible for other actors like endorsers, i.e. private economic and societal interests. Due to its legal competencies in policy fields other than external trade policy strictu sensu, Scotland can make use of intra-state coordination mechanisms, for instance to obtain information about the positions of other member-states in the EU TPC, something EU lobbyists find the hardest to discover. This is why we argue that the introduction of the transceiver concept helps the classification of types of roles actors can play in a chain of delegation.
The contribution we thus offer is conceptual and theoretical in that we complement principal-agent literature with a typology of the roles of regional actors in a federal kind of setting. Our second contribution is empirical in that we find that regional authorities (such as Scotland), can act as if they were endorsers (private economic or societal interests) but, at the same time, can forge access to channels of information they are not formally part of in policy fields for which they hold internal competences.
Although we have developed the transceiver concept from the EU-context, we see potential for its application to other regional political actors elsewhere, like for instance the role of Canadian provinces or US states in the US federal trade policy.
Of course, our research has distinct limits and also raises questions that we have not addressed. One of the limits of preference driven frameworks of analysis like principalagent is that they presuppose the empirical observability of actors' preferences being expressed. Very often, however, regional authoritiesjust like EU member-statesdo not receive any input from endorsers, resulting in the absence of any clear policy position taking form (see Adriaensen 2016). Formal powers thus only become useful when the actor has a specific interest at stake and a position to push forward in the negotiations a prerequisite for this is that the actor is knowledgeable about economic interests. Furthermore, the actual use of those formal powers is warranted only when the region or member-state in question has an interest to defend that is opposite to what the agent was planning to defend. Also, we have not systematically sought to map whether or not region-specific endorsers and regional transceivers attempt to leapfrog actors in the chain of delegation and directly address the negotiating agent itself, the European Commission. Both for such endorsers and transceivers this may well be a costly and risky strategy, and the agent may not necessarily want to engage in such interaction, let alone act upon it.
More specifically related to our empirical research, the timing of interviews could have had an effect on the results, as the hype about TTIP had sparked attention from regional authorities and activated them. Follow-up research would be needed to check whether regions remain active, or whether some go back to sleep, in line with the general ebb and flow of politicization of EU trade agreement negotiations.

Notes
1. This paper builds on previous work by the authors in non-peer reviewed outlets. We thus reuse (parts of) texts published in Kersschot (2016aKersschot ( , 2016b; as well as Kerremans (2013, 2014). 2. We use the term regions to refer to 'micro-regions' ('sub-national' or 'sub-state' entities-below the nation-state). These are territorial areas that are smaller than the state to which they belong (e.g. Flanders, North Rhine-Westphalia, Catalonia, …). This term is opposed to 'macro-regions' (supra-national and thus above the nation-state), which are large territorial units comprising different states (for example, the EU, Mercosur, …) (based on Rowe 2011, 215; Van den Brande 2008, 10). 3. Depending on the type of negotiations and the article of the Treaty on the Functioning of the European Union that applies (art. 207 or art. 218), the agent/negotiator could also be another party than the Commission (e.g. the Council Presidency, EU High Representative for Foreign Affairs), and this could be potentially different according to the topic at hand. In order not to overly complicate the explanation however, we use the example of the Commission negotiator as agent. 4. When an agent intentionally pursues his own interests, and not the ones of the principal, this is called shirking. When the agent unintentionally deviates from the principal's preferences, due to structural incentives, this is termed slippage (Pollack, 1997a). 5. Voters can be conceived of as 'ultimate principals'. We are well aware that by leaving out voters and parliaments that we drastically simplify the policy process. For analytical purposes however, it is useful to rely on the political-economy assumption that collective organization and lobbying of economic sectors can be taken to primarily influence the government's positions in trade policy making (De Bièvre and Dür 2005; Poletti 2011). 6. http://stats.oecd.org/OECDregionalstatistics/#story=0. 7. With the Brexit-vote the division of powers in the UK, also specifically on trade policy, has resurfaced. 'Many of the policy areas concerned have been legislated for at European level. As the UK leaves the EU, these powers will return to the UK, where under proposals included in the European Union (Withdrawal) Bill' (Walker 2017, 6-7). 8. https://www.scotlandeuropa.com/.