Developing and sustaining new regional industrial paths: investigating the role of ‘outsiders’ and factors shaping long-term trajectories

ABSTRACT

This article casts light on the development of new regional industrial paths. We explore factors explaining why regional industries with similar early path development trajectories may exhibit diverging outcomes in the longer run and pay particular attention to the role of ‘outsiders’ in the initiation and further development of regional industrial paths. Drawing on a comparative case study of IT industries in Linköping and Karlskrona, two medium-sized Swedish city regions, we find that the inflow of outsiders was an important driver of early path development processes. However, we find that the interplay between regional preconditions and arriving outsiders, and between outsiders and existing actors, substantially shaped the long-term sustainability of the industrial paths in our study. In particular, the role of agency in fostering positive self-reinforcing mechanisms and structure–agency dynamics are highlighted as key factors for understanding how new industrial development paths are unfolding in the longer term.

KEYWORDS:

• Regional industrial change
• agency
• path development
• sustaining industrial paths
• outsiders

1. Introduction

The question of how and why new industries emerge in some places and not in others is a long-standing topic in economic geography (see e.g. Storper and Walker 1989). Recent contributions, in particular within the strand of ‘evolutionary economic geography’ (EEG) literature (see e.g. Boschma and Frenken 2006; Martin and Sunley 2006; Boschma and Martin 2010; Martin 2010), have enhanced our understanding of how pre-existing regional industrial structure, institutional configuration and local knowledge interaction patterns condition and stimulate innovation and the emergence of new industrial growth trajectories.

Path dependence and new industrial path development, defined as the emergence and renewal of industries in cities and regions (see Grillitsch, Asheim, and Trippl 2018; Dawley 2014; Simmie 2012), have become key concepts in economic geography. Studies have given attention to which regional structures are most conducive to new industrial path development (Isaksen and Trippl 2016b) and through which mechanisms they emerge (Martin and Sunley 2006). It has also been argued that existing preconditions are important determinants of new path development, and that the exploration and analysis of these factors should be at the top of the research agenda (Neffke, Henning, and Boschma 2011). Furthermore, current conceptual and empirical analyses have started to give attention to the role of agency in new regional industrial path development and how actors may influence the regional environment to be more enabling for the emergence of new paths (Dawley 2014; Dawley et al. 2015; Simmie 2012; Steen 2016; Miörner and Trippl 2017).

However, we argue that insufficient attention has been given to factors explaining why regional industries with similar early path development trajectories may exhibit diverging outcomes in the longer run. In addition, whilst contributions have provided important insights to the role of key actors at different spatial scales, the role of ‘outsiders’, understood as actors originating from other regions, remains relatively unexplored (Trippl, Grillitsch, and Isaksen 2018), despite studies suggesting that such actors may be an important source of new industrial pathways (Neffke et al. 2018). Hence, our contribution lies in exploring what role is played by outsiders in the emergence and development of new industrial paths, and investigating what matters for sustaining the dynamics of an industrial path after its initial emergence.

More specifically, we compare the rise and further development of the IT industry in two Swedish medium-sized cities, namely Linköping and Karlskrona. These two manufacturing cities, located in the southern part of Sweden, both became key centres of the Swedish IT industry in the 1990s but have since then experienced different path development trajectories. Drawing on 38 in-depth interviews with key stakeholders from industry, academia and the policy and support system, we explore how key actors of change, in particular outsiders, have set in motion processes that culminated in the emergence of the IT industry in Linköping and Karlskrona, and why these two regional industrial paths exhibit different long-term development despite similar early path development trajectories.

2. Conceptual framework

Questions of how and why new industries emerge in different locations have been long-standing points of enquiry in economic geography. In an influential contribution, Storper and Walker (1989) argued that firms in the early stage of industry emergence experience locational freedom and ‘windows of locational opportunity’ (see also Boschma 1997). In this view, new industries need generic capabilities and basic institutions, which are too widely available in space to explain why new industries emerge in particular locations. However, studies of path dependence and path development in economic geography have put more emphasis on the relationships between existing structures and the emergence of new industries, and on the casual mechanisms through which new paths emerge.

Early work on path dependence in economic geography (Grabher 1993; Martin 1999; Hassink 2005) focused on explaining continuity, the persistence of economic patterns and (negative) lock-in. In terms of providing explanations for how trajectories were ‘de-locked’, studies emphasised the role of historical accidents and exogenous shocks (Martin 2010; Bathelt and Boggs 2003; Hassink 2005, 2010; Meyer-Stamer 1998). More recent contributions have started to highlight questions of how, why and where new paths emerge (see e.g. Martin and Sunley 2006; Martin 2010; Simmie 2012; Dawley et al. 2015).

New regional industrial path development is defined here as the rise of new regional industrial activities of economic relevance (see e.g. Grillitsch, Asheim, and Trippl 2018). The keywords here are ‘new’, referring to activities not previously engaged in by regional actors, ‘regional industrial’, referring to industrial activities of economic relevance in the region, ‘path’, referring to the rise of a critical mass of interrelated actors and activities stimulating self-reinforcing effects and (positive) lock-in, and ‘development’, referring to the importance of taking into account underlying agentic processes rather than perceiving new paths as solely the outcomes of certain structural conditions.

Research on new path development has given particular attention to decisive factors and conditions in the regional environment in which new path development comes about (Neffke, Henning, and Boschma 2011; Boschma and Martin 2010; Martin and Sunley 2006; Martin 2010; Fornahl et al. 2012; Isaksen and Trippl 2016b). However, even though Martin and Sunley (2006) highlighted ‘transplantation from elsewhere’ as a possible source of path development already in their seminal paper, there have been few studies investigating exogenous sources of new paths (for exceptions, see e.g. Trippl, Grillitsch, and Isaksen 2017; Dawley et al. 2015).

Whilst being grounded in evolutionary thinking, the ‘path development debate’ is taking place at the intersection of several strands of literature in economic geography. Studies have combined EEG and institutional approaches in economic geography to understand path development in different contexts and multi-scalar influences. Most notably, innovation system approaches, in particular regional innovation systems (RIS) and technological innovation systems (TIS), have been used to understand regional structural change and the formation of new industries (see e.g. Isaksen and Trippl 2016b; Martin and Simmie 2008; Binz, Truffer, and Coenen 2016). Furthermore, studies have drawn upon the global production networks (GPN) approach (Coe et al. 2004; Yeung and Coe 2015), considering relationships between actors and institutions at different spatial scales (see MacKinnon 2012; Pike et al. 2016; MacKinnon et al. 2018). In this paper, insights from different strands of literature are incorporated both for understanding regional preconditions and the role of exogenous actors, as will be discussed below.

New regional industrial paths are the outcome of actions taken by agents embedded in regional environments, which can be both enabling and constraining (Martin 2010). It is possible to situate studies of new path development between structural accounts and approaches emphasising agency (see Isaksen and Jakobsen 2017). The former is reflected in literature arguing for the embeddedness of actors in social structures (Uzzi 1997), which generally downplays the role of strategic agency, particularly the role of non-firm actors (Binz, Truffer, and Coenen 2016). Studies emphasising agency, on the other hand, treat actors as being able to exercise intentionality in their activities, more or less independently of structure (Garud and Karnøe 2001). It is however increasingly argued that such actors are still influenced by the structures in which they are located, for example by mobilising resources to create more favourable structural conditions. Studies of new regional industrial path development are paying increasing attention to such structure–agency dynamics (Isaksen and Jakobsen 2017; Grillitsch and Sotarauta 2018).

According to Martin (2010), new industrial paths might take one of two stylised trajectories. The first is to take the form suggested by the traditional canonical model of path dependence, in which the path is self-reinforcing without any substantial evolution or endogenous change, ultimately leading to lock-in, continuity and stasis. This type is likely to co-evolve with a constraining regional environment. The second is to take a more open and dynamic form, with self-reinforcing effects beneficial to continuous evolution and endogenous change. To determine which trajectory an emerging path will take in the longer run, it is necessary to examine the interplay between regional preconditions and the role of agency in fostering processes leading to positive self-reinforcing development mechanisms (Martin and Sunley 2006).

2.1. Regional preconditions

Instead of a ‘historical accident’, as in the canonical model of path dependence, recent work highlights pre-existing regional structures (Martin 2010) or initial conditions made up by the historical evolution of path dependent trajectories (Simmie 2012). Such place-specific preconditions inherited through previous local economic development constitute environments in which local agents are embedded (Martin 2010). A range of studies have investigated different regional preconditions and their influence in new industrial path development, such as the industrial composition of the region, and organisational and institutional support systems (see e.g. Isaksen and Trippl 2016b; Boschma and Frenken 2011a; Dawley 2014).

Knowledge, competences and other resources built up in other industries can be used in new industrial activities, and regions endowed with a diverse industrial structure, preferably with a high degree of (related) variety, are argued to offer enabling environments for new paths to emerge (Boschma and Frenken 2011a). Furthermore, institutions shape regional economic activities (Martin 2000; Scott 2010) and differences in institutional contexts are important determinants of innovation activities across space. Studies have demonstrated how formal institutions such as government policies and regulations can play an important role in path development (Dawley 2014; Dawley et al. 2015), but have also pointed to the conducive influence of informal institutions such as a favourable ‘climate’ for entrepreneurship and common visions among different stakeholders (Fritsch and Wyrwich 2014; Zukauskaite, Trippl, and Plechero 2017). Finally, the role of knowledge organisations, such as universities, R&D units, vocational training facilities and support agencies such as incubators, cluster initiatives, networking platforms and other intermediaries, have been highlighted as important in processes of new path development (see e.g. Isaksen and Trippl 2016b).

Consequently, the local environment can also limit the scope of new industrial activities, for example due to lack of combinable or convertible resources, too much competition in existing activities (Brezis and Krugman 1997; Martin 2010; Boschma and Martin 2010) or underdeveloped institutional and supporting frameworks.

However, whilst different regional contexts are seen as being more or less enabling for the emergence of new paths, it is important to remember that regional preconditions need to be adapted throughout the path development process. Supporting structures are typically not present at the early emergence of a new industry but ‘gradually evolve as the local industry develops and as processes of positive lock-in consolidate the industrial path and reinforce its momentum’ (Martin and Sunley 2010, 83).

2.2. Agency in new industrial path development

Until recently, the role of agency has been largely neglected by scholars in evolutionary economic geography (see e.g. MacKinnon et al. 2009; Dawley 2014; Neffke et al. 2018). Studies have started to emphasise the role of the purposeful agent who ‘mindfully deviates’ from existing paths, a perspective founded in a socio-economic ‘hybrid theory’ (Simmie 2012), drawing to a large extent on work by Garud and Karnøe (2001). In this strand of literature, existing pathways are seen as the ‘platforms for potential departures’ (Simmie 2012, 756) and it is argued that new paths do not always require the breaking of old ones. In other words, economic agents such as inventors, innovators and entrepreneurs (Simmie 2012) may use localised resources to engage in processes with the purpose of creating new paths.

Martin (2010) highlights local agents engaging in experimentation, and competing against other agents, implying that private actors draw on potentially favourable initial conditions but also are exposed to selection mechanisms through competition. An emerging strand of literature deals with the role of non-firm actors (Tanner 2014; Binz, Truffer, and Coenen 2016; Simmie, Sternberg, and Carpenter 2014), for example the role of policy actors at different spatial scales (Dawley 2014). In addition, it has been demonstrated how actors are involved in ‘adjusting’ the initial conditions and constructing capacities for path development in parallel with the emergence of early activities within a new industrial path (Dawley 2014; Tanner 2014), for example by adapting the regional support system for innovation (Miörner and Trippl 2017).

2.2.1. Outsiders

Despite findings supporting the idea that ‘outsiders’, actors moving into the region, play an important role in the formation of new industries (Saxenian 2007), the literature has focused extensively on endogenous potentials for new path development and has largely neglected exogenous inputs. However, outsiders may actually be a source of more radical change than existing regional actors (Neffke et al. 2018) and recent studies have increasingly highlighted the role of outsiders as triggers for the emergence of new regional industrial paths (Hedfeldt and Lundmark 2015; Stockdale 2006; Kalantaridis and Bika 2006). In this paper, outsiders are defined according to previous studies (see Trippl, Grillitsch, and Isaksen 2017) to include entrepreneurs (Brekke 2015; Saxenian 2007) and other individual actors, but also organisations such as firms or R&D units relocating to the region.

Outsiders are often not embedded in existing regional structures, which might equip them with a greater freedom of constraints originating from old incumbent structures. However, the inflow of outsiders triggering new industrial path development does not necessarily mean that long-term sustainable industrial growth paths are created. If actors arrive to exploit existing conditions, such as low production costs or different forms of subsidies, the interplay between the new industry and the regional environment might be limited. On the other hand, if actors engage in interactive processes and exploit learning opportunities in the region, positive spillovers are more likely to occur (Cantwell and Iammarino 2005). In other words, their success is dependent on becoming sufficiently embedded in their respective regional context (Binz, Truffer, and Coenen 2016; Vale and Carvalho 2013). There is reason to believe that not only the success of relocating outsiders but also the success of the regional industrial path in which they engage depends upon such embeddedness.

Actors act upon opportunities they perceive in the regional environment, and for outsiders this includes a shift of location to another region, in order to mobilise and exploit resources (Ács, Autio, and Szerb 2014). However, varying preconditions not only provide different resources for actors to mobilise but also attract actors with different actor attributes. Various typologies for different types of entrepreneurship traits have been presented in the literature, such as intrinsic versus extrinsic motivations for entrepreneurship, and entrepreneurship driven by necessity or opportunity (Carsrud and Brännback 2011). These have in common that entrepreneurs differ in their individual motivation, and that these differences influence the outcome of their entrepreneurial activities. Similar observations can be made when it comes to organisations; the reasons behind the resettlement of firms may range between access to new markets, locally available knowledge or other resources and the existence of favourable institutional or organisational support configurations. However, some organisations, such as national authorities or universities, might become established in a region as a result of policy decisions at other spatial scales and not so much as a result of certain regional conditions.

Thus, as some actors are characterised by a high degree of mobility (Stam et al. 2009), their regional presence can be assumed to be dependent on the existence of favourable preconditions for their activities. Over time, structure–agency dynamics might lead to a co-evolution of motivation and structural conditions, leading actors to become embedded and anchored in the region. However, it might also be the case that the regional environment, for various reasons, no longer provides the resources for which the outsider came to exploit, leading to departures in seek of new opportunities. This situation can be expected to be more likely to occur if the outsiders never became sufficiently embedded in the region.

Not all outsiders, however, arrive in the region in seek of economic opportunities arising from regional preconditions; they come for a wide range of other reasons. For individuals, such reasons range between family conditions, job opportunities, quality of life (Florida 2005) and so forth. It is for example widely acknowledged that most entrepreneurial activities take place in the ‘home region’ of the entrepreneur (Stam et al. 2009). In other words, even if they are outsiders in the sense that they are relatively new to the region, their locational decision is not always based on the availability of exploitable resources in relation to the new path. Outsiders relocating to these premises come with new knowledge and experiences, but when they become engaged in new path development activities they are already established in the region. This has potential implications for the long-term sustainability of new paths, as it could be expected that paths consisting of activities by such actors should be more sustainable than if actors are more opportunity-driven in the sense of relocating to exploit certain regional preconditions.

2.3. Fostering self-reinforcing development mechanisms

Actors interact with existing regional structures in different ways. For example, actors might exhibit different degrees of ‘fit’ when it comes to informal institutions in the region, for example in terms of entrepreneurial culture (Beugelsdijk 2007). This interplay depends not only on actor attributes but also on the absorptive capacity of the existing regional firms (Cohen and Levinthal 1990; Giuliani and Bell 2005). Furthermore, existing preconditions may influence the type of outsiders that are attracted to the region. Actors with different motivations, intentions and other attributes will most likely differ in the way which they, intentionally or unintentionally, adapt supportive institutional structures (Isaksen and Trippl 2016b) or ‘cultivate’ self-reinforcing effects (Sydow, Lerch, and Staber 2010).

Agency, and in particular the arrival of outsiders experiencing some ‘structural freedom’, can widen the scope of possible outcomes within a certain structural context (Pike et al. 2016) and trigger change processes, spurring the emergence of new industrial paths in seemingly unfavourable environments. Agency can also contribute by ‘strategically manipulating’ self-reinforcing mechanisms (Karnøe and Garud 2012). As actors are continuously adapting and changing the regional environment in which they are embedded, the interplay between regional preconditions and processes of agency may or may not foster self-reinforcing development mechanisms, which are crucial for the long-term sustainability of new paths.

First, the development of path-specific formal and informal institutions which embed the new path into a local trajectory is argued to be a potential source of positive lock-in (Martin and Sunley 2006). Such institutional configurations include, for example, political awareness, common visions, and expectations, which are manifested through, for example, regional strategies (Cooke 2001; Dawley 2014). In such processes, key actors can play important roles as institutional entrepreneurs or place leaders (Sotarauta and Pulkkinen 2011), setting regional visions and mobilising support from various stakeholders. It has also been shown that self-reinforcing expectations might provide incentives stimulating new entrants (Steen 2016). However, these processes can play out differently depending on the dynamics between existing regional actors and outsiders. If existing regional actors are ‘in the lead’, it can be expected that outsiders are more likely to adapt and become embedded in the region, whilst it is harder for a group of outsiders to influence existing regional actors.

Second, self-reinforcing mechanisms such as local external economies of industrial specialisation (Martin and Sunley 2006, 2010) are closely related to the development of an organisational support structure fostering collective learning, collaborative networks and economic interrelatedness. Miörner and Trippl (2017) have demonstrated how key actors, both public and private, play important roles in shaping supportive elements and adapting the organisational support system of a region, by creating new support organisations but also by influencing the activities of existing ones to become aligned with the emerging path, or by using existing structures in new and possibly unforeseen ways. In such processes, the power and knowledge held by key actors proved important, pointing to the potential uneven relationship between outsiders equipped with less knowledge about existing conditions than existing regional actors.

In summary, the analytical framework outlined above highlights the continuous interplay between preconditions and agency throughout the path development process. Agents, and in particular outsiders, may engage in adjusting the regional preconditions to support new activities, but our framework highlights the role also played by agency in fostering self-reinforcing development mechanisms and in the gradual evolution of supportive conditions, crucial to the long-term success of new paths. Whilst outsiders are indeed argued to experience a greater freedom of constraints arising from regional preconditions, it is important to stress that preconditions still matter: as the environment shaping early path development activities, in influencing what type of outsiders are attracted to the region, and in providing interpretative leeway for actors fostering self-reinforcing mechanisms.

3. Method

In this study, we use a qualitative case study methodology (Yin 2014; Flyvbjerg 2006) to gain in-depth insights into the role of outsiders in new path development and how new paths can be sustained after their initial emergence. Our research aims for an in-depth analysis of a current social phenomenon, which makes case studies particularly useful (Yin 2014). In addition, in order to reduce the risk of selection biases (Eisenhardt 1989), we do a comparative analysis between two cases.

3.1. Case selection

For illustrating the role of outsiders in new path development and how paths are sustained after their initial emergence, the cases of Linköping and Karlskrona have been selected for several reasons. First, in the 1990s, both regions housed two of the best-known IT clusters in Sweden. Second, both regions had a strong manufacturing industry in the 1940s, but the regional preconditions for new path development in the 1980s were quite different. Linköping has undergone a remarkable transformation over the past century from a small town of rural character into one of the largest cities in Sweden. Karlskrona, on the other hand, is known as an industrial city with a long history in naval and military-related industries. In other words, these cases represent so called ‘paradigmatic cases’, which can help us to answer questions of ‘how’ and ‘why’ by identifying differences and similarities between the two.

Table 1 gives an overview of the two cases comprising the sample, and a comparison with the Swedish average.

Table 1. Overview of the two cases comprising the sample.

Location	Population	Population development	Economic base in 2010	Traditional economic base in 1960s	University education level
Linköping	1950: 83,641 1990: 122,268 2010: 146,416	1950–1990: +46% 1990–2010: +20%	IT cluster, aviation cluster	SAAB AB (computer, aviation), NAF	1985: 13.8% 1997: 24.0% 2011: 34.0%
Karlskrona	1950: 63,668 1990: 59,054 2010: 64,032	1950–1990: ‑9% 1990–2010: +8%	Manufacturing industries; IT cluster	Navy, dockyard, electronics	1985: 9.2% 1997: 16.5% 2011: 25.0%
Sweden	1950: 7,046,920 1990: 8,590,630 2010: 9,415,570	1950–1990: +22% 1990–2010: +10%	Advanced industry (Ericsson, Volvo, Sandvik, Alfa Laval)	Manufacturing industry, forest industry, steel and mining	1985: 10.1% 1997: 16.5% 2011: 24.0%

Source: based on data from Statistics Sweden (SCB)

In our study, we compare and analyse the development of IT industries in the two regions. Figure 1 shows the composition of the IT industry in Linköping and Karlskrona. For Karlskrona, years 1997–1998 are left out owing to changes in the classification of IT activities, leading to incomparability with previous and following years. The figure illustrates how the IT industry in Linköping is largely dominated by IT services, while the IT industry in Karlskrona is mixed, with a slight trend since 2003 towards IT manufacturing. It should be noted that in Karlskrona IT manufacturing has been increasing since 1991 in total numbers but the rise and subsequent fall in the total number of IT employees is due to changes in the number of employees in IT services. In Linköping, the number of employees in IT manufacturing has been slowly decreasing over time: from 527 employees in 1991 to 221 employees in 2010.

Figure 1. Share of total IT sector employment by IT services and IT manufacturing. Numbers above the columns indicate the total number of IT employees per year. Years 1997–1998 for Karlskrona are left out owing to inconsistencies in the data (see above).

Source: based on data from Statistics Sweden (SCB). SNI2002 sector codes; Manufacturing: 3001, 3002, 3130, 3210, 3320, 3330; Services: 7210, 7221, 7222, 7230, 7240, 7250, 7260.

3.2. Data collection

Our empirical data was collected over a five-year period between 2010 and 2015. It consists of interviews, city chronicles and official documents from the regional and national governments. The city chronicles were in particular helpful to gain insights on the historic development of Linköping and Karlskrona and enabled to go further back in time than interviews alone would have allowed (see Almroth and Kolsgård 1981; Lokalhistoria 1999; Wirén 1986). Altogether, we conducted 38 in-depth interviews. Each interview lasted between one and a half and three hours. Interview partners included both entrepreneurs who started IT-related companies (in Linköping during the past 35 years and in Karlskrona during the past 25 years) and representatives of firms in the IT industry. It should be noted that the actors had changing roles throughout the decades: university staff and students first became entrepreneurs and later local investment managers or board members of other technological start-ups. In order to get a broader picture of the development process of the IT industry, we also interviewed representatives from the municipal administration, universities and established local companies. Table 2 gives a short summary of the interviewees and also indicates the share of outsiders, as previously defined in the theoretical framework. We deliberately did not base our selection of interview partners on whether or not they were outsiders but selected key actors in initiating and sustaining the IT industry in the respective region. During the interviews it became apparent that most of the key actors were outsiders which moved to the regions to study or work.

Table 2. Summary of interviewees.

	Occupation	Region	Outsiders
IT industry (21)	Entrepreneurs	Linköping: 9 Karlskrona: 12	Linköping: 9 Karlskrona: 10
Municipality (5)	Civil servants	Linköping: 3 Karlskrona: 2	Linköping: 2 Karlskrona: 0
University (7)	Researchers/academic staff	Linköping: 4 Karlskrona: 3	Linköping: 3 Karlskrona: 3
Established firms (5)	Management level	Linköping: 3 Karlskrona: 2	Linköping: 3 Karlskrona: 2

3.3. Data analysis

Each interview was transcribed, and we used thematic analysis to extract information linked to the emergence of the local IT industry (Guest, MacQueen, and Namey 2012). The aim of the data analysis was to understand and explain the phenomenon at hand, namely the emergence and further development of a local IT industry, and to stipulate a set of causal links about how and why this phenomenon occurred. In that sense, we searched for patterns and insights in our data that seemed promising. Going through each transcript, common themes between the two cases emerged, such as the role of outsiders and the role of policy actors. These themes have been used to structure the case analysis. The analysis is presented below and is illustrated with quotes from some of the interviews. These quotes have been translated from Swedish by the authors.

While the interplay among different actors is of great concern in the process of discovering how a focus on agency can extend our understanding of new path development, our aim was to produce a contextualised story. Instead of telling a story about certain individuals and their activities, the analysis is based on a deeper investigation of certain processes or common themes, which proved to be central for the emergence and further development of the local IT industries in Linköping and Karlskrona. We used these local processes to bring together and highlight the activities of different actors, using several concrete examples. This is not to focus on several random ‘historical accidents’ but to show how the activities of many different actors reinforced each other and culminated in the emergence of the local IT industry. Major strengths of a detailed chronological sequence are the ability to trace changes over time, such as the importance of different actors and their activities, but also the relationships among events (Yin 2014).

4. Empirical case study

4.1. Regional preconditions

Drawing on material gathered through an examination of city chronicles (Almroth and Kolsgård 1981; Wirén 1986; Lokalhistoria 1999) and official documents from the regional and national governments (Gov. bill. 1987, SOU (Statens offentliga utredningar) 1989), we were able to construct a chronological sequence of historical accounts providing a background to the regional preconditions in each region. The preconditions in the two regions may seem similar. In the 1930s, both regions hosted manufacturing industries. These were in turn dominated by a few major players. In Linköping, the military aeroplane company SAAB AB was the main actor in the local industrial setting, whilst Karlskrona was dominated by the naval industry and its suppliers. Because of the national importance of their activities, both were considered matters of national security; these firms were characterised by closed organisational structures with a large share of in-house activity. In other words, formal regulations shaped the possibilities for interaction and inter-industrial knowledge exchange among local firms.

However, whilst Karlskrona had been dominated by military activities since the sixteenth century, Linköping’s military industry emerged in modern times. In 1909, the establishment of the Swedish Railroad Company (ASJ) by the Uggla brothers represented the first larger manufacturing company in Linköping. Owing to a change in market demand, the company later diversified into aeroplane manufacturing. Sweden’s largest aeroplane manufacturing company, SAAB, acquired the aeroplane division in the 1930s and at the same time relocated its own aeroplane manufacturing to Linköping. Instead of a process of continuous expansion, Linköping went from being a rural town to, at least in some sense, becoming ‘Sweden’s aviation capital’ practically overnight. Suppliers were attracted to Linköping but, as a military company, SAAB had much of the R&D and production in-house and engaged in little interaction with other companies (Eliasson 2010). As a one-company town, the success of Linköping was now very much linked to the success of SAAB. The demand for military aeroplanes increased drastically after the Second World War and the company was in desperate need for large numbers of qualified employees. The local labour market did not match the needs of SAAB and the required nationwide recruitment resulted in a constantly increasing number of highly educated ‘outsiders’ to the city, attracted by job opportunities provided by existing firms. The rise of aeroplane manufacturing was thus associated with the attraction of highly skilled engineers, indicating the early role played by outsiders already in shaping what would serve as the preconditions for the emergence of an IT industry at a later stage.

As a Swedish naval base, Karlskrona traditionally had a strong dependence upon support from the national governmental, initially owing to the ownership structure of the naval dockyard and later because of national policies supporting the naval base (Engstrand 2003). After the Second World War, heavy manufacturing plants were established in Karlskrona, as a result of incentives provided by the national government. Such relocations could be observed until the 1970s. For these manufacturing plants, such as LM Ericsson (1947), Vibroverken (1960) and Uddcomb (1971), the link to the navy and its infrastructure was a crucial factor for the choice of location. In other words, these firms were attracted by, but also took part in shaping, the regional preconditions in which the IT industry would later emerge. However, these were pure production units, with their headquarters and strategic divisions located elsewhere. In the 1980s, these plants were producing standardised products on a shrinking market. This became obvious when the economy stagnated, and the local industry could be labelled as more or less outdated owing to structural change.

4.2. New industrial path development: the emergence of IT industries in linköping and karlskrona

4.2.1. The establishment of universities

As described above, the first wave of outsiders was attracted to Linköping in nationwide recruitment campaigns by SAAB, starting in the 1950s. At the same time, the municipality expressed its wish for a higher education institution. Together, SAAB’s CEO and a civil servant successfully lobbied to bring a technological university college to Linköping, consequently leading to the establishment of Linköping University (LiU) in 1969.

It took a lot of time and a lot of other resources to look for qualified engineers. So, at some point the CEO said that we have to educated people here.

(Linköping, SAAB employee A)

As a one-company town, SAAB’s technological profile was strongly reflected in the newly established university’s educational programmes. The rapid progression in the development of aeroplane technology, for example the introduction of navigational computers, spurred a process of diversification into IT and electronics. In 1950s, SAAB diversified into mainframe computers for private industry and this division spun out as Datasaab in 1978. In parallel, LiU received Sweden’s first IT professorship. SAAB donated one of its first computers for educational purposes to LiU in order to enable the education of suitable employees for SAAB. However, the strict secrecy of a military company did not allow for joint development projects or other forms of cooperation.

SAAB was a closed company at that time, but still had a strong part in forming LiU to what it is today. Who knows if we ever would have gotten such a strong technical profile. And this technical profile prepared the foundation of future start-ups.

(Linköping, university employee A)

Hans Meijer, the vice chancellor of LiU and himself an outsider to the region, had the intention to go against the prevailing norms in academia and wanted to create a university with strong links to private industry. At that time, collaboration between academia and private industry was generally frowned upon, but, searching nationwide, he actively recruited academics who were known for their collaboration with private industry. In that sense, Hans Meijer targeted the adaptation of the regional institutional setting and thus contributed to adapting the regional preconditions to become more enabling.

The vice chancellor had a very strong idea how the university should look like and he would only recruit people who fit within his idea of the university. Professors were very young as well. He wanted to have young and open staff.

(Linköping, university employee B)

In other words, it was not the creation of the university itself which proved to have a significant impact on the development of a new industrial path but how the organisation shaped the institutional setting in the region.

In Karlskrona, the local industry did not have the same impact on the establishment of a local university college. The most significant initiative was taken by the mayor of the neighbouring municipality of Ronneby, Roland Andersson, and resulted in the establishment of Soft Centre in 1987, which was a centre for education, R&D and business-related software development. Soft Centre was highlighted in a proposition to the government as contributing to a unique milieu for a university college. Shortly thereafter, Blekinge Institute of Technology (BTH) was officially established in 1989 with campuses in both Karlskrona and Ronneby. Hence, the role of the established local industry in attracting BTH was minor. The university did evolve into an important local player, mainly owing to initiatives taken by the vice chancellor, Per Eriksson, who himself was recruited from a university in another region. He recruited academic staff from all over Sweden, who brought ideas about a new entrepreneurial university with a close relationship with industry and project-based educational programmes. One interviewee stated that:

The vice chancellor was very, very active, and he saw the need and potential of working with telecom in a higher qualified level, in other words in a research-oriented milieu at the university, but at the same time always in a applied manner.

(Karlskrona, entrepreneur B)

The profile of BTH was very much influenced by the software development profile of Soft Centre in Ronneby. In Karlskrona, the local interests were in favour of a technological education in relation to traditional industry. Hence, Campus Karlskrona was located geographically close to heavy industry in the periphery of Karlskrona, but it was not really embedded in the existing industrial structures. In that sense, there was a certain mismatch between, on one side, the local interests in Karlskrona and, on the other, local interests in Ronneby and the government’s proposition, which argued for a software profile favouring new industries.

In both cities, national decisions leading to the creation of new knowledge organisations, in terms of LiU and BTH, were an important factor enabling early path development activities. However, it was the activities of certain key actors, in this case the vice chancellors, and their efforts shaping the regional institutional setting, that led to a more enabling regional environment, rather than the establishment of universities per se.

4.2.2. The inflow of outsiders

In the case of Linköping, three important points can be made. First, the establishment of a university led to a great inflow of new people (i.e. outsiders) to a little town. University staff and students originated from all over Sweden. Second, owing to the vision of the vice chancellor, the university staff were not only geographical outsiders but also academics with a different mindset, shaping informal institutional conditions which were not typical in the Swedish academic setting at the time. Tailored to the needs of the local private industry, the core of the educational programmes were new, applied courses rather than theoretical ones. These new courses and programmes required a new type of academic, one which enjoyed experimentation and entrepreneurial activities. Third, these new programmes would also attract a different kind of student, one which had an open mind and was willing to take on new educational programmes.

Only a few could really claim to be locals by birth. People were moving to Linköping from all over Sweden. People came from different backgrounds but they spoke the common language of engineering.

(Linköping, SAAB employee A)

In summary, Linköping attracted outsiders with specific characteristics, promoting experimentation, risk-taking and openness. These individuals were attracted by an academic environment that was different from the prevailing norms, rather than entrepreneurial opportunities per se. This highlights the role of regional conditions in attracting outsiders with a certain skill- or mindset.

The strong relation between private industry and LiU led to the establishment of university spin-offs from the mid-1970s onwards. The first university spin-offs were created by university staff originating from joint projects with private industry. Many of these projects required practical solutions to real problems and involved little academic research. Hence, the idea to create their own companies originated from opportunities emerging in the immediate proximity to the entrepreneur. This proved to be a role model for others to follow and by the beginning of the 1980s, for example, most staff from the Department of Image Processing were involved in different spin-offs: Imtek AB, Contextvision and IDA Infront AB, to mention just a few. Soon students also created start-ups. Between 1981 and 1984, altogether 35 spin-offs were created, most with a focus on IT. This meant that there was an increasing number of technological entrepreneurs with small IT companies active in the region.

In the case of Karlskrona, the initial inflow of outsiders was related to the transition from the old, analogue NMT standard to the new, digital GSM standard. The Swedish market for mobile operators was heavily regulated, allowing only for two operators to offer their services. However, the shift from NMT to GSM opened up the possibility for a third mobile operator in Sweden. The company that obtained the licence was NordicTel (which later became Europolitan and was then acquired by Vodafone and finally by Telenor). Nordic Tel was formed by a network of businesspeople in Stockholm, but the new GSM licence was conditioned by the national government, requiring the company obtaining the licence to operate from outside Stockholm. It was decided that the new company, including its headquarters, would be located in Karlskrona. As a result of the establishment of NordicTel, the arrival of additional outsiders, both entrepreneurs and firms, followed. These experienced a high degree of structural freedom, being largely disconnected from existing regional industrial structures, and triggered the early stage of path development in the region.

However, whilst experiencing a high degree of freedom and relocating to exploit the valuable knowledge and skill base existing through BTH, these outsiders did not engage much in shaping local institutions or support structures. In parallel, local politicians and public actors, whilst of course welcoming these new activities, had very limited knowledge about the needs of the new industry. For example, the local commissioner at the time expressed in our interviews that:

We initially did not have a clear but rather vague understanding what this company [Nordic Tel] was all about.

(Karlskrona, civil servant A)

Simultaneously, Ericsson in Karlskrona shifted from manufacturing mobile devices to software development. A new division, called Ericsson Software Technology, was formed, but it had difficulties recruiting skilled labour locally. Furthermore, during the 1990s, several IT companies from other Swedish regions relocated to Karlskrona. This was fuelled by the successful branding of Karlskrona as an IT city, not the least because of the local cluster initiative TelecomCity, which highlights the importance of organisational support structures in attracting outsiders in path development processes.

Whilst the traditional industry had little interaction with BTH, the companies from the new industry considered the university an important partner for cooperation and, when the need for skilled labour increased, the university became a natural collaboration partner. The local CEO of Ericsson Software Technology, Jan-Åke Kark, contacted the vice chancellor, Per Eriksson, at BTH and together they developed educational programmes and a strategy for staff exchange between the university and the private sector.

He [Per Eriksson] was very foresighted in recruiting qualified personnel. The university was able to recruit people because he gave them a broad arena, it wasn´t crowded. Those who were skilled were able to create something really good.

(Karlskrona, university employee C)

Both Linköping and Karlskrona thus experienced an inflow of outsiders. These outsiders were not embedded in existing regional structures, which meant that they were less constrained by the existing incumbent structures. In both regions, outsiders also played an important role in shaping the institutional setting to enable path development processes. However, owing to a better fit between preconditions and the new activities in Linköping, one can argue that the emerging IT industry quickly became embedded in existing structures. In Karlskrona, the IT industry started as a distinct, parallel path with few connections to the established industry. In that sense, outsiders have been important triggers for the emergence of a new regional industrial path, albeit to various degrees of ‘fit’. There were also some differences in the outsiders’ intentions. In the case of Karlskrona, outsiders relocated to exploit business opportunities that had arisen through a combination of national policy decisions and local organisational support structures (i.e. Soft Centre and related initiatives), while outsiders in the case of Linköping relocated to study and work there and played a role in the early path development process first after establishing themselves in the region.

4.3. Sustaining new paths: embedding the IT path in the regional environment

4.3.1. The rise of political awareness

In Linköping, the public authorities quickly became aware of the increasing entrepreneurial activities and several efforts were made to promote this dynamic process in an effective way. Their proactive approach led to the establishment of Teknikbyn, an intermediary providing office space for rent to small companies, in 1981. Three years later, this was further developed into the Mjärdevi Science Park, one of the first science parks in Sweden, in order to meet the specific demands of the more technologically oriented entrepreneurs.

Ever since the local economy started to stagnate, public actors in Karlskrona were determined to create work opportunities, but the municipal focus was on the traditional industries. The interviews captured what also could be described as a form of cognitive lock-in hindering industrial renewal.

The municipal decision to locate the university geographically close to the heavy industry in the outskirts of Karlskrona shows how the political mindset was geared towards the revival of the traditional industry.

(Karlskrona, university employee D)

One entrepreneur illustrates that the inflow of people from outside the region and the establishment of new companies then started to create a certain dynamic milieu that did not experience constraints from the existing environment.

Incubators and new possibilities for venture capital developed out of this. It was like a parallel track to the existing traditional industry. Everything was possible, you know!

(Karlskrona, entrepreneur D)

However, whilst outsiders were important in terms of engaging in entrepreneurial activities and new firm formation in the early phase of path development in both Linköping and Karlskrona, at a certain point it becomes crucial to embed these activities in the broader regional setting. In both regions, the municipality and the local politicians were positive towards these new activities, albeit to different degrees and with different intentions. In the case of Linköping, the municipality wanted to provide a favourable environment for new firm formation, while in Karlskrona it was promoting outsiders entering the region with an established company. This highlights the importance of joint visions and political awareness in shaping the possibilities for sustaining new paths, as will be discussed in the next section.

4.3.2. Institutionalisation of the entrepreneurial phenomenon

As the number of entrepreneurial start-ups in Linköping increased, an entrepreneurial support structure started to emerge, both at the university and on the municipal level. In 1994, the Entrepreneurship and New Business Development Programme (ENP) was initiated at LiU to train individuals in starting up technology-based or knowledge-intensive companies. The programme was designed to help aspiring entrepreneurs with practical advice on how to develop a business idea, business plan and so forth.

In the 1990s, most of the public entrepreneurial support structure was created: different foundations, holding companies and other types of organisation were founded in order to provide support for nascent and novice entrepreneurs. Nowadays, the number of such organisations is around a dozen: the Foundation for Small Business Development (SMIL), InnovationskontorEtt, Exportrådet, Innovation Bridge, Teknikbyn, the Center for Innovation and Entrepreneurship (CIE), LEAD Incubator, Novare, ECI, Linktech and University Holding AB, just to mention a few. In an attempt to co-ordinate the efforts of all organisations with public funding being geared towards entrepreneurship, Growlink was initiated by public actors as an umbrella organisation. In other words, it was local public actors, rather than outsiders, than took on the role of co-ordinating visions, strategies and different parts of the organisational support structures. In 2013, Linköping was the only municipality in Sweden which had such an organisation responsible for co-ordinating the innovation system. The idea was that each organisation would specialise in the particular needs of the start-up. As the start-up goes through different development phases, different specialist organisations would assist their progress.

In addition, the emergence of a large number of university spin-offs, started by staff and students, soon created a self-reinforcing entrepreneurial milieu. Entrepreneurs from the 1980s and 1990s are now appointed board members of start-ups, are mentors in product development or work as consultants for different local support organisations.

The established companies are very open towards the new emerging companies, because they come from the same system. They have the feeling that it is payback time. They had help when they started their companies so they want to help others to succeed.

(Linköping, entrepreneur A)

Not only are local entrepreneurs role models; they are now actively shaping the entrepreneurial environment for novice entrepreneurs. This unusually high level of institutionalisation reflected a long tradition of practical entrepreneurship programmes at the university as well as the co-ordinated support organisations, and it created an entrepreneurial environment rather different from 30 years earlier. One of the interviewees stated that the entrepreneurs are ‘loved to death now’, expressing that there are a lot of actors involved in providing institutionalised support for entrepreneurial activities. In other words, the IT industry in Linköping is now benefitting from self-reinforcing development mechanisms originating from a plethora of organisations providing assistance in a wide range of challenges faced by new and established firms. Furthermore, the institutional framework in the region has been shaped to provide both formal and informal supporting institutions, not least an entrepreneurial climate favouring experimentation and risk-taking among both new firms and incumbents.

In Karlskrona, Jan-Åke Kark, the CEO of Ericsson Software Technology, and Per Eriksson, vice chancellor at BTH, approached the local government in the early 1990s to stimulate actions making use of the unique local combination of hosting a hardware telecom manufacturer, a software engineering company and a mobile network operator. The initiative resulted in a strategic network, named TelecomCity, in 1993. The main goal of the establishment of TelecomCity was to co-ordinate the interests and efforts of the university, IT industry and municipality. The network was an important support structure for the newly emerging IT industry, and it supported the mobility of staff; accompanying spouses were assisted in their search for jobs and joint social activities such as business idea contests were initiated. In other words, whilst Linköping developed endogenous capacities for developing and sustaining the newly emerged path, Karlskrona continued to focus on the attraction of outsiders, becoming dependent on nurturing the attractiveness of the regional environment to sustain the path development process.

One might now argue that the role of outsiders in Linköping became less important than in Karlskrona. But the institutionalisation of the entrepreneurial phenomenon was possible owing to a constant flow of outsiders and the development of endogenous capacities through the integration of these outsiders. LiU still attracts a large number of workers and students from outside the region who stay in the region after graduation. Karlskrona, as a smaller region without endogenous capacities for developing and sustaining the newly emerging path, was more dependent on a high inflow of outsiders for compensating the lack of self-reinforcing development mechanisms developed within the region.

4.3.3. Sustaining entrepreneurial dynamics?

By the mid-1990s, the IT industries in both Linköping and Karlskrona had developed into two of the largest clusters in Sweden. However, the impact of the IT crisis and the ‘burst of the bubble’ did however differ substantially between the two regions.

In Linköping, the burst of the IT bubble resulted in a decline in the number of IT employees and number of IT companies, but the local IT industry quickly proved to be dynamic and resilient enough to recover quickly. Outsiders in the form of established companies with their headquarters in Stockholm, e.g. Motorola, Nokia, Microsoft and Ericsson, had previously started to open units at the Science Park in order to take part in the local activities. When the IT crisis peaked, Nokia closed down its unit. However, most of its employees did not want to move to Finland but started their own companies, e.g. Centerio and 27M. Similar things happened when Ericsson downsized in the region. In other words, outsiders that were once attracted as workers from all over Sweden stayed in Linköping and started their own companies as a response to the crisis. In that sense, the strategic decisions of larger companies to downsize meant an injection of new companies into the local IT industry.

In Karlskrona, the burst of the IT bubble had immediate negative consequences on the local industrial dynamics and economy. BTH experienced a reduced demand for software education and the inflow of IT students decreased. Europolitan/Vodafone was acquired by Telenor and their headquarters were moved to Stockholm. In addition, Ericsson Software Technology moved its strategic units, such as the R&D unit, back to Stockholm. In other words, the two major players withdrew from TelecomCity, and the municipality took over the administration of the network, which had lost much of its dynamic character.

This withdrawal was very obvious for Karlskrona’s IT cluster. It was like letting the air out of a balloon.

(Karlskrona entrepreneur B)

However, not only did the major private players leave Karlskrona; the charismatic vice chancellor of BTH also decided to move on to other commissions and he was replaced by others, who could be described as less visionary.

The university witnessed a loss of pioneering spirit and commitment, and faced a growing bureaucracy.

(Karlskrona university employee C)

The development of the IT cluster in Karlskrona was driven by a handful of key actors, mainly outsiders, and as these outsiders left the region no one took their place. The emergence of the IT industry, driven by two large companies, did not lead to self-reinforcing development mechanisms to the same extent as in Linköping and never became sufficiently embedded in the existing local economy.

We didn’t manage to create the kind of milieu or spirit needed to make the IT cluster develop; maybe Karlskrona was too small.

(Karlskrona entrepreneur C)

It is important to state that, whilst the IT industry in Karlskrona lost much of its dynamic character, the number of employees is still substantial in terms of its share of the local economy. However, rather than being a centre of R&D and technological development, the local IT industry now employs people doing back office functions, IT technicians and other, less technologically sophisticated, functions.

One of the main differences between Linköping and Karlskrona was the type of outsiders that entered the respective regions. In Karlskrona, they were mainly established organisations which wanted to exploit business opportunities, whilst in Linköping outsiders were mainly individuals who entered the region to study or work. The key actors also had different degrees of mobility. Once the firms and entrepreneurs in Karlskrona had exploited business opportunities, they moved on to other places, while the outsiders in the case of Linköping stayed and continued their lives in the region. We can also conclude that, owing to their different intentions, the outsiders in Karlskrona were less embedded in the regional structures and therefore less inclined to stay.

In order to summarise the empirical discussion, Figure 2 gives an overview of the most central events in both cities.

Figure 2. Chronological sequence: (a) Linköping. (b) Karlskrona.

5. Discussion and conclusion

Through a comparative case study of IT industries in two Swedish regions, this paper has highlighted the role of outsiders in the emergence and development of new industrial paths, and has provided insights to what factors matter for sustaining the dynamics of an industrial path after its initial emergence. Our empirical findings are summarised in Table 3 below.

Table 3. Summary of the empirical findings (own summary).

	Linköping	Karlskrona
Preconditions – Industry structures	Manufacturing industry dominated by large companies Aeroplane manufacturing	Manufacturing industry dominated by large companies Dockyard, electronics
Path initiation – Importance of outsiders	Entrepreneurial vice chancellor Attraction of university staff, students and employees Power of many: agency distributed between many actors	Entrepreneurial vice chancellor Attraction of established entrepreneurs and firms Power of few: agency concentrated in a few actors
Sustaining path – Importance of outsiders and policy actors	IT industry fits into technology-oriented economic structure Embedded actors Cooperation between policy actors and entrepreneurs	Few links to the dominating traditional industries Relocation of actors Policy actors working individually

Our analysis strongly supports the view that outsiders can play crucial roles in the early phase of path emergence, both in terms of being key actors in setting the scene for path development processes and in terms of ‘providing fuel’ to the path development process at an early stage. In both regions, new universities were led by vice chancellors who went against the prevailing norms in academia and had clear agendas of collaboration with private industry, and thus shaped the institutional setting in the region. Furthermore, the arrival of outsiders proved to be an important driving force of path development throughout the early stage. However, the processes differed between the two regions. In Linköping, the emergence of the IT industry was initiated mainly through start-up activities taking place locally but to a substantial extent being performed by outsiders. In Karlskrona, on the other hand, the IT industry emerged through the relocation of established entrepreneurs and large companies.

A possible explanation for these different patterns can be found when analysing the regional preconditions in the two regions, in terms of existing industrial structures, institutions and organisational support structures. Both regions were dominated by large manufacturing companies, but whilst Karlskrona mainly hosted manufacturing units of firms with headquarters in other regions, Linköping was the home of R&D units with a strong engineering tradition. This influenced the potential for path development in terms of providing combinable assets and a strong innovation support structure, which resonates well with existing knowledge about new path development (see e.g. Martin 2010; Neffke, Henning, and Boschma 2011).

However, our study demonstrates that regional preconditions do not only provide the point of departure for endogenous path development activities, but also influence the regions’ attractiveness to different types of type of outsiders. We observed that, in the case of Karlskrona, established firms and entrepreneurs were attracted to the region based on the existence of exploitable resources, in combination with incentives provided by the national government. In Linköping, however, individuals relocated to the region for a wide range of reasons, including a vibrant labour market and opportunities in academia.

Our investigation of the role of outsiders in path development, as presented in this paper, clearly illustrates the need to pay closer attention to structure-agency dynamics in path development research. Even though preconditions may work in favour of attracting outsiders to the region, these may fail to become embedded in the broader regional setting. This is illustrated in the case of Karlskrona, as when entrepreneurial opportunities had been exploited, outsiders relocated to exploit new opportunities elsewhere. In Linköping, on the other hand, outsiders developed connections to existing industries as well as with public innovation organisations.

The link between regional preconditions and agentic processes is also apparent when analysing factors influencing how new paths are sustained in the longer run, i.e. how self-reinforcing development mechanisms are fostered throughout the path development process. We have shown that the interplay between existing preconditions and intentional agency can shape path-specific informal institutions, for example through the development of joint visions and strategies. Also, local policy efforts can play an important role in the formation of an organisational support structure specific to the needs of the IT industry, fostering collective learning and collaboration. In other words, the role of established local actors and outsiders can be complementary, with outsiders driving the mechanisms of path development at an early stage and local policy actors strategically developing self-reinforcing development mechanisms through activities reflecting their knowledge about local preconditions. Without the support of established local actors, such as policy actors or established firms, outsiders may have difficulties shaping institutional arrangements and organisational support structures. This confirms the idea that fostering self-reinforcing development mechanisms is more likely to be successful if established local actors are in the lead. Our study points to the ability to foster such development mechanisms as a key explanatory factor for the difference between our two cases.

This study resonates with studies emphasising the role of exogenous inputs for new path development (Dawley 2014; Trippl, Grillitsch, and Isaksen 2017) and highlights the importance of a continued engagement with the interplay between existing regional preconditions and the varying role of agency in new path development, as well as in economic geography broadly. Our findings strongly support the argument that outsiders can play a crucial role in new path development. Nevertheless, whilst outsiders are indeed less constrained by existing structures, the importance of a well-functioning interplay of preconditions and outsiders, that is, between structure and agency, is crucial for the long-term success of new paths. Equally important is the interplay between established regional actors and outsiders, who are more or less embedded in the region. This implies that, in the analysis of how self-reinforcing development mechanisms are stimulated, it is equally important to acknowledge not only the role of certain agents in ‘fostering’ such mechanisms but how structural preconditions and agency work in tandem to produce certain outcomes.

Our analysis is based on a comparison of only two cases, both situated in the Swedish national system, which limits the direct generalizability of our results. Our findings have relevance when investigating other IT industries, or other Swedish regions, but more empirical studies would be needed to assess how patterns might differ between industries or regions in different contexts. However, our results are arguably analytically generalizable, in terms of highlighting decisive factors and conditions influencing new industrial path development processes, which can serve as a useful analytical point of departure for future empirical studies.

Acknowledgments

The authors gratefully acknowledge the contribution by Michaela Trippl to an earlier version of this paper. We would also like to thank Mikhail Martynovich for providing us with quantitative data on the IT industry in Sweden, and Thomas Niedomysl for providing valuable comments. Any errors remain the authors’ responsibility.

Disclosure statement

No potential conflict of interest was reported by the authors.