External relations of an Industry 4.0 cluster: the case study of the Hamburg aviation cluster

ABSTRACT Based on the case study of the Hamburg aviation cluster (HAv), this paper touches upon the nature of cluster processes in digital transformation (Industry 4.0). The discussion is carried out with reference to the concept of hubbing, while highlighting the difference between classic internationalization. The aim of the paper is to unearth the nature of the cluster’s external expansion in the digitally reshaped era by investigating the case of an advanced aviation cluster officially branded as an Industry 4.0 cluster.


INTRODUCTION
This paper touches upon the process of cluster external relations in the context of digital transformation (4th Industrial Revolution; Industry 4.0; I4.0).As the literature on a mutual 'cluster-I4.0'relation is only emerging, more in-depth contextualized research seems necessary (Bailey et al., 2020;De Propris & Bailey, 2021;De Propris & Bellandi, 2021;Grashof et al., 2020;Hervás-Oliver et al., 2020;Sedita & Ozeki, 2021).I4.0, by enabling the fusion of real and virtual worlds, stands for the far-reaching digital transformation of business models (Kagermann et al., 2013).It heralds the manufacturing of self-organized, multi-agent systems with big-data base feedback and coordination; systems which are interconnected, integrated, autonomous and can work in real time (Culot et al., 2020).The pervasive and disruptive nature of I4.0 technologies redraws global value chains (GVC) in terms of value creation and geographical presence (Propris & Bailey, 2021), and the emerging literature demonstrates that I4.0 bears the danger of increasing rather than reducing regional disparities (Greef & Schroeder, 2021).
There is a clear need for more knowledge regarding the likely I4.0 impact on clusters that have been commonly seen as crucial for economic development (Malmberg & Maskell, 1999;Porter, 2000).The processes of the international expansion of clusters deserve special attention in this respect as they are considered necessary for innovativeness and competitiveness (Miguelez et al., 2019;Turkina et al., 2016).
Based on the case study of the Hamburg aviation cluster (HAv), this paper touches upon the nature of cluster processes in the time of digital transformation (I4.0).The discussion is carried out with reference to the concept of hubbing (Bathelt et al., 2004;Njøs et al., 2017), while highlighting the difference to classic internationalization.This paper aims to unearth the nature of the cluster's external expansion in the digitally reshaped era by investigating the case of an advanced aviation cluster officially branded as an I4.0 cluster. 1  The remainder of the paper is structured as follows.First, it introduces the concept of the Fourth Industrial Revolution in a cluster context.Next, it considers the peculiarities of the cluster external relations by highlighting the nuances of hubbing and internationalization.The following section describes the method applied and introduces the HAv.The next section discusses the nature of the international expansion of clusters in the I4.0 era by elaborating on the findings of the HAv case.The final section concludes.

THE FOURTH INDUSTRIAL REVOLUTION AND CLUSTER CONTEXT
I4.0 means a digital transformation of business models and new value creation (Muscio & Ciffolilli, 2019).It involves a set of interlinked and interdisciplinary technologies and heralds a far-reaching integration of processes and systems and the emergence of the industrial internet, integrated industry or collaborative production (Isaksen et al., 2020).Few reviews so far have covered the dynamic nature of this topic in a comprehensive manner (Maresova et al., 2018).I4.0 stands for the combination of real and virtual worlds and breakthrough innovation in production systems.It comprises a set of revolutionary technologies which can transform, if not disrupt, the markets.I4.0 is the result of complex interactions and coordination between technical and social aspects in pursuit of the digital transformation of production (Prodi et al., 2022).Far-reaching consequences of I4.0 imply that industry borders will become more blurred; competitive advantageonly transient, whereas ecosystemsnot a single firmwill be the new unit of competitive analysis (Lanteri, 2021).There is no consensus about what I4.0 stands for, but regardless of the definition used, scholars tend to agree that it will have an impact on the geography of knowledge production (Balland & Boschma, 2021).Lund and Vildåsen (2022), drawing on the sociology of expectations and regional path development, showed how expectations created by the global I4.0 narrative trickled down into national and regional policies.
Clusters, on the other hand, are geographical hubs of related companies representing a hybrid between pure hierarchies and markets (Maskell & Lorenzen, 2003).Morosini (2004, p. 307) defines a cluster as a 'socioeconomic entity characterized by a social community of people and a population of economic agents localized in proximity … who work together … to generate superior products and services'.Cluster membership is associated with an array of benefits (González-Torres et al., 2020).Grashof et al. (2019) show that clusters can provide a suitable environment for radical innovations, and such are usually perceived as related to I4.0.This can be attributed mainly to the existence of a specialized labour market and spillovers of tacit knowledge.Such diffusion is essential for collective learning and innovation (Otto & Fornahl, 2010).Knowledge transfers easier thanks to face-to-face daily contact are of utmost importance for radical innovations (Braunerhjelm et al., 2017) and firms tend to locate in particular places for the development of technological breakthroughs (Castaldi et al., 2015).
There are different channels through which clusters can facilitate the advancement of I4.0 (Götz & Jankowska, 2017;Hervás-Oliver, 2021).Manufacturing firms face rapid changes associated with 'I4.0' (Karlsen et al., 2019), but thanks to cluster location and intermediary's assistance, they can access and further spread this external knowledge among themselves, which seems of utmost importance in the dynamic I4.0 era.Idiosyncrasies of complex knowledge creation fundamental for I4.0 can be reconciled with unique features of innovative processes typical for clusters (Baker et al., 2021;Corradini et al., 2021;McPhillips, 2020;Micek et al., 2022).New business models triggered by the Fourth Industrial Revolution, epitomized by the concepts of a connected company with vanishing boundaries or digital business ecosystems, could be found in mechanisms associated with clusters (Balland & Boschma, 2021;Benitez et al., 2020;Jankowska et al., 2021;Li et al., 2021;Prodi et al., 2022).I4.0 implies a shift towards highly adaptive networks of interrelated entities and stipulates the reorganization towards the platform structures.In this context, clusters seem to be predestined to be nodes in such configurations (Bathelt & Li, 2014;Hervas-Oliver, 2019).They can also serve as promising policy tool organizing the implementation of the Fourth Industrial Revolution and securing more coordinated business digital transformation (Hervás-Oliver, 2021).For information technology (IT)-assisted and globally spread operations of I4.0 co-location remains still essential (Busch et al., 2021).Clusters can be harnessed as laboratories for I4.0 experiments (Götz & Jankowska, 2017;Hervás-Oliver, 2021).They offer a favourable environment stimulating knowledge creation and its further diffusion, function as a tool enabling the smooth implementation of advanced projects and can be seen as key centres in the emerging I4.0 landscape (Balland & Boschma, 2021;Miguelez et al., 2019).Previous research (its OWL cluster -Intelligent Technical Systems OstWestfalenLippea flagship I4.0 German cluster) suggest that we will witness mutually reinforcing interdependencies between clusters and I4.0 (Götz, 2021).On the one hand, cluster thanks to the advantages offeredpecuniary agglomeration economies including local labour market, knowledge spillovers, proximity or institutional framework can facilitate the I.40 implementation (Götz, 2019;Hilpert, 2021;Isaksen et al., 2020).Though idiosyncrasies of I4.0 (Yang & Gu, 2021) may impact the functioning of the cluster as they require a more interdisciplinary and integrative approach with the provision of industrial commons, that is, a bundle of regional and industry idiosyncratic assets (Pisano & Shih, 2009), and development of related varieties, that is, dynamic and complementary externalities originating in similar industries (Frenken et al., 2007).Besides, processes of cluster stretching facilitated by digital technologies should be harnessed for upgrading the core competences of cluster.Clusters might thus become more cross-sectoral and interregional which can materialize via hubbing (geographical scale) or blending (industrial scope) (Njøs et al., 2017).
Clusters may contribute to the development of I4.0, yet they are reshaped and remodified by the digital transformation as well (Pagano et al., 2020).Certain co-evolution is taking place (Benner, 2021;Blažek et al., 2019;Isaksen et al., 2020), which makes the enquiry into I4.0 and clusters relations even more relevant and urgent.

THE CLUSTERS' EXTERNAL RELATIONS: HUBBING AND INTERNATIONALIZATION
This paper posits that hubbing and internationalization are two different (though not mutually exclusive) types of external cluster relations.When initiating this study, these relations were originally conceived in terms of classic internationalization, yet the empirical findings and rich material gathered showed that they may differ from commonly applied concepts of export and import or inward and outward foreign direct investment (FDI).In fact, they draw on multiple collaboration projects, alliances of knowledge exchange but also formal, institutionalized collaborations under umbrella programmes or platforms.
The literature on external relations of clusters is undoubtedly rich but seems to be rather incoherent and disorganized, often mixing different terms.Given this haziness, when discussing the cluster external relations, one needs to take a more nuanced approach and differentiate between hubbing and internationalization (Figure 1).
The latter implies revealed competitiveness, it aims at exploiting competitive advantages and means that local firms have a competitive edge, whereas the localization is attractive.Such an approach is reflected in the rich literature on local firms expanding abroad and/or foreign investorsinward FDI (IFDI) and outward FDI (OFDI)coming in (Belussi & Hervás-Oliver, 2017;De Marchi et al., 2019;Dohse et al., 2018;Pietrobelli & Rabellotti, 2011;Pla-Barber & Alegre, 2007).On the other hand, hubbing involves the need to sustain or increase competitiveness.It aims more at exploring competitive advantages.Thus, it takes the form of concluding partnerships, and alliances, creating observatory membership, seeking foreign knowledge, sourcing foreign know-how and hence developing industrial commons.Such an approach is covered not only by the literature strictly devoted to hubbing seen as scale expansion and part of stretching (Bathelt et al., 2004;Njøs et al., 2017) but also by the novel studies on cluster internationalization (Gold, 2019;Grashof et al., 2021) or inter-clustering, cluster-to-cluster (C2C) (Goerzen, 2018).
Hubbing, which signifies the expansion of scale (Njøs et al., 2017), is a process that clusters can undergo and represents one of two dimensions of stretching, besides blending, that is, expansion in the industrial scope.The literature on clusters stressed that this category has expanded from its original concept leading to some detachment from Porter's genuine concept (Fløysand et al., 2012;Malmberg & Power, 2006;Martin & Sunley, 2003;Tödtling & Trippl, 2005).Already the Leading-Edge Cluster Competition programme (Abbasiharofteh 2020) can best illuminate the open interpretation of the cluster concept as the size of their geographical areas and distances within cluster boundaries differ substantially among clusters (Cantner et al., 2015).Hubbing could also be seen in the context of inter-clustering (Goerzen, 2018;Lorenzen & Mudambi, 2015).Focused on knowledge-sharing and learning as specific inter-organizational relations (Franco & Esteves, 2018) it contributes to regional competitiveness (Dohse, 2007;Schüßler et al., 2013).
Radical innovations with which I4.0 is commonly associated imply that knowledge pieces are combined in a new way (Fleming, 2001), which may happen when inventors merge their experience with external expertise (Desrochers, 2001).These external relations might refer to various stages of knowledge generation and collaborative relationships can encompass the exploration, integration and exploitation of new ideas (Balland et al., 2013).Richardson et al. (2012), studying the more formal clusters, argue that the diffusion of the internationalization of knowledge can take place via regular events.Thanks to the fact that firms meet regularly, they get greater familiarity and trust with one another, which in turn may enable exchanging knowledge and contribute to internationalization.
Hubbing might be related to the concept of strategic coupling of regional and extra-regional assets which leads to the emergence of new economic activities (Chandrashekar & Bala Subrahmanya, 2019;MacKinnon et al., 2019).Strategic coupling explains the global-local articulation of production networks (Yeung, 2018) and can be understood as a process that occurs when complementary effects between regions and global production networks (GPNs) spread over time.But hubbing, as one dimension of external relations, needs to be distinguished from cluster strategic coupling, which emphasizes the (mis)match between cluster assets and the lead firm's needs (Hassink, 2021) and recently aims at reinvigorating the stagnating debate on cluster development path (Harris, 2021).Hubbing seems to ignore the needs of the lead company, can be positive or negative, and does not always have to follow the strategic consideration aiming at 'coupling', and might be depicted by 'simple' city sprawl.De facto strategic coupling can be characterized by reference to both internationalization and hubbing, as it remains agnostic about exploitation or exploration of competitive advantages and can accommodate both assets exploitation and exploration, although these differences do not seem well articulated.The approach proposed in this paper may thus enrich the discussion on strategic coupling which emerges as a catch-all phrase, an extremely voluminous and wide concept (Henry et al., 2021) warranting the need to develop and propose some classifications.
Summing up, a broader set of 'cluster external relations' encompasses classic internationalization and hubbing (see Appendix A in the supplemental data online).The latter as geographical expansion (along with blending, i.e., industrial expansion; Njøs et al., 2017) may intensify as a result of digital transformation I4.0 (Götz & Jankowska, 2017) yielding less concentrated and more diversified clusters, which epitomize the expected 'I4.0-facilitated' modification of the cluster concept (Bathelt & Li, 2014;Turkina et al., 2016).

THE CASE OF THE HAV CLUSTER: METHOD APPLIED AND INTRODUCTION
Adoption of a narrative literature review derives from the scope of this statu nascendi phenomenon studied (Gancarczyk, 2019).Besides a critical literature review, this paper draws on the qualitative empirical method, which can produce rich descriptions, explanations and interpretations of phenomena (Michailova & Mustaffa, 2012;Ryan et al., 2020).This research is exploratory because I4.0 and cluster studies have not been the subject of comprehensive and systematic analyses.A case study approach has been used with partially structured, in-depth interviews with experts as the primary source of empirical data.This qualitative approach enables researchers to explore new and complex phenomena (Dominguez & Mayrhofer, 2017;Zaefarian et al., 2016) while the case study offers valuable contextual insights (Martineau & Pastoriza, 2016).
As stated by Mees-Buss et al. ( 2019), the purpose of a case study is not statistical generalizability, but to offer analytical generalizability, that is, providing an opportunity to understand the nature of phenomena, revise or refine existing concepts, and explain how and why events occur.A phenomenon-driven selection of the case study (Fletcher et al., 2018;Schwarz & Stensaker, 2014) is inspired by concrete phenomenon (in this case, I4.0 relation to clusters, in particular, the official label of the I.40 cluster).The chosen approach can be classified, as suggested by Fletcher et al. (2018), as the third path based on an earlier theory, which plays a sensitizing role and is used in a very loose way to direct research, offering a specific perspective.Following the Flyvbjerg (2006) classification, this paper builds on information selection that maximizes the usefulness of information from small samples.In-depth interviews with experts, analysed inductively, provide deep-seated information and help to answer the 'how' and 'why' questions that relate to the nature of the phenomenon (Eisenhardt & Graebner, 2007;Yin, 2009).Certain scepticism towards the qualitative approach and case study permeates the literature on international business, economic geography and also regional studies.Nevertheless, there seems to be growing recognition as to the value-added they might offer (see Appendix B in the supplemental data online).Calls for more methodological pluralism are now accompanied by a growing perception of the need for theoretical pluralism which in a way rehabilitates the status of qualitative approach and case study (Cornelissen et al., 2021;Reuber & Fischer, 2021;Tsang, 2022;Welch et al., 2022).The presented study should fit into this discussion and could be seen as perhaps part of this ongoing debate.The aim of this research was exploration drawing on the induction/abduction approach rather than the verification of concrete hypotheses or confirmation of existing theories (Welch et al., 2022).Knight et al. (2021) argue that scholars have favoured confirmatory approaches in deductive theory-building, whereas there is a clear need for more exploratory research using both qualitative and quantitative data.This paper could add to this ongoing vivid debate on the role of theorizing from case studies.
The Hamburg aviation cluster (HAv) has been chosen for the analysis because it is officially one of 40 'I4.0 branded' clusters out of more than 400 identified by Cluster Platform Deutschland. 2 It is a winner of prestigious Leading-Edge Cluster Competition (Cantner et al., 2015) and honoured with various awards.The activities of the HAv represent the sectors of aeronautics, aviation and aerospace.HAv is the world's third-largest aviation cluster after Seattle (Boeing) and Toulouse (Airbus headquarters).Three main entities -Airbus, Lufthansa Technik and Hamburg Airportare accompanied by 300 plus small and mediumsized enterprises (SMEs), suppliers and service providers with more than 40,000 highly qualified employees.HAv members include both large manufacturing companies and small consulting firms, all of which active in the metropolitan region and are integrated into HAv's supply chain.This business ecosystem is accompanied by educational and academic institutions -associations  External relations of an Industry 4.0 cluster: the case study of the Hamburg aviation cluster 2167

REGIONAL STUDIES
The field study relies on primary and secondary sources and encompasses participant observation and in-depth semi-structured interviews, the scenario of which was literature-based but flexible to allow new topics to emerge (Kasabov, 2015).Those interviewed were representatives of the HAv office (the team of managers involved in coordination of cluster activities), Hamburg City (the Ministry of Economy, research institutions, and managers or chief executive officers (CEOs) of various companies, mainly small and medium-sized, often start-ups active in consulting, training, manufacturing and design), as well as scholars (researchers from HSU).The sample of firms was pretty heterogeneous with producers of aircraft components, propulsion systems, aeroplane engines or integrated mobility solutions, as well as providers of fabrication services.This stratification guarantees the diversity of research participants.The 26 interviewers conducted in mid-February and March-June 2019: in person, at the premises of companies or institutions, and by phone, as teleconferences were anonymized and classified as cluster representatives (CR), cluster experts (CE), cluster companies (CC), cluster officials (CO) and cluster scholars (CS) (respectively, seven CR, four CE, 12 CC, two CS and one CO) (for more details, see Appendix C in the supplemental data online).
Based on the typology proposed by Welch et al. (2011), two basic dimensions were incorporated: a causal explanation that ensures the internal validity of reasoning, and a context that provides a full description of the case, which is particularly vital as I4.0 encompassing digitalization, connectivity and automation is bound to contextspecific variations (Culot et al., 2020).In addition to seeking to meet the criteria of contextual explanation, this research fits into relational research design that factors in the setting and interdisciplinarity (Bathelt & Glückler, 2018).In light of the shortcomings of the qualitative approach, one should refer to the ongoing scholarly debate where the case studies are gaining recognition or even being moved into the mainstream (Knight et al., 2021;Welch et al., 2011Welch et al., , 2022)).The intention of this research was to explore and to find out (certainly not verify); and by recognizing some empirical regularities, to come up with a contextualized explanation (certainly not a fully fledged theory) which according to Welch et al. (2022) can be seen as emerging theorizing.Often practised decontextualization implies the risk of reductionism, misinterpretation and losing sight of outliers.Suggested contextualized explanation, which attributes that context has explanatory power, can be achieved by process research, historical research, the extended case method or configurational theorizing.Following Cornelissen et al. (2021), theorizing can be understood as encompassing different practices, each of which produces its own distinct knowledge claims.Different forms of theorizing, as practices, effectively provide different forms of understanding; some are geared towards explanation, others interpretation or emancipation.All three involve researchers using different styles of reasoning.Drawing on such an approach this study is guided by both explanatory interestto reveal the fundamental forces, and to some extent by interpretive interestaiming at reflexively revealing the interdependencies of studied phenomena and social meanings.Inspired by Tsang (2022), this paper could be seen as an identification of empirical regularities aiming at contextualized explanation (as context is seen not as a 'hindrance', but as having explanatory power; Welch et al., 2022).

DISCUSSION: UNEARTHING THE NATURE OF THE INDUSTRY 4.0 CLUSTER'S EXTERNAL RELATIONS
Desk study, combined with participant observation and semi-structured guided interviews (for selected quotations, see Appendix D in the supplemental data online), yielded insight into the peculiarities of cluster external relations in the I4.0 era (Figure 2) which might be confronted with the available literature.
This empirically grounded interpretation and contextualized explanation of external relations undertaken by HAv as an I4.0 cluster enable identifying certain regularities, in particular the scope of I4.0 impact on cluster external relations, the multiscalarity of these processes, the actors affected and the aspect of longevity (i.e., a long-term, future-oriented adaptability).The obtained findings demonstrate the following: . Longevity is a primary attribute of external cluster relations and hubbing (Buxbaum-Conradi, 2018; Dohse et al., 2018).International expansion should be naturally sustained after the initial support expires and adapt to changing circumstances.Besides, it needs to be jointly developed (owned) by as many cluster members as possible (Andersson et al., 2019).Only then can hubbing or other external ties last long and be truly self-sustaining.The need for adaptability, which combines both resilience and flexibility of external relations, is also mirrored in the cluster firms' business behaviours and factored in the long-term strategy of many SMEs, as they prefer a long-time approach to establish and conduct relations with outside partners. .External cluster expansion, including its specific form of hubbing, is obviously affected by the ongoing digital transformation (Strange & Zucchella, 2017).The pervasive nature of I4.0 implies that it is regarded not only as know-how or a set of modern technologies which can be the subject of exchange.It is also a tool facilitating distant cooperation, which allows the partners to overcome the disadvantages of distance.I4.0 is also raising concerns regarding, for instance, common challenges, such as future skills.The interpretative flexibility (Vernay et al., 2018) of I4.0 implies that the topics covered in clusters' international cooperation might be indeed broad (Lund & Vildåsen, 2022). .This cluster expansion (hubbing) happens at many levels (Asheim et al., 2018;Benner, 2021;Hassink & Gong, 2017) and ranges from the very local, even municipal, level, with plants located in different districts of the city, through the regional, national to finally the European and global levels reflected in truly transnational partnerships. .International cluster expansionhubbingalleviates liability issues (Aldrich & Auster, 1986;Baum & Oliver, 1991).In particular, it helps SMEs to overcome the liability of un/(under)connectedness and that of smallness (Moreira et al., 2019).It empowers them and allows them to emancipate vis-à-vis another influential cluster member, hence reducing the asymmetry.
Based on the exact citation collected during the interviews, the codes were derived and proposed.They formed the basis for emerging higher order constructs, that is, the categories that describe the nature of cluster external relations, hubbing, in particular, and which are discussed in detail in the following sections.

LONGEVITY (LONG-TERM OWNED ADAPTABILITY)
The key challenge in terms of the cluster's geographical expansion tuned out to be the longevity of its relations.'In all these international projects usually knowledge is being exchanged, but the problem is how to sustain this often unfortunately superficial cooperation, which tends to die soon after the funding stops' (cluster scholar 2).
In order to sustain these ties, which quickly ends when the funding expires, a bottom-up strategy that reflects members' genuine needs must be developed.It is becoming apparent that to stay competitive, it is necessary to bring new ideas from the outside.It is essential to link to the outside world and use the services of international partners.However, as HAv experts say, it is a prerequisite to strengthen local assets, assure endogenous growth and only then focus on internationalization which may serve as an extra facilitator of this growth.
In line with the concept of internationalization considered as a C2C cooperation, HAv is initiating partnerships with its foreign partners such as Canada, Brazil and Portugal within the framework of the ministerial project (Bundesministerium für Bildung und Forschung -BMBF).The areas of likely collaboration and perspective partners, however, always derives from negotiations with all HAv stakeholders.So far, only cooperation with Canada has proved successfulmembers have jointly defined the areas of their interest and fields for potential collaboration focusing on the early stages of the aircraft life cyclejoint scientific cooperation.The European Aviation Cluster Partnership (EACP) (n.d.), coordinated by HAv, and which is a link to other selected clusters, including many from Germany, is intended to facilitate a joint submission of applications and the implementation of European Union-funded projects.
Cluster experts stress the critical role of internationalization, but they underline the necessity to assure that it is a two-way street and the need to avoid unintentional harmful sharing of knowledge with competitors.Thus, international partnerships supported by HAv aim at securing this win-win situation, implying that only a fraction of received calls to set up partnerships materializes, which reflects that priority is given to quality, not quantity.'We have 10-15 international delegations but are selective, focus more on quality we signed only few MoU but indeed demand is high' (cluster representatives 1, 3, 6 and 7).This approach reflects the issue of alliance saturation when the cooperation capacity reaches the upper limit starting to produce more negative consequences (Goerzen, 2018).

Specificity of nexus I4.0-external relations
Digitization is an opportunity to accelerate international expansion.New ideas and business strategies are emerging that force companies to seek international solutions and establish contacts with the outside world.I4.0-related technologies facilitate obviously remote communication with other clusters, but besides being simply a tool connecting dispersed partners, they are first and foremost regarded in terms of specific knowledge and know-how which is being shared.
Skype and all modern ICT [information and communication technology] tools enable distant communication, but it matters to meet, to have face to face contacts … cooperation encompasses not only R&D [research and development] projects but also works on supply chain and recently also HR [human resources] cooperation.This has been initiated to grasp more knowledge of how to deal International R&D collaboration supported in HAv intends to establish new cooperation platforms.However, the cooperation includes not simply research projects, but also downstream activities such as personnel or HR-related collaboration with the aim to obtain new knowledge concerning the diversified upcoming digital challenges.
HAv strategy of internationalization does not aim at attracting FDI.For example, in the case of Canada, the aim is to enter into new research projects.The innovative scientific collaboration will be the first stage, followed by exploring market opportunities and investment chances.HAv tries to help companies to get to know market chances and provide them with necessary contacts.The international expansion comprises diversified channelspartnerships with Canada, Portugal and Brazil are the most important, whereas cooperation with the UK (reinforced due to the Brexit uncertainties) or Japan has been advancing.Typically, HAv helps and assists companies in the first phase of exploratory cooperation, facilitating the posting of entrepreneurs or offering information on investment opportunities.However, the explicit aim is neither to rise export or OFDI by cluster companies, nor to attract more IFDI.This may also be due to the specificities of the aviation industry, where international expansion is more about connecting to GPNs and building relations within value chains.

SMEs' unconnectedness
The role of a cluster in facilitating internationalization derives from the prevalence of SMEs experiencing often the liability of smallness and un-or (under)connectedness. Participation in international projects assisted by HAv managers leads to certain liberation of small companies from other large players, makes them more resilient and reduces their excessive dependence.'I think for some entities here in Hamburg it is quite interesting development because of the danger of consequences of relying so much on Airbus' (cluster expert 3).
Efforts expediating internationalization are indispensable for SMEs, which lack resources and the ability to undertake their own projects abroad (Yamamura & Lassalle, 2019).Thus, it is vital to provide them with an insight into the foreign market and improve access to distant contractors.'We as first tier provider for OEMs worldwide have own ties but for small entrepreneurs who do not have business contacts in the world, cluster assistance in this respect is critical' (cluster company 11).Although international actions can positively affect cluster members' visibility and publicity it remains unknown how they contribute to genuine internationalization, and it is difficult to see the exact, measurable results or the effects of such actions.Too close ties with one large company, in this case, Airbus, can constitute an over-dependency risk.Hubbing and the expansion of international relations may hence contribute to diversification that enhances SMEs' resilience in case of a negative shock arising from this asymmetry.It opens up new opportunities for SMEs and facilitates access to other clients that they can connect to or to other aviation clusters.
Almost all members recognize HAv assistance in internationalization; however, despite such opportunities available, smaller companies often do not have sufficient resources to exploit fully these occasions as they are understaffed and cannot commit more assets to get engaged.It is up to each firm to decide if and how international expansion fits into its own strategy.SMEs agree that external relationships established thanks to HAv support serve as the first step and even if not immediately, these initiatives form the basis for future relations as from such early meetings, and discussions valuable business cooperation can emerge.However, there are also sceptical voices which stress that most of SMEs operate very locally, and it is challenging for them to access other markets, to supply to foreign customers.HAv members admit also that the whole aerospace community as such is becoming more diversified and international than it used to be twenty years ago with new players entering the market.

Region plugged into multiscalar networks
Hubbing of HAv is mirrored in Airbus specific and close linkages to France, the UK or Spainthe relations which are, though, best epitomized by the coopetitionthat is, simultaneous cooperation and rivalry.Thus, hubbing occurs thanks to Airbus n-tier partners (Bräuninger et al., 2010) and the entire region is connected to the GVC/GPN and inserted into broader exchange networks (Bathelt & Li, 2014).Firms' local cluster embeddedness coexists with explicit global network engagement (Chaminade et al., 2020).In fact, as there are only a few aviation hubs worldwide, they need to be somehow connected and digital transformation is seen as a new tool for more international openness of such places.'Aviation is dominated by few large firms like Boeing or Airbus, so there is less a question of where to relocate but rather what region fits best into my current project, my needs?' (cluster representatives 1, 3, 6 and 7).
HAv geographical expansion happens at various levels.Locally, HAv stretching is mirrored in the metropolitan area of Hamburg, with Airbus facilities sprawling in Finkenwerder, Stade and Buxtehude, while HAv's formalized strategy focuses on internationalization with central target regions such as Brazil, Canada or Japan.HAv has hence developed jointly with its members a suitable strategy aiming at connecting members with future cluster partners worldwide.The internationalization strategy focuses on intercontinental target regions and consists of a conceptual and implementation phase, also financed by the BMBF.To strengthen the position of the industry, HAv has initiated and is coordinating the EACP which started in 2009, with 24 aviation organizations in 11 countries and is comprises more than 43 members from around 18 countries.The EACP works in a decentralized manner, based on working groups, temporary consortia and ad hoc partnerships.It aims to improve global competitiveness in Europe through intensive cooperation between clusters.The EACP serves as a platform for mutual exchange and a launchpad for international projects.This cooperation is part of the inter-cluster alliances (Goerzen, 2018) which facilitate the flow of knowledge, reducing the problem of missing links felt particularly by SMEs' suffering the liability of unconnectedness (Baum & Oliver, 1991).Cluster management with inter-cluster alliances, therefore, provides the necessary bridge function for resource-poor SMEs.
Apart from the EACP, the HAv cooperates with clusters from other industries in Hamburg and with other leading aviation clusters throughout Germany.For instance, the Supply Chain Excellence Initiative (SCEI) framework was initiated with the aim to develop stable and agile supply chains at all value-added steps, whereas the new national project REALISE aims to develop the runway-independent automatic launch and landing system, which represents a revolutionary concept for the future of aviation.These initiatives might be perceived as examples of inter-clustering (Cusin & Loubaresse, 2018;Franco & Esteves, 2018).
Summing up, the results underline that sustainable development here depicted as longevity is a common thread and an overarching aspect binding all types of external activities carried out by local stakeholders (Bellandi et al., 2020).External relations are essential for SMEs, suffering the unconnectedness liability, because they empower and help them to emancipate (Baum & Oliver, 1991).International expansion of the cluster in the times of I4.0 implies new forms of cooperation, while the I4.0 is the subject of such a cooperation, as well as a tool facilitating it.It is related to the integration of the cluster into global chains and networks (De Marchi & Di Maria, 2019), while external outgoing relationships are implemented on many levels (Hassink & Gong, 2017).
The obtained results may contribute to the existing literature on clusters by developing a more differentiated approach to clusters' external relations at the times of I4.0.Hubbing in HAv implies different channels of opening the cluster to the outside world: hosting non-Hamburg members, maintaining Airbus networks, the EACP activities or strategic ties with selected foreign clusters.The findings of HAv show that establishing external relations is more about searching for an extra competitive edge than exploiting the existing, already revealed, one.It is about concluding partnerships, developing alliances and hence jointly tackling challenges faced in the I4.0 times (Pagano et al., 2020).Hubbing comes in different forms and runs at different scales ranging from purely city sprawl to global partnerships.It needs to derive from members' preferences to ascertain that efforts made, and mechanisms implemented, will be naturally sustainable.For SMEs, experiencing often insufficient external links, international assistance provided by cluster organizations could serve literally as the 'door opener'.Though, it is often the case that such opportunities cannot be properly harnessed by local firms due to resource constraints.Worth stressing is a mature and sober approach to hubbing demonstrated by cluster members.While acknowledging the need for foreign expansion they also remain cautious as to the risk of 'defocusing'.
This research touches upon the issue of cluster openness and its local roots by discussing the multiscalarity of hubbing (Bathelt & Taylor, 2002).The covered issues address the recent scholarly trend to regard clusters as dynamic and complex systems, whose evolution depends on the capacity to reach outside to reap and diffuse new knowledge (Belussi & Hervás-Oliver, 2017;Pagano et al., 2020).Evolution, however, must not be seen as the 'path-dependent-deterministic-going-through-subsequent-stages', but rather as a complex system adaptation and ability to transform and transmute (Trippl et al., 2016, p. 276).The findings obtained may have useful implications, as they could provide insight and practical guidelines for other similar initiatives (Hervás-Oliver et al., 2020) and enrich the studies devoted to the effects of cluster policies (Cantner et al., 2019;Rothgang et al., 2019) or policies for clusters (Abbasiharofteh, 2020).

CONCLUSIONS: LIMITATIONS AND CONTRIBUTION
This paper relies on the narrative literature review, combined with the empirical qualitative research method and a case study to find out the nature of clusters' external relations in the I4.0 era.The obtained results reflect recent calls for a transformative industrial strategy, which needs to be developed both nationally and regionally and should use different scales (Bailey & De Propris, 2020).Diverse mechanisms such as networks, markets, hierarchies or spillovers should be harnessed at multiple geographical scales to help cluster entities access foreign knowledge and trigger combinatorial knowledge creation processes, enabling local transformations (Bellandi et al., 2020).
This paper has classic shortcomings that derive from the qualitative nature of the research.Researchers recognize more often the risks of decontextualization, hence an attempt was to offer a contextualized explanation of the I4.0 cluster's external relations.Some information might seem as inadequately provided, which only illuminates the problem of qualitative and case study research of how to reconcile the rich description of context with concise generalized results (Knight et al., 2021;Tsang, 2022).Certainly, one of the challenges is deciding 'what to show and what to tell' (Pratt, 2009).Simply showing the data can make the paper overly descriptive, whereas just telling about the data can suggest the results are unsubstantiated.The single-case study method makes it difficult in generalizing and results in possible bias in interpreting interview transcriptions (Richardson et al., 2012), despite internal validity, the transferability to other cases, that is, the external validity is difficult to assess.This study covers only one industry, in a specific metropolitan region, hence it is indeed idiosyncratic.Nevertheless, the purpose of this qualitative study might be described as the identification of regularities aiming at contextualized explanation (as context is seen not as a 'hindrance', but as having explanatory power; Tsang, 2022;Welch et al., 2022).Future research testing these relationships in larger samples seems inevitable.Drawing on Feldman and Storper (2018), it could be argued that adopting a relational research approach that integrates the context, path dependency and contingency factors yields more authentic and empirically grounded findings, which may serve as a departure point for prospective research.
Despite the limitations, the presented study may enrich the existing literature on cluster internationalization by proposing a more nuanced view of this process (Zubiaurre et al., 2020).The peculiarities of hubbing processes as investigated here represent a general concept of the evolution of clusters (Fornahl et al., 2015) and may contribute to the existing literature in this respect.This research adopts the meso-level analysis, which, as it seems, is somewhat neglected in the dominant literature focusing on either microeconomic aspects of the digital revolution (improved firms' productivity, lower manufacturing costs) or macroeconomic dimension (cushioning problems of an ageing population, enabling urban production).
This research sits at the intersection of international business and economic geography (Hervás-Oliver et al., 2020) or evolutionary economic geography and GPNs which have run on parallel tracks with limited cross-fertilization (Rodríguez-Pose, 2021) and can complement the recent research combining strategic management and the regional studies literature stressing that regions have to develop dynamic capabilities to successfully adapt to big disruption such as I4.0 (Bianchi et al., 2019).It also answers the call for more diversified, open and qualitative research in economic geography (Hassink et al., 2019).
As the research landscape on geographical aspects of digital transformation remains rather modest (Castelo-Branco et al., 2019) the obtained results can contribute to the literature by advancing a more nuanced approach to cluster external relations in the I4.0 time.Despite the growing interest in the exploration of business digital transformation, there seems to be a research gap with regard to the spatial dimension of this process (Castelo-Branco et al., 2019;De Marchi et al., 2017).The research on I4.0, in relation to clusters, is only emerging (Hervás-Oliver, 2019;Hervás-Oliver et al., 2019).This study can be perceived as an encounter between localization and globalization and digitalization, which also speak to recent calls for a new industrial policy in I4.0 (Bianchi et al., 2019) which is place-based and integrated working at the intersection of technological change and territories (Bellandi et al., 2020).This paper might enrich the existing literature on clusters' expansion, advance our knowledge regarding the I4.0 in the meso-perspective and contribute to a better understanding of hubbing, that is, the scale expansion which depicts the evolutionary and dynamic approach to cluster study (Abbasiharofteh 2020).It may thus advance our understanding of the cluster and I4.0 co-evolution.

Figure 1 .
Figure 1.Simple map of cluster external relations: proposed typology.

Figure 2 .
Figure 2. Key aspects of cluster external relations: Hamburg aviation cluster (HAv) case.