Rebordering Europe in the Ukraine War: community building without capacity building

Abstract How has the European Union (EU) developed in reaction to the Russo-Ukrainian war, its most severe military threat since the end of the Cold War? In a ‘bordering’ analysis of political development, this article studies changes in the closure and control of the EU’s boundaries with Russia and Ukraine as well as its internal boundaries between 2013 and August 2022. It finds that the EU’s response has hitherto consisted in a regulatory process of community building without concomitant capacity building. Whereas the EU has increasingly closed its boundaries with Russia, it has progressively opened its boundaries towards Ukraine. By contrast, the war so far has not had a discernible impact on the EU’s internal boundary configuration and its authority and capacity for boundary control. These preliminary findings are in line with the EU’s developmental path as a ‘regulatory state’. More than half a year into the invasion, they stand in contrast to ‘bellicist’ expectations of centralised capacity building in response to military threats.

culmination of Russian revisionism, which threatens the sovereignty of former Soviet Union republics and allies, opposes their membership in the EU and NATO, and rejects the principles of the contemporary European order. The Ukraine war thus constitutes the most significant military threat to European integration in the post-Cold War era. How does it affect the EU? Will it alter the course of European integration significantly?
Bellicist theories of political development (Hintze 1994;Riker 1964Riker , 1987Tilly 1975Tilly , 1990Weber 1978) suggest that it might. They have long claimed that the exigencies of warfare have been the decisive factor in the emergence and prevalence of the modern state. Daniel Kelemen and Kathleen McNamara recently attributed the 'incomplete, uneven, and dysfunctional development' of key EU policies (Kelemen and McNamara 2022: 965) to the fact that the EU polity was formed in a gradual process of market-making rather than in the 'crucible of war' . Correspondingly, the Russian invasion of Ukraine holds the potential to remake the EU (McNamara and ). Yet the bellicist expectation of EU-level centralisation and administrative, financial, and capacity-building in response to the Ukraine crisis may be based on a misconstrued historical analogy and a flawed understanding of the drivers of EU political development (Eilstrup-Sangiovanni 2022; Freudlsperger and Schimmelfennig 2022;Genschel 2022).
In this article, we approach the effects of the Russo-Ukrainian war on EU political development from a 'bordering' perspective. Following the seminal work of Rokkan (1974) and Bartolini (2005) on the linkage between boundary formation and political development in modern state building and European integration, we theorise boundary change in the EU. We conceive of boundaries as functionally differentiated institutions that regulate the movement of persons and products across territorial systems. In a bordering perspective, changes in the closure and control of the internal and external boundaries of a polity are important markers of political development. According to Rokkan and Bartolini, the ability of political units to both close and control their boundaries to the outside world is a necessary precondition for internal consolidation and development.
Specifically, we classify the Russian attacks as a boundary shock that challenges the identity, autonomy, and security of the EU's political community. Such attacks typically trigger processes of 'external rebordering' (Schimmelfennig 2021. For one, external rebordering consists in the increasing closure of the polity's external boundaries towards the attacker. In addition, the community insiders close ranks. However, external rebordering is significantly shaped by the regulatory nature of the EU polity, in which the executive control of boundaries and resources needed for this purpose remain largely with the member states. Moreover, most member states organise their collective defence and the pooling of the relevant military capacity in NATO rather than the EU. We therefore expect the EU response to the Ukraine crisis to consist predominantly in the regulatory closure (vis-à-vis Russia) and opening (vis-à-vis Ukraine) of the EU's external boundaries.
Empirically, we use a novel dataset of EU boundary configurations to describe the development of the EU boundaries with Russia and Ukraine between 2013 and 2022 in comparison with the EU's internal boundaries. We show that the EU's boundaries with Russia and Ukraine stood at a similar level of closure before the war. Since 2014, they have diverged strongly in a dual movement of increasing closure towards Russia and gradual opening towards Ukraine. The exclusion of Russia and the inclusion of Ukraine were accompanied by decreasing and increasing cross-border transactions, respectively.
At the same time, we do not observe major changes in either the internal boundary configuration of the EU or the control of its external boundaries. More than eight years after the annexation of Crimea and more than half a year into the Russian invasion, the Ukraine crisis has not triggered a centralisation of EU decision-making, nor has it caused significant administrative, financial, or coercive capacity-building at the EU level. The EU has responded to the Russian challenge in line with its standard institutional template of a 'regulatory state' (Majone 1996). As the war is ongoing and its effects on the EU are still unfolding, it is too early to draw firm conclusions. So far, however, we find that the Russo-Ukrainian war is remaking Europe, indeed, but differently than both bellicist and Rokkanian theories of state-building suggest. Neither has the war ushered in a new era of European state-building, nor has the closure of external boundaries triggered a movement towards internal institutional consolidation. Whereas the Russian invasion has had a major impact on EU community building, its effects on EU capacity building have been minor. 'Bordering' encompasses all activities of boundary making and management. We define boundaries as both territorial and functional institutions. Boundaries consist of rules that regulate the movement (entry and exit) of subjects (persons) and objects between territorial organisations. As functional institutions, they differ by the type of transactions they regulate. Conventionally, the literature distinguishes economic, cultural, political, and military boundaries in line with the respective functional subsystems of territorial social systems (Bartolini 2005: 13-20;Rokkan 1974: 42).

Bordering, polity formation and the Ukraine crisis
The institutional configuration of each boundary consists in a combination of closure and control. Closure determines to what extent the rules facilitate or restrict exits and entries. Open borders allow for unrestricted movement; closed borders prohibit exit or entry. Control refers to the legal competence and the resource-dependent capacity to make, implement, and enforce these rules. Finally, boundary congruence refers to the overlap of functional boundaries. Congruence is high if different functional boundaries delimit the same territories and if they are equally closed and controlled. In sum, the 'boundary configuration consists in the constellation of closure, control, and congruence across the economic, cultural, political, and military boundaries' of a territorial system (Schimmelfennig 2021: 315-316;cf. Bartolini 2005: 15-16).
Stein Rokkan and Stefano Bartolini pioneered the analysis of political development as boundary formation. Rokkan started from the assumption that the control of boundaries affects 'the configurations of political resources inside each territory' (Rokkan 1974: 43). In his account, the relative difficulty of boundary control shaped the diverse trajectories, timing, and forms of state building in Europe. Following Hirschman's (1970Hirschman's ( , 1978 argument that dissatisfied individuals turn to 'voice' when 'exit' opportunities decrease, Rokkan further theorised a systematic effect of the increasing control of external territorial boundaries on the internal development of representative institutions and democratisation (Rokkan 1974: 49-53).
Bartolini (2005) elaborated Rokkan's seminal work and extended it to European integration as a new phase in Europe's political development. In his analysis, controlling the external boundaries of a territory and locking in actors and resources strengthens in-group identities and behavioural conformity. It reduces the options of societal actors to withdraw from territorial public policies, increases pressures for territorial social cooperation, and secures the resources required for effective collective action. It facilitates internal learning and trust building, and it generates demand for the legitimation of and participation in political authority (Bartolini 2005: 40-7). In sum, effective boundary control and the concomitant reduction of exit opportunities set in motion a mutually reinforcing process of political structuring (strengthening voice opportunities) and political production (public goods provision).
Conversely, open and incongruent boundaries and the exit opportunities they generate lead to political de-structuring and a decline in the scope and effectiveness of political production (Bartolini 2005: 53). In Bartolini's analysis, this has been the effect of European integration. By lowering, removing, and differentiating the internal boundaries between the member states, and by failing to compensate internal 'debordering' through external boundary control and supranational capacity building, the EU has weakened national democracies and welfare states (Bartolini 2005: 242-5, 369-81).
Rokkan and Bartolini focus on the effects of boundary formation on the internal political structuring of territorial political systems and their ability to produce public goods. They theorise and analyse how a given boundary configuration shapes political development. By contrast, they have little to say on the emergence and change of the boundary configuration itself. In addition, their analyses focus on the nation-state rather than the development of the EU. Rokkan explained the historical political development of European nation-states, and Bartolini was mainly interested in the effects of European integration on national political development.
We propose a bordering analysis of European integration that complements Rokkan and Bartolini. Our analysis focuses on boundary change rather than boundary effects, and it does so for the EU as an emerging polity rather than for the nation state. We make an explicit distinction between the two major dimensions of political development: community building and capacity building. In a bordering perspective, boundary closure is a measure of political development as community building. Institutional barriers for the exit and entry of persons and objects not only signify the closeness and distance (or inclusion and exclusion) between the political communities on both sides of the border; they also facilitate or impede the cross-border communications and transactions that help transnational communities to form (Deutsch 1957;Puchala 1970). In addition, boundary control is a measure of political development as capacity building. A 'durable shift in governing authority' (Orren and Skowronek 2004: 123) and the increase in centralised administrative, fiscal and coercive capacities are widely used attributes of political development.

War and boundary closure
For the analysis of boundary change, we build on elements of the 'rebordering Europe' framework (Schimmelfennig 2021). It assumes that boundary development is driven by the scale-community dilemma widely theorised in the literature on global and multi-level governance (e.g. Dahl 1994: 27-32;Hooghe and Marks 2016: 7-19). Open and functionally differentiated boundaries increase the benefits coming with scale. They typically improve factor allocation, allow for economies of scale, and enhance individual freedoms. At the same time, however, they tend to undermine the identity, solidarity, and security of political communities. Mass exits and entries weaken the bonds of identity, reduce the willingness of individuals to contribute to the public good and engage in social sharing, and increase their opportunities to avoid taxation and redistribution. Finally, open boundaries governed by fragmented authorities and low control capacity fail to fulfil their protective function and compromise the security of the community. We assume that it takes a 'boundary shock' to produce boundary change. Transboundary crises (Boin et al. 2014;Christensen et al. 2016) are the most important sources of such shocks. They originate on one side of the boundary and generate a sudden and sizable change in the nature and rate of cross-boundary transactions, thereby reducing boundary performance, i.e. the ability of boundaries to regulate entries and exits as intended. The transnational credit squeezes that triggered the eurozone crisis, the surge in migration movements across the Mediterranean that produced the EU refugee crisis, and the Covid-19 pandemic are recent transboundary crises that created boundary shocks to the EU and its member states.
Boundary shocks trigger a political process of boundary contestation and reconfiguration. Depending on the nature of the boundary shock, its effect on boundary performance and the politicisation it generates, boundary change takes different directions. One distinction is between 'debordering' and 'rebordering' (Popescu 2012: 69-77). 'Debordering' refers to the opening of boundaries, the reduction of boundary controls, and the territorial and functional differentiation of boundaries. By contrast, 'rebordering' implies higher entry and exit barriers, the strengthening of boundary control and the concomitant capacities, and more boundary congruence. In addition, the bordering process can lead to a reconfiguration of either the internal boundaries or the external boundaries of a polity.
If transboundary crises expose scale deficits, they create debordering pressures. For instance, economic stagnation may generate demand for the expansion of markets, and a pandemic may generate calls for international research facilities and joint procurement of medical equipment. Conversely, if transboundary crises challenge the identity, autonomy, or security of a political community, they typically generate rebordering pressures. Demand for the closure, congruence, and enhanced control of boundaries arises.
Military transboundary crises potentially expose both scale deficits and community threats. Bellicist theories of state formation typically attribute the prevalence of the territorial, Westphalian state to the war-making necessity of establishing a central monopoly over the extraction of militarily relevant resources and command of the means of physical violence -and to the disappearance of polities that were too small and decentralised to survive (Tilly 1990;cf. also Spruyt 1994). According to realist balancing theories (Walt 1987;Waltz 1979), states faced with military threats initially strive to increase their own military capabilities to re-establish the balance of power. They join military alliances (external debordering) if national capabilities alone prove insufficient.
At the same time, military crises constitute the ultimate challenge to the security, autonomy, integrity, and potentially identity of a political community. In response, externally threatened polities resort to a clearer demarcation of their boundaries with the attacker, raise the barriers for the movement of people and products, and establish stronger boundary control. External rebordering does not only affect the military boundary. To the extent that open economic, political, and cultural boundaries are perceived to provide the attacker with resources and influence, they experience rebordering as well (Gowa 1995).
Moreover, military threats and wars relate to social dynamics of in-group inclusion and out-group exclusion (Wimmer 2013). First, internally and externally differentiated integration (Schimmelfennig and Winzen 2020), which creates fuzzy and malleable boundaries between members and non-members, gives way to pressures for a clear demarcation of insiders and outsiders, the abolishment of grey areas, and uniform integration. Second, whereas the attacker is subjected to 'othering' , ostracism, and social exclusion that cuts across the board of functional boundaries and cross-border interactions, the attacked benefits from the collective identification and solidarity of the community that manifests itself in the opening of boundaries. Finally, security crises shift the scale-community dilemma towards the community pole. Faced with a fundamental threat to their order, polities renounce the scale benefits accruing from flexible and open boundaries and pragmatic transactions with the attacker but invest in the defence of their autonomy and integrity and in the solidarity among its members. External rebordering thus consists in a dual movement of boundary reconfiguration: closure against community outsiders, and openness and congruence among community insiders.

War and boundary control
In addition to these general expectations, any explanation of the EU's boundary development in reaction to military transboundary crises needs to take into consideration the peculiar origins and nature of the EU polity as a 'regulatory state' (Majone 1996). After the failure of the European Defense Community in the mid-1950s, European integration left the path of federalist polity formation built around common military capacities. Henceforth, the common defence of the member states was organised in NATO, whereas the European Communities focussed on economic integration. Supranational European institutions have played a predominantly regulatory role in making a common market. Whereas they are delegated an important legislative and judicial enforcement role to remove at-the-border and behind-the-border barriers to the movement of persons and products and to ensure fair competition, they lack major administrative, fiscal, and coercive capacity. The administrative implementation of EU regulation, the compensation of economic losers through the welfare state, and the means of macroeconomic stabilisation remained with the member states (Majone 1996). EU external governance also focuses on the extension of the internal market and the projection of regulatory rules (Bradford 2020;Lavenex and Schimmelfennig 2009). Even when functional pressures arising from the completion of the internal market incited the EU to venture into areas of 'core state powers' (Genschel andJachtenfuchs 2014, 2016), it upheld this emphasis on regulation over capacity-building. It created the common currency together with a regulatory regime for national budgets but refrained from extracting the resources necessary to forge a common fiscal policy, and it created the Schengen free-travel area together with a regulatory regime for asylum, law enforcement and visa policies but refrained from the creation of administrative capacities for supranational asylum management and coercive capacities for the control of external borders.
This developmental path has a significant impact on EU boundary formation policies -and in particular on its reaction to military crises. The EU focuses on the regulation of boundary closure. It legally introduces and enforces the rules that regulate the entry and exit of persons and products across its internal and external borders in a large majority of functional domains. For this purpose, it enjoys considerable legislative authority, coupled with high judicial authority, over boundaries. By contrast, the executive control of boundaries, the implementation of EU boundary rules and their everyday application, remains largely with the member states. This is also true for the administrative, financial, and coercive resources required for effective boundary control. Whereas the EU often plays a complementary role, providing technical facilities, information, and financial support, the bulk of the border control 'hardware' is national.
This division of labour in boundary control is subject to change, especially when transboundary crises expose capacity deficits. In response to the 2015 refugee crisis, for instance, the EU has increased its financial help to the most affected member states, its administrative support and monitoring through the European Union Asylum Agency, and the budget for Frontex, its external border enforcement agency, to enable it to purchase its own equipment and expand its staff. Yet because most member states organise their collective defence and the pooling of the relevant military capacity in NATO, the EU is unlikely to react in a similar way to military crises like the Russo-Ukrainian war. We therefore expect the EU response to the Ukraine crisis to consist predominantly in the regulatory closure and opening of boundaries, rather than an increase in the supranational executive boundary control authority and capacity.
To sum up the empirical implications of our theoretical considerations, we expect to observe three major patterns in EU external rebordering in response to the Russo-Ukrainian war. First, the EU-Russian boundaries are increasingly closed and cross-border transactions decrease. Second, EU-Ukrainian boundaries become increasingly open and cross-border transactions increase. Correspondingly, boundary congruence between the EU and Ukraine increases. Taken together, the closure of the EU-Russia boundaries is the mirror image of the opening of the EU-Ukraine boundaries, indicating the exclusion of Russia from, and inclusion of Ukraine into, the EU's political community. Finally, EU-internal boundaries and EU control of the boundaries with Russia and Ukraine remain largely stable. Political development in reaction to the Russian attacks on Ukraine changes the contours of the EU's political community, whereas capacity-building is absent or weak.

Case and data
In this section, we briefly introduce the historical background and the conceptual foundations of our empirical analysis. After the dissolution of the Soviet Union, the EU offered all its successor countries Partnership and Cooperation Agreements (PCA), assuming they would embark on a transition to democracy, market economy and closer cooperation with the EU. The PCAs with Russia and Ukraine were signed in 1994, entered into force in 1997 (Russia) and 1998 (Ukraine) and resembled each other in their basic structure and aims. Before the Russo-Ukrainian war, they remained the main legal basis for the relations of both countries with the EU. They established a framework for dialogue and cooperation in a broad array of sectors and provided for the liberalisation of trade but did not include a free trade area or regulatory alignment with the EU.
In the 2000s, the paths of Russian and Ukrainian relations with the EU started to diverge. Whereas Ukraine became part of the European Neighbourhood Policy (ENP, 2004) and the Eastern Partnership (2009), which sought to promote gradual and differential sectoral integration with the EU without the prospect of membership, Russia insisted on its own separate format based on symmetrical and high-level intergovernmental relations -and not predicated on the adoption of EU rules. Instead, in 2005, the EU and Russia agreed on 'road maps' for 'common spaces' , which mirrored the ENP in scope but produced few tangible boundary changes (except for readmission and visa facilitation agreements in force since 2007). By contrast, in 2008, the EU and Ukraine started negotiations on an Association Agreement (AA) including a Deep and Comprehensive Free Trade Agreement (DCFTA), which committed Ukraine to aligning its legislation to that of the EU in return for increasing access to the EU market and participation in EU policies.
The conclusion of the agreement became a matter of domestic and international contestation and the immediate prelude to the Russo-Ukrainian war. The EU halted the signing of the treaty over the authoritarian turn of the Ukrainian government under President Yanukovych, whereas Russia threatened economic sanctions should Ukraine choose EU association over a customs union with Russia. In November 2013, Ukraine rejected concessions to the EU and stopped preparations for signing the association agreement. This policy reversal caused the Euromaidan protests that removed President Yanukovych from power in February 2014, installed a pro-European government and triggered the Russian military intervention. In early March 2014, the EU condemned 'Russia's unprovoked violation of Ukrainian sovereignty and territorial integrity' , applauded 'the courage and resilience shown by the Ukrainian people' and vowed to 'stand by Ukraine' . 1 In the same month, the EU and Ukraine signed the political provisions of the association agreement.
To investigate how the war in Ukraine affected the European Union's institutional development, we collect novel data on the internal boundaries between EU member states and the Union's external boundaries vis-à-vis Ukraine and Russia. 2 We choose 2013 as the starting point of our investigation to assess over-time changes from the pre-crisis period over the 2014 Russian invasion to the full-scale attack of 2022. Our unit of analysis is the country-year-dyad. This means that we gather annual data on the configuration of three territorial boundaries: between EU member states, between the EU and Ukraine, and between the EU and Russia. Even though there is some variation among EU member states, most notably due to the differentiated integration of the Area of Freedom, Security and Justice, we treat the EU as a unitary actor. Per country dyad, we thus have ten points at which we measure the boundary configuration, from 2013 to 2022. While we assess the boundary configuration at year-end for the period between 2013 and 2021, we code the values for 2022 at the beginning of August. The 2022 values are thus to be seen as preliminary and potentially subject to change until year-end.
We follow the pertinent literature (Bartolini 2005;Rokkan 1974;Schimmelfennig 2021) in distinguishing four functional boundaries of political systems: economic, political, cultural, and military. In a first step, to operationalise these abstract concepts, we define the objects (goods/artefacts) and subjects (persons) which cross each of these four boundaries and whose movements are affected by a given boundary's institutional configuration. In line with our codebook, the political boundary of a territorial unit, for instance, affects the movements of four types of subjects (residents, refugees, non-governmental political agents, government officials) and one object (personal data). The military boundary, to take another example, regulates the cross-border movements of two subject types (military personnel, police/law enforcement personnel) and two types of objects (dual-use technology and military weapons). As we equally specify boundary-crossing subjects and objects for the economic and cultural boundary, our dataset encompasses all types of objects and subjects regulated by EU law referring to Ukraine and Russia between 2013 and 2022. We thus cover all relevant functional boundaries. All in all, our dataset comprises twelve subject types and nineteen object types. Ultimately, this provides us with 930 country-year-dyads (N = 930). Gauging a given boundary's openness for cross-border movements in this manner allows us to trace the evolution of the EU's external and internal boundaries over a period of ten years. We provide a full list of the boundaries analysed in this article in Table A1 of the appendix.
In a second step, we code the level of boundary closure for each of these boundary-crossing subjects and objects, i.e. a boundary's permeability for cross-boundary movement. Per subject/object, we code the level of closure twice, once for movements from the respective other territorial unit into the EU (entry), once for movements from the European Union into the respective other territorial unit (exit). To this end, we constructed a six-point categorial scale that ranges from fully open (0) over mostly open (1), partly open (2), partly closed (3), and mostly closed (4), to fully closed (5). 3 In order to assess the value of closure for each dyad, we screen the relevant primary, secondary and tertiary legislation of the European Union. The cross-boundary movements of one and the same object or subject can be differently regulated depending on the boundary at hand. For instance, while the cross-border supply of steel and its derivatives is subject to internal market regulation between EU member states, it is regulated by the existing association agreement between the EU and Ukraine, and it is affected by the current sanctions regime in transactions between the EU and Russia. We code the level of boundary closure based on the relevant regulation that applies to a given boundary-crossing subject or object at a given point in time. Note that, since we are interested in the EU reaction to the Ukraine war, we only code the EU regulation of the boundaries and do not take into account Russian and Ukrainian measures. For instance, whereas the EU did not restrict agricultural exports to Russia after the annexation of Crimea, Russia imposed sanctions on imports from the EU. From an EU point of view, exit has thus remained relatively open, whereas in fact the boundary has been closed for these products.
In a third step, we assess the degree of the EU's control over an internal or external boundary, i.e. the centralisation of boundary control authority in the EU's multilevel system. To this end, we distinguish between legislative, executive, and judicial control. 4 • Legislative control: Our categorial measure of legislative control encompasses both the pooling of decision-making powers in the Council and the degree of task delegation to supranational bodies (Börzel 2005;Leuffen et al. 2013). It ranges from an absence of EU-level policy coordination (0), over intergovernmental coordination (1), intergovernmental cooperation (2), joint decision-making under unanimity (3), joint decision-making under qualified majority (4), to supranational centralisation (5). • Executive control: We distinguish two dimensions of executive control. The first dimension, executive competence, describes the legal authority to implement boundary legislation. It ranges from national competence (0), over intergovernmental (1) and joint competence (2), to supranational competence (3). The second dimension of executive resources describes the origin of the fiscal, administrative, and technical capacities used to implement boundary legislation. It ranges from national capacities (0), over supporting (1) and joint capacities (2), to supranational capacities (3). • Judicial control: Lastly, we assess whether the treaties foresee the possibility of infringement procedures and preliminary references in the regulation of a given boundary. We then distinguish between national jurisdiction (0) and EU jurisdiction (1).
Note that boundary closure and boundary control are conceptually and empirically independent from each other. Whether boundary control remains with the national level or is transferred to the Union does not imply more or less permeability for cross-border transactions, and vice versa.

Empirical analysis
In the following, we present the results of our empirical analysis. We first turn to the impact of the Ukraine crisis on the dyadic boundary relations between the EU and Russia and Ukraine, respectively. In a second step, we investigate the degree of control that the EU enjoys over its external boundaries, and the dynamics of the relationship between closure and control.
Step 1: closure Our data shows that Russia's aggression against Ukraine has had a sizable impact on the external boundaries of the European Union. Figure 1 shows the evolution of the average closure levels across the three boundary dyads of interest. The picture that emerges corroborates our broad initial expectations.
First, the EU has increasingly excluded Russia from participating in the European political community. Over the observed period of ten years, the average closure level of the EU-Russia boundary has moved from the category of partly open (2013 value: 2.21) to partly closed (2022: 3.04). The bulk of this movement towards an increased impenetrability of the EU-Russia boundary occurred after Russia began its all-out war against Ukraine in 2022.
Second, the EU has increasingly included Ukraine in its political community. While the closure of the EU-Ukraine boundary was nearly on the same level as the EU-Russia boundary at the beginning of our observation period, it decreased steadily ever since. In terms of our categories, the EU-Ukraine boundary went from partly open (2013: 2.09) to mostly open (2022: 1.43). This movement towards increased openness occurred in a relatively continuous fashion, with the entry into force of the EU-Ukraine association agreement in 2016/17 and the 2022 war the most important moments of opening.
Third, the internal boundaries between EU member states have remained largely unaffected by the prolonged crisis. We provide a detailed overview of the EU's internal boundary configuration in appendix Figure A1. Throughout the entirety of the observation period, the boundaries of the European political community stayed constantly open (2013: 0.77, 2022: 0.70). One notable boundary opening in response to the Ukraine crisis is provided for in the Gas Supply Security Regulation of 2017 (Regulation 2017(Regulation /1938. This act repealed an earlier regulation of 2010, after a 2014 stress test 'on the short-term resilience of the European gas system analysed the effects of a partial or complete disruption of gas supplies from Russia' and revealed sizable deficits (Recitals 4 and 5 of Regulation 2017/1938). Whereas the 2010 regulation only stipulated that national emergency measures must not harm gas supply to other member states, the 2017 regulation demands active solidarity: as a last resort, EU member states must now help each other out to guarantee supply to the most vulnerable gas consumers.
Another notable exception is the slight increase in the congruence of the EU's external boundaries in response to the crisis. Denmark ended its treaty-based opt-out from the Common Foreign and Security Policy (CFSP) in 2022 and enacted a law informally mirroring the Temporary Protection Directive, from which it continues to have a formal opt-out. The EU also decided sanctions against Russian individuals consensually even though some member states do not fully participate in Schengen or have opt-outs from justice and home affairs policies. These formal and informal moves towards boundary congruence are in line with the expectation of external rebordering in response to military threats.
We will now look at the EU's respective boundary dyads with Russia and Ukraine in more detail. Here, we also refer to the development in cross-boundary transactions to show that changes in the rules of closure have had practical consequences.

EU-Russia boundary
We first turn to the dyadic boundary relation between the EU and Russia ( Figure 2). When eyeing the over-time development of the average level of boundary closure, it is immediately apparent that the EU's reaction to the Ukraine crisis of 2014 was relatively reluctant. After the Russian annexation of Crimea in March, the EU began to gradually phase in a sanctions regime, which initially targeted involved political and military officials with travel bans and asset freezes. Due to disagreements among member states in an area requiring unanimity (Howorth 2017;Karolewski and Cross 2017;Orenstein and Kelemen 2017), it took the EU until late June 2014 to impose broader economic restrictions. Only the downing of the passenger plane MH17 in mid-July served to galvanise the EU's crisis reaction. Within days, the EU agreed on a comprehensive package of sanctions limiting access to EU capital markets for state-owned Russian banks, imposing an embargo on exports of arms and dual-use goods, and limiting Russian access to sensitive technologies in especially the oil sector (Seibel 2017).
The EU's relatively subdued reaction to the 2014 crisis is also reflected in the constant, and sometimes even increasing, number of boundary transactions between the EU and Russia in the years between 2014 and 2021. The volume of bilateral trade in industrial and agricultural goods as well as services remained by and large unaffected. Commodity trade, and especially EU imports of natural gas, coal, oil, and steel, even witnessed steep increases after 2014. Only a handful of sectors took substantial and sustained hits from the EU sanctions and the general souring of the bilateral relationship. The import of financial services from Russia decreased significantly and bilateral flows of foreign direct investment, both from and to Russia, dwindled. According to SIPRI data, EU member states have indeed exported no arms to Russia since 2014 (although Greece purchased Russian surface-to-air missiles in 2019). Lastly, while the number of Russian students enrolled in European universities and the number of Russians applying for asylum in the EU remained relatively constant, less and less Russians were granted Schengen tourist visas (2013: 6,893,346; 2019: 4,054,086). This was long before the EU began mulling a general visa ban for Russian citizens in the summer of 2022. When Russia launched its full-scale attack against the whole of Ukraine in February 2022, the EU was well-prepared and acted as swiftly as decisively to close its external boundaries. In marked contrast to 2014, the EU immediately adopted comprehensive and incisive sanctions that touched upon the bulk of boundary transactions between the EU and Russia. Beyond asset freezes and diplomatic restrictions against an ever lengthier list of individuals (extending its sanctions list from 203 individuals in December 2021 to 1229 in July 2022), the various sanctions packages adopted in 2022 mandate, inter alia, the freezing of Russian central bank assets, the exclusion of most Russian banks from the SWIFT international payments system, a ban on direct investments from and in Russia, a prohibition on almost all Russian goods and services imports, a prohibition of high-technology and strategically important exports, a prohibition of coal and steel imports, a ban on seaborne oil imports, a closure of the EU's airspace for Russian aircrafts and of European ports for Russian ships, and a ban of Russian state-owned media in the EU. For most categories of subjects and objects, it is still too early to tell the impact of the 2022 EU sanctions regime on boundary transactions, though it is likely to be significant. There are, however, signs that the new sanctions will weigh more heavily on actual transactions than the 2014 regime. Even though, for instance, the EU has not banned imports of natural gas from Russia and, in the March 2022 conclusions of the European Council, 5 has pledged independence of Russian gas by 2030 only, imports of Russian gas during the first half of 2022 have sunk to the lowest level since the beginning of our observation. While natural gas imports from Russia increased heavily after the 2014 crisis, peaking at an average 14,072 million cubic metres per week in 2019, this number decreased to 7724 million cubic metres in the first seven months of 2022.
Our data also provides a variety of further insights into the development of the EU-Russia boundary since 2014. First, the EU has consistently restricted entry from Russia (2022: 3.76, mostly closed) more strongly than exit to Russia (2022;2.32, mostly open), whether for objects or subjects. This is possibly motivated by a desire to curb Russian profits while safeguarding as many of Europeans' (business) opportunities in Russia as possible. Again, we did not code Russian countersanctions, which indeed imposed mirroring boundary restrictions on entries into Russia in many instances, but the incongruence of entry and exit closure on the EU side remains. Second, until 2022, the EU has consistently treated the entry of subjects from Russia (2021: 3.46, partly closed) more restrictively than the entry of objects (2021: 2.96, partly closed). Only in reaction to the war, the EU closed its boundary for objects much more comprehensively (2022: 3.95, mostly closed) while keeping the boundary for subjects untouched. In general, apart from the travel bans imposed on an increasing number of Russian state officials, all EU boundary closures since 2014 have affected the movement of objects rather than subjects. This might reflect a desire to limit Russia's economic benefits from its transactions with Europe while keeping the door open for societal exchange. If, however, future sanctions should impose a Schengen visa ban for Russian citizens, the EU-Russia boundary would become significantly more impermeable for subjects, too. Third, the EU has consistently treated transactions over the military (2022: 4.5, fully closed) and, at least for entry, cultural (2022: 4.0, mostly closed) boundaries more restrictively than transactions over the economic (2022: 3.5, mostly closed) and political (2022: 3.0, partly closed) boundary. The high levels of closure of the military and cultural boundary reflect the EU's desire to halt the Russian military effort and the spread of disinformation within the Union. At the same time, all functional boundaries between the EU and Russia have become less permeable in reaction to the 2022 war, with the economic boundary seeing the steepest increase in closure (2021: 2.5, partly closed; 2022: 3.5, mostly closed). As of August 2022, all functional boundaries of the EU are either partly closed (economic, political), mostly closed (cultural) or fully closed (military) for entries from Russia.

EU-Ukraine boundary
In terms of its average level of closure, the EU boundary with Ukraine has become significantly more open since 2014, providing a mirror image to the boundary with Russia ( Figure 3). The EU has increasingly allowed Ukraine to participate in the European political community, culminating in the country's designation as a candidate for EU accession in 2022. The opening of the EU-Ukraine boundary proceeded in steps, rather than solely in reaction to the Russian invasion of February 2022. The first opening occurred with the entry into force of the EU-Ukraine association agreement (AA) in 2016 (trade-related provisions) and 2017 (full AA). The AA heralded a movement towards more openness of the EU-Ukraine boundary (2015: 2.08, partly open; 2017: 1.83, partly open), and did so across a variety of different functional boundaries. The DCFTA it contained affected a wide range of economic transactions and eased boundary restrictions for both subjects (service providers) and objects (industrial goods, agricultural goods, commodities, digital services, investment). The AA equally facilitated transactions across the political (political agents, government officials) and cultural (students, tourists, family) boundaries. The granting of visa-free entry into the Schengen area to Ukrainian citizens in 2017 had a particularly profound effect on the openness of the political and cultural boundaries between the EU and Ukraine. The military boundary, in turn, remained untouched and mostly closed (2017: 4.0).
The increasing openness of the EU-Ukraine boundary in the years after 2014 is also reflected in an increasing level of economic, political, and cultural transactions between both sides. Bilateral exchanges of industrial and agricultural goods as well as services reached new heights in the years after 2014. The same goes for the direct investments of European companies in Ukraine. In 2012, Russia was still Ukraine's most important trading partner, accounting for 26 percent of its goods exports and 32 percent of imports. In the meantime, the EU has become Ukraine's most important trading partner by a wide margin, representing 43 percent of Ukrainian goods exports and 41 percent of its imports. Cultural exchanges between the EU and Ukraine increased, too. Millions of Ukrainians have profited from visa-free entry into the Schengen area since 2017, and the number of Ukrainian students enrolled at European universities has increased significantly, from roughly 18,000 in 2013 to 49,000 in 2017. In addition, political boundary transactions between the EU and Ukraine have increased. Since 2014, the EU advisory mission Ukraine (EUAM) has supported the country in reforming its civilian security sector. Between 2014 and 2021, the EU also provided more than 5 billion Euros in macro-financial assistance to Ukraine, adding onto roughly 2 billion Euros of humanitarian aid. The second and more momentous step towards an opening of the bilateral boundary between the EU and Ukraine occurred in reaction to the Russian invasion of February 2022 (1.42, mostly open). This time, the opening also affected transactions across the military boundary. While the military boundary has remained fully closed for subjects (military personnel), the EU relaxed its former restrictions on arms exports to support the Ukrainian military effort against the Russian invasion forces. Already from 2018 onward, according to SIPRI data, a handful of EU member states (Czechia, Estonia, France, Lithuania, Poland) had begun to deliver mostly light weaponry to Ukraine. Responding to the Russian invasion of 2022 then, the EU coordinated its response and all EU member states (except for Cyprus, Hungary, and Malta) began providing military equipment and weapons bilaterally to Ukraine, with three member states (Austria, Bulgaria, and Ireland) providing equipment only (Clapp 2022). In addition, the EU used the 'European Peace Facility' (EPF) to channel 2.5 billion Euros directly into Ukraine for the purchase of lethal weapons and other military equipment. The Council had created the EPF as an off-budget instrument in 2021, to finance EU military missions and to assist third states or international organisations with defence-and security-related expenses. For the current budgeting period (2021-27), the EPF is allocated 5.7 billion Euros.
The EU also eased many persisting restrictions on the economic, cultural, and political boundaries. In early March 2022, the Council invoked the Temporary Protection Directive (Council Directive 2001/55/ EC) for the first time since its conclusion in 2001. Under temporary protection, all persons displaced from Ukraine enjoy a right to residence of up to three years in the EU, access to employment, health care, social assistance, housing, and the freedom to move freely within the Union. As such, the temporary protection directive affects the freedom of entry for subjects across the economic (workers, service providers), political (refugees, residents), and cultural (family, students, tourists) boundaries. According to UNHCR data, more than 3.8 million Ukrainians have received temporary protection in the EU as of August 2022. In addition to its EPF funding, the EU has also provided another 2.2 billion Euros in macro-financial assistance, 120 million Euros in grant support for state and resilience building, 500 million Euros in direct budget support, and 867 million Euros in humanitarian aid. The EU also removed most of its persisting restrictions on the movement of economic objects. In March already, the EU integrated the electricity grids of Ukraine and Moldova into the Continental European Grid, which also allows for reverse flows and safeguarding Ukrainian energy supply from Europe. In May, the EU temporarily removed all persisting restrictions, i.e. tariffs and quotas, on imports of industrial goods, agricultural goods, and commodities from Ukraine.

Step 2: control
In a second step, we look at the control dimension of the EU's internal and external boundaries. To this end, we construct a control index that ranges from 0 (= full national control) to 1 (= full supranational control). For the index, we standardise and sum all four control variables (legislative, executive competence, executive resources, judicial) and divide them by their overall number. Figure 4 provides an overview of the development of our control index for the three boundaries of interest across the entire period under investigation.
Three observations stand out: First, the degree of EU control over its external boundaries remains by and large constant. While the EU's control over its internal boundaries (0.61) and its external boundary with Ukraine (0.56) stagnates over time, it even decreases slightly for its boundary with Russia (2013Russia ( : 0.56, 2022. This goes to show that an increase in external closure does not necessarily go hand in hand with an increase in internal control. Second, the EU's control over the internal boundaries between member states is slightly higher than over external boundaries with third countries. This has to do with the legal basis of the respective boundary regulations. Whereas internal boundary transactions are regulated as part of the internal market and thus part of the core of supranational (legislative and judicial) control, the external relations of the EU, at least where they transcend the Common Commercial Policy (CCP), constitute a more intergovernmentally controlled 'core state power' (Genschel and Jachtenfuchs 2016) area. Third, over both its internal and external boundaries, the EU has attained a by and large medium level of supranational control (average control score over the entire observation period: 0.57). This medium level of control reflects the discrepancy between high standardised average scores for legislative (0.70) and judicial (0.86) control on one hand, and the relatively low levels of control over executive competence (0.30) and executive resources (0.41) on the other. This constellation is typical for the EU 'regulatory state' (Majone 1996) in which the supranational level enacts and adjudicates laws but the national level enforces them (see Figure A1 in the appendix for additional graphs on internal boundary control).
We also checked for bivariate correlations between our variables. The analysis corroborates our overall finding of the EU's regulatory boundary state. Figure 5 shows that we find merely a weak, and slightly negative, correlation (R 2 : 0.08) between the average level of closure and the average level of control. While legislative and judicial control are highly and positively correlated (R 2 : 0.78), we find only weak correlations for legislative control and both executive competence (R 2 : 0.12) and resources (R 2 : 0.00). This shows the divergence between supranational regulatory and national positive statehood in the EU multi-level system very impressively. Interestingly, executive competence and executive resources are only weakly correlated either (R 2 : 0.20), demonstrating that EU executive resources go hand in hand with both national and supranational competence. Finally, we find moderate negative correlations between entry and exit closure and legislative control (R 2 : 0.23 and 0.28). Hence, the more closed the external boundaries of the EU, the lower the level of legislative (and, by extension, judicial) supranational control. This corresponds to our observation that hard boundary closures (sanctions) fall under the CFSP and are thus primarily decided on an intergovernmental basis. The correlation is therefore stronger on the boundary between the EU and Russia (R 2 : 0.37; see the downward movement in Figure 4 and the positioning of the green triangles in Figure 5).

Conclusions
Russia's war against Ukraine has shaken Europe's post-Cold war order to the core. It has shattered Europeans' belief in the peaceful resolution of conflicts and the inviolability of borders, their confidence in economic integration as a way to overcome interstate rivalry, and their hope for an amicable coexistence between Europe and Russia. Vladimir Putin's attempt to annihilate the Ukrainian state and nation and to roll back the expansion of the liberal European international community and order has triggered swift and decisive changes in the EU's boundary closure policies vis-à-vis Russia and Ukraine, signifying a major shift in community building.
The threat of war is the most profound boundary shock a political system can be subjected to. The war against Ukraine threatened not only an associated country of the EU but has been widely interpreted as a direct threat to especially its Eastern and Northern European members. As expected by bordering theory, the EU has adjusted its external boundaries in response to this exogenous shock. Making out Russia as the aggressor and standing with Ukraine, the EU has progressively closed its formerly partly open economic, political, and cultural boundaries with Russia. It did so rather reticently in 2014, yet ever more decisively in 2022. In turn, the EU has gradually opened its economic, political, cultural, and ultimately also its military boundaries to Ukraine. A first important step was the 2016/17 entry into force of the EU-Ukraine association agreement. Russia's all-out invasion in 2022 then heralded a comprehensive opening across all functional boundaries. As much as the EU has excluded Russia from the European political community, it has closed ranks with Ukraine, culminating in the latter's designation as an accession candidate in June 2022. In Eastern Europe, the Russian invasion has put an end to the EU's strategy of external differentiation, in which the post-Soviet countries including Russia were to pursue flexible issue-specific integration with the EU below the threshold of membership. Instead, it has pushed the EU to sever ties with Russia (and its close ally Belarus) while committing to the eventual membership of Ukraine, Moldova and Georgia. The EU has both more sharply demarcated the limits of and increased the boundary congruence within its political community.
But has Russia's war also changed the course of European integration with regard to capacity-building political development? So far, our analysis suggests, there are few signs to believe so. According to the bordering analysis of political development, pioneered by Stein Rokkan (1974) and Stefano Bartolini (2005), a political system's ability to close and control its transactions with the outside world is the central precondition of its internal development. External closure, in this view, enables societal sharing and an effective production of public goods. This logic applies to the present-day EU as it has to the development of the European nation-states. While historically the EU has predominantly acted as an institutional device to deborder the nation-state, it has begun to gradually consolidate its external boundaries in reaction to the polycrisis of the 2010s (Schimmelfennig 2021).
Defying Rokkanian expectations, however, the closure of the EU boundaries with Russia has not been accompanied by a concomitant internal debordering and increase in supranational control. The boundaries between EU member states have remained largely unaffected by the crisis in its vicinity. The sole notable exceptions consist in the 2017 Gas Supply Security Regulation, which represents an important step towards internal debordering by demanding active solidarity among EU member states, and Denmark's decision to end its opt-out from the CFSP. The EU has also not acquired novel competences and resources to control its boundary transactions with Russia. Our data even shows a moderate inverse trend: Despite increased closure, supranational control has decreased slightly over time. This reflects the locus of the regulation of the EU-Russia boundary, which has shifted more and more from the supranationally governed realm of the Common Commercial Policy to the intergovernmentally controlled CFSP. Ultimately, the EU has reacted to the Ukraine crisis in line with its longstanding regulatory template. It decides over the closure of boundaries, but their control remains primarily in national hands.
We write this article roughly six months after the beginning of the war. Obviously, the evidence we present is of a preliminary nature. The duration and outcome of the war are highly uncertain. Longer-term effects, spill-overs and unexpected events might still vindicate bellicist or Rokkanian expectations. So, how likely is it that the conclusions we have drawn will stand the test of time? As for the EU's boundaries with Russia, we are confident they will remain closed for the foreseeable future. Their closure has acquired an enormous functional scope and regulatory depth. Bar regime change in Russia, a swift end to the war, or a grand post-war settlement over the European political order, none of which seems to be currently in the cards, their reopening remains unlikely. The EU's opening towards and inclusion of Ukraine seems equally enduring and will probably progress further. Having kicked off the accession process, the EU has firmly entered the territory of the 'community trap' (Schimmelfennig 2001), with Ukraine making a strong case that its fight is also in defence of Europe and its values.
The further development of the EU in response to the war seems more open-ended. So far, the Union has fared well with its regulatory approach and its reliance on national capacities. Russia's war has not posed a challenge to the inner workings of the EU as for instance the Euro crisis did, which endangered the very survival of a core achievement of European integration. It need not remain this way. In past crises, the EU relied initially on regulation, too, yet solely to the point where the costs of the regulatory status quo began exceeding the costs of supranational capacity-building. European states elites are no diehard sovereignty maximisers but, by and large, functional problem-solvers (Freudlsperger and Jachtenfuchs 2021). Where EU-level capacities seem inevitable to stem a crisis and regain control over the Union's external boundaries, elites tend to regard their creation as the lesser evil. The most likely candidate for internal change is, of course, energy policy. From joint gas procurement over novel EU-funded pipelines to debt-backed fiscal compensation, a lot seems possible if Russian disruptions continue to endanger Europe's energy supply. Supranational capacities could also serve to solve problems in other domains. In foreign policy, they could help overcome arising national divergences; in internal security policy, they could safeguard the proper enforcement of EU sanctions; in border policy, they could help to inhibit a potential intrusion of weapons, fighters, and saboteurs into the EU. What remains highly improbable though, against the backdrop of the reinvigorated presence of NATO, is any meaningful EU involvement in territorial defence. Even in the face of Russia's war machine, the EU is as unlikely to become a Westphalian state as ever.

Notes
boundary (1): explicit permission of nearly all cross-border provisions with light conditions, e.g. permanent authorization through 'passporting'; partly open boundary (2): explicit permission of a majority of cross-border provisions with moderate conditions, e.g. temporary authorization; partly closed boundary (3): explicit prohibition of a majority of cross-border provisions, permission only under exacting conditions, e.g. ad-hoc authorization; mostly closed boundary (4): explicit prohibition of nearly all cross-border provisions, permission only under exceptional circumstances; fully closed boundary (5): explicit and unconditional prohibition of all boundary transactions. 4. For descriptive summary statistics, see Table A2