Revisiting path-as-process: agency in a discontinuity-development model

ABSTRACT

In the literature on regional industrial path development, the path-as-process perspective conceptualizes the emergence, evolution, transformation, and decline of regional industries in the long term. However, critical questions about the role of agency in path development and transformation remain open, partly due to a frequent empirical focus on isolated episodes. This article argues that path development should be seen as a long-term sequence that includes episodes of path development interrupted by occasional episodes of transformation. These transformative episodes are driven by agency within a changing or stable structural context. Such a discontinuity-development model focuses attention on how and why agency patterns change and on which practices agents employ during critical junctures. An empirical vignette on the long-term development of tourism in Eilat, Israel, illustrates how the model can be applied and elucidates methodological challenges for further empirical research.

KEYWORDS:

• Evolutionary economic geography
• path development
• path transformation
• agency
• institutional work
• tourism

JEL:

• Codes: B52
• O31
• O43
• R11

The literature on regional industrial path development (e.g. Baumgartinger-Seiringer, Miörner, and Trippl 2021b; Blažek et al. 2020; Grillitsch, Asheim, and Trippl 2018; Martin and Sunley 2006; Tödtling and Trippl 2013) examines how regional industries emerge, evolve, decline, and transform. Yet, understanding the long-term evolution of regional economies requires a historical perspective (Henning 2019) that covers not just isolated episodes of path development and transformation but their sequence (Grillitsch, Rekers, and Sotarauta 2021b). Martin’s (2010) path-as-process model offers such a long-term perspective but remains vague in analysing changing ‘patterns of agency’ (Sotarauta et al. 2021, 93) between path development episodes. In particular, how different patterns of agency such as change agency (Grillitsch and Sotarauta 2020), reproductive agency (Bækkelund 2021; Grillitsch, Asheim, and Nielsen 2021a), or maintenance agency (Henderson 2020; Jolly, Grillitsch, and Hansen 2020) tip the balance towards certain forms of path development remains a critical question.

Hence, this article aims at shedding light on (i) how paths can be conceptualized in a long-term perspective that connects radical discontinuity with gradual development, (ii) how agency affects the shifts between these episodes. The article argues that for a more precise understanding of long-term regional evolution, we should look at sequences of path development interrupted and linked by occasional path transformation. Such a sequence does not follow a deterministic logic but can unfold in various ways (Baumgartinger-Seiringer, Miörner, and Trippl 2021b; Strambach and Halkier 2013). Specifically, a focus on changing agency patterns during critical junctures of transformation (Bækkelund 2021; Beer, Barnes, and Horne 2021) can significantly improve our understanding of long-term path development (Grillitsch, Rekers, and Sotarauta 2021b).

Empirically, such a discontinuity-development model offers a framework for refining the analysis of long-term path development and agency patterns at critical junctures and thus responds to the methodological challenges for analysing agency Grillitsch, Rekers, and Sotarauta (2021b) identify. To do so, this article takes inspiration from an empirical strategy of ‘temporal bracketing’ (Langley 1999, 703) which has proven useful for an evolutionary perspective (Stephens and Sandberg 2020).

The article starts by reviewing the path development literature and the path-as-process perspective before proposing a simple path discontinuity-development model. In a brief empirical vignette, the article interprets approximately six decades of history of the tourism sector in Eilat, Israel, as a sequence of path development and transformation and thus illustrates features of the model and highlights methodological challenges. The article closes with an outlook on further research.

Regional industrial path development and transformation

The path development literature builds on the fundamental assumption that path dependence is an evolutionary process (Garud and Karnøe 2001; Martin 2010; Martin and Sunley 2006) and produced a typology of path extension, branching, diversification, creation, importation, and upgrading (Grillitsch, Asheim, and Trippl 2018; Isaksen, Tödtling, and Trippl 2018; Isaksen et al. 2019; Martin and Sunley 2006; Tödtling and Trippl 2013) as well as path contraction, downgrading, delocalization, and eventual disappearance (Blažek et al. 2020). Recently, major changes of paths have been conceptualized as path transformation (Baumgartinger-Seiringer, Miörner, and Trippl 2021b; Miörner and Trippl 2019) and the relationships between different paths have been discussed (Frangenheim, Trippl, and Chlebna 2020; Hassink, Isaksen, and Trippl 2019).

However, empirical studies tend to regard path development as isolated episodes or at most as a sequence of few episodes (e.g. Bækkelund 2021; Dawley 2014; Isaksen et al. 2019; MacKinnon et al. 2019; Miörner and Trippl 2019; Sotarauta et al. 2021). There are partial exceptions though. Some studies address longer sequences covering up to roughly five decades (e.g. Binz and Gong 2021; Isaksen 2015) and consider up to three specifically designated episodes of path development (e.g. Jolly, Grillitsch, and Hansen 2020), though some remain vague in defining specific forms of path development (e.g. Binz and Gong 2021). Those studies that refer to only one episode show that even a single episode of path development can span two decades or more (e.g. Miörner and Trippl 2017; Rekers and Stihl 2021) which suggests that a historical perspective going beyond isolated episodes needs to span longer periods (Henning 2019).

A focus on isolated episodes risks neglecting the nexus between path development and transformation and makes it difficult to analyse and compare changing patterns of agency in and between these different episodes. Save a few studies that focus on transformation (e.g. Baumgartinger-Seiringer, Miörner, and Trippl 2021b; Miörner and Trippl 2019), such a focus on isolated episodes privileges ‘“on-path” evolution’ (Martin 2010, 9) over processes of radical change that transform paths from one form into another although Martin (2012) stresses the relevance of both types of evolution, radical discontinuity and gradual development.¹

Empirically, looking at isolated episodes limits the possibilities for inter-temporal replication of findings that enables analysing simultaneous and bidirectional structure-agency dynamics through temporal bracketing (Langley 1999). Hence, there is a need for a more long-term perspective suitable for conceptualizing shifts between forms of path development (Blažek et al. 2020) and the ‘dependence between successive paths’ (Martin and Sunley 2006, 427) which remains underexplored (MacKinnon et al. 2019).

Conceptualizations that view the development of regional industries as a life cycle such as cluster life cycle models (Bergman 2008; Harris 2021; Menzel and Fornahl 2010) allow for the possibility of different episodes to follow each other but face criticism due to a certain determinism and neglect of context and agency (Martin and Sunley 2011; Stephens and Sandberg 2020; Trippl et al. 2015). The idea of an adaptive cluster life cycle that allows for different trajectories proposed by Martin and Sunley (2011) further widens the concept. Though life cycle models and the path development literature are complementary (Benner 2020; Harris 2021), the comparability of forms of path development with life cycle stages is limited because path development does not presume any inherent cyclicality. Hence, in an adaptive conceptualization, path development may be better captured as a sequence with a potentially unlimited number of successive episodes linked by occasional transformation which can include various stages and dynamic processes of agency (Baumgartinger-Seiringer, Miörner, and Trippl 2021b; Boschma et al. 2017; Dawley 2014; MacKinnon et al. 2019; Miörner and Trippl 2017, 2019; Simmie 2012).

Path-as-process and the interplay of agency and structural conditions

Path development can be understood as a long-term sequence with successive episodes (Garud and Karnøe 2001; Martin and Sunley 2006; Simmie 2012). Martin’s (2010) path-as-process model offers a long-term perspective on path development with different stages following each other. Instead of seeing path dependence as ‘canonical’ (p.4), this model defines path development as a process with different outcomes that can lead either to stasis and constrain the conditions for growth or to dynamic adaptation that enables further growth in a new round of path preformation, creation, and development (Martin 2010). In this sense, the model offers a non-deterministic long-term perspective of understanding sequences of path development.

The path-as-process model offers a simplified account of what happens at the end of a path development episode by either generating some form of renewal or decline (Martin 2010). In reality, the succession of different path development episodes is much more complex. As Blažek et al. (2020, 1467) put it, ‘the evolutionary trajectory of a particular industry in a particular region might consist of multiple and swinging shifts.’ Given the possibility of path plasticity (Strambach and Halkier 2013), these shifts do not occur suddenly but are the outcome of gradual processes of path transformation (Baumgartinger-Seiringer, Miörner, and Trippl 2021b; Miörner and Trippl 2019).

Path transformation is not a specific form of path development (Miörner and Trippl 2019). Rather, processes of path transformation occur every time a path transforms itself from one form into another. In a long-term perspective, these path transformations will occur occasionally and repeatedly, and what precisely goes on within these episodes is likely to be a contingent (Bathelt and Glückler 2003), agency-driven set of mechanisms (Baumgartinger-Seiringer, Miörner, and Trippl 2021b; Boschma et al. 2017; Hassink, Isaksen, and Trippl 2019; Miörner and Trippl 2019; Steen 2016).

The idea of path development as a process provides a larger frame for a focus on path transformation (Miörner and Trippl 2019). As Garud, Kumaraswamy, and Karnøe (2010) argue, in contrast to the ontology behind canonical path dependence, a path-as-process perspective follows a contextualized ‘“insider’s” ontology’ (p.761), implying that it is less exogenous events or chance that lead to transformation but agents engaging in what Garud and Karnøe (2001) call ‘mindful deviation’.² Hence, path development and transformation are the results of the complex interplay between structure and agency (Archer 1982; Giddens 1984). Agency can be defined as ‘intentional, purposive and meaningful actions, and the intended and unintended consequences of such actions’ (Grillitsch and Sotarauta 2020, 707) and understood within the framework of Emirbayer and Mische’s (1998) intertemporal activities of interpreting the past, imagining the future, and acting upon the conditions of the present (Grillitsch, Asheim, and Nielsen 2021a; Rekers and Stihl 2021; Steen 2016). Structure and agency are linked through practices (Giddens 1984; Sewell 1992; Stephens and Sandberg 2020) such as those conceptualized by Lawrence and Suddaby (2006) under the different forms of institutional work.

Following Garud and Karnøe’s (2001) call for an agency-focused process perspective, Martin and Sunley (2006) criticize the neglect of agency in the concept of path dependence. Still, the critical role of agency in transformation is not in the focus of Martin’s (2010) path-as-process model which lacks an explanation for what precisely leads to the bifurcation towards either renewal or decline and notably for agency-driven processes of transformation (see also Dawley 2014). Hence, a long-term perspective that sees path development and transformation as subsequent and repeated episodes of evolution has to address the role of agency in relation to the socio-economic structure (Sewell 1992) in which agency takes place (Garud and Karnøe 2001).

Some contributions address the role of agency in the path-as-process perspective (e.g. Dawley 2014; Garud, Kumaraswamy, and Karnøe 2010; Simmie 2012). Bækkelund (2021) argues that agency changes during path development and notably during critical junctures and conceptualizes the role of different forms of agency in Martin’s (2010) model by placing specific forms of agency at a given stage. Still, it remains largely open how and why forms of agency evolve, succeed or compete with each other, and how they become dominant in a new mix of agency patterns that generates a new episode of path development.

As part of the methodological challenges to analysing agency they identify, Grillitsch, Rekers, and Sotarauta (2021b) define a ‘main phase’ (309) in path development as an episode between critical junctures and stress that these critical junctures open possibilities for alternative development trajectories, while agency plays a role in shaping which one prevails. However, the challenge remains ‘to bring forth convincing explanations for why one path rather than another emerged’ (Grillitsch, Rekers, and Sotarauta 2021b, 310).

Agency can take various forms that can combine (Bækkelund 2021; Grillitsch, Rekers, and Sotarauta 2021b). Grillitsch and Sotarauta (2020) argue that three interlocking types of agency (innovative entrepreneurship, institutional entrepreneurship, place-based leadership) form a ‘trinity of change agency’. Following Lawrence and Suddaby (2006) and Henderson (2020), Jolly, Grillitsch, and Hansen (2020) add ‘structural maintenance’ agency that refers to ‘resisting novel activities and adapting to change incrementally’ (p.179). While this definition emphasizes the negative consequences of non-transformative agency (see also Baumgartinger-Seiringer 2021; Boschma et al. 2017), the related but apparently more positive term of reproductive agency can be understood to emphasize how agents contribute to stabilizing and entrenching development processes (Bækkelund 2021; Grillitsch, Asheim, and Nielsen 2021a). While their outcomes might differ, both maintenance and reproductive agency can be summarized as stability agency.

Lawrence and Suddaby’s (2006) institutional work approach offers an umbrella over constructive and destructive forms of agency (Fuenfschilling and Truffer 2016; Kivimaa and Kern 2016) by defining a number of practices agents employ to interact in different ways with the structures surrounding them. These practices include, inter alia, constructive practices of advocacy, educating, or network creation, destructive ones of undermining assumptions, and stability-oriented ones of enabling or mythologizing (Lawrence and Suddaby 2006). Different constructive, destructive, and stability-oriented practices can exist, coexist, and compete in path development and transformation and create friction and contestation (Baumgartinger-Seiringer 2021; Baumgartinger-Seiringer, Miörner, and Trippl 2021b; Baumgartinger-Seiringer et al. 2021a; Boschma et al. 2017; Frangenheim, Trippl, and Chlebna 2020; Fuenfschilling and Truffer 2016; Kivimaa and Kern 2016; Miörner and Trippl 2019).

Agency in path development and transformation occurs in a context that includes structural conditions such as institutions or human and natural assets (Baumgartinger-Seiringer, Miörner, and Trippl 2021b; Baumgartinger-Seiringer et al. 2021a; Chen 2021; Frangenheim, Trippl, and Chlebna 2020; MacKinnon et al. 2019; Trippl et al. 2020). Changes in structural conditions enable new possibilities for agency, thus pre-defining critical junctures (Bækkelund 2021) and opening up ‘opportunity spaces’ (Grillitsch and Sotarauta 2020). Hence, the mix of agency at any given point in path development is the result of different, competing, and contested practices that aim at either constructive or destructive change or stability in a context of changing or stable structural conditions.³

A path discontinuity-development model

Combining the path-as-process perspective with repeated transformation due to changing patterns of agency and resulting in variegated forms of path development yields a model that builds on previous path-as-process conceptualizations (Bækkelund 2021; Baumgartinger-Seiringer, Miörner, and Trippl 2021b; Martin 2010), puts a particular focus on the relationship between discontinuous ‘dramatic’ evolution on the one hand and gradual ‘developmental’ evolution on the other hand (Martin 2012, 184), and attempts to explain this relationship through agency patterns in their structural context with agents’ practices linking structure and agency. Such a model refines the analytical sequence between development episodes and critical junctures (Grillitsch, Rekers, and Sotarauta 2021b) and draws on conceptualizations of path transformation (Baumgartinger-Seiringer, Miörner, and Trippl 2021b; Miörner and Trippl 2019).

Figure 1 illustrates a stylized, hypothetical example of a path that includes episodes of positive and negative path development with occasional transformation.⁴ Hence, Figure 1 demonstrates the logic behind the discontinuity-development model. Apart from those forms of path development that lead to the emergence of a regional path (path creation and importation) or to its demise (path disappearance and delocalization), a discontinuity-development model understands all other forms of path development as temporary episodes that begin and end in transformation. Hence, the model follows the call for identifying development episodes and critical junctures between them to analyse the role of agency in them (Grillitsch, Rekers, and Sotarauta 2021b). Parallel paths are subject to interpath relationships that can be competitive, supportive, or neutral (Frangenheim, Trippl, and Chlebna 2020; see also Benner 2021; Hassink, Isaksen, and Trippl 2019).

Figure 1. A stylized example of the path discontinuity-development model. Source: author’s elaboration.

The path is affected by either changing or stable structural conditions (Baumgartinger-Seiringer, Miörner, and Trippl 2021b; Baumgartinger-Seiringer et al. 2021a) that define the opportunity space for agency (Grillitsch and Sotarauta 2020), notably at critical junctures (Bækkelund 2021). At the same time, changing agency patterns during path transformation contributes to changes in structural conditions (Garud and Karnøe 2001), including by creating, modifying, reorienting, recombining, or destroying assets (Baumgartinger-Seiringer, Miörner, and Trippl 2021b; Chen 2021; MacKinnon et al. 2019; Miörner and Trippl 2019; Trippl et al. 2020). This is why structural conditions and notably assets are ‘both the outcome of previous rounds of regional economic development and the platform for future ones’ (Baumgartinger-Seiringer, Miörner, and Trippl 2021b, 178).

In contrast to life cycle models (e.g. Bergman 2008; Harris 2021; Menzel and Fornahl 2010), a discontinuity-development model does not imply any pre-defined regularity in the sequence of episodes. Instead, the model assumes, first, that episodes of path development will at some point be succeeded by transformation and, second, that path development and transformation will often follow the stages of initiation by pioneers, acceleration by further agency, and consolidation through a critical mass (Baumgartinger-Seiringer, Miörner, and Trippl 2021b; Bergman 2008; Menzel and Fornahl 2010; Simmie 2012). Given the ontology of a path-as-process model that places agency in the centre of evolution (Garud, Kumaraswamy, and Karnøe 2010), the model regards changing agency patterns as the fundamental mechanism behind the succession of path development and transformation episodes, drawing on Bækkelund (2021) who addresses the role of different forms of agency in the phases of the path-as-process model and acknowledges gradual shifts between them. What happens during path transformation is different forms of agency evolving, emerging, and possibly competing in contested processes and subject to power relations (Baumgartinger-Seiringer 2021; Baumgartinger-Seiringer, Miörner, and Trippl 2021b; Baumgartinger-Seiringer et al. 2021a; Boschma et al. 2017; Frangenheim, Trippl, and Chlebna 2020; Miörner 2020; Miörner and Trippl 2017, 2019). Structural conditions can either be stable or change, thus providing opportunity spaces for different agency patterns. Agency and structure are linked through concrete practices. These practices include those enumerated by Lawrence and Suddaby (2006) for institutional work⁵ which are understood here in a more generic sense that goes beyond the strictly institutional level and encompasses change agency forms such as innovative entrepreneurship or place-based leadership in addition to institutional entrepreneurship as well as maintenance and reproductive agency.⁶ For the remainder of this article, these agency patterns are simplified as constructive and destructive agency and stability agency.⁷

Table 1 proposes a heuristic for how practices of institutional work can be used in different agency patterns under either changing or stable structural conditions and how they can contribute to path development or transformation, thus providing a lens for what happens within the episodes in the model.

Table 1. Agency patterns and practices in the path discontinuity-development model.

Agency patterns	Structural conditions
	Changing	Stable
Constructive change agency	Advocacy, defining, educating, identity construction, mimicry, network construction, normative changes, theorizing, and vesting can facilitate transformation	Advocacy, defining, educating, normative changes, theorizing, and vesting can increase the likelihood of transformation
Destructive change agency	Moral disassociation and undermining assumptions can facilitate transformation	Disconnecting sanctions and moral disassociation can reinforce gradual (notably negative) path development
Stability agency	Demonizing, deterring, and policing can counter pressures for transformation and contribute to gradual path development	Embedding, enabling, mythologizing, policing, and valorizing can reinforce gradual path development

Source: author’s elaboration drawing on Baumgartinger-Seiringer et al. (2021a), Binz and Gong (2021), Grillitsch and Sotarauta (2020), Fuenfschilling and Truffer (2016), Jolly, Grillitsch, and Hansen (2020), Kivimaa and Kern (2016), and Lawrence and Suddaby (2006).

As Table 1 illustrates, many practices of constructive change agency can facilitate transformation under changing structural conditions while some of them can increase pressures for transformation under stable structural conditions. Practices of destructive change agency can facilitate transformation (Baumgartinger-Seiringer, Miörner, and Trippl 2021b; Kivimaa and Kern 2016) but can also reinforce forms of gradual path development that will often be negative (Blažek et al. 2020). As for stability agency, some practices can contribute to gradual development despite changing structural conditions while others will be more apt under stable structural conditions.

While the model offers a heuristic for empirical research on the long-term sequence of path development and transformation, several caveats are necessary. First, due to the ‘duality of structure’ (Giddens 1984), whether structural conditions change or remain stable is not exogenous but can, at least in part, be a result of the agency patterns described in Table 1. The interplay of structure and agency is related to the temporality under analysis (Archer 1982; Grillitsch, Asheim, and Nielsen 2021a) which has to be considered in a long-term perspective. Second, the heuristic propositions made in Table 1 do not imply any determinism but have to be seen in a contingent and contextual way (Bathelt and Glückler 2003). Despite its simplifications, the model can help focus empirical attention on the long-term nature of path development with its sequence of gradual change and radical transformation and on the concrete practices driving changes in agency patterns between episodes, as the empirical vignette in the next section illustrates.

Empirical vignette: tourism in Eilat, Israel

The empirical vignette sketches the path tourism in Eilat, Israel’s Red Sea resort, witnessed over more than six decades. The most recent episode analysed is the path transformation and resulting incipient path development after the aviation agreement signed in 2013 that liberalized air traffic between Israel and the European Union (Benner 2022; Reich 2015) until the onset of the COVID-19 pandemic in early 2020.

This vignette is meant as a paradigmatic single-case study (Flyvberg 2006) that draws on empirical material gathered during a preceding research effort (Benner 2022) and recoded and reinterpreted here in view of path development through the lens of the discontinuity-development model. The case study builds on 20 semi-structured interviews with tourism stakeholders and experts (Helfferich 2019) from Israel performed either on the phone or in internet calls between February and November 2020, in some instances supplemented by written clarifications.⁸ Interviewees included representatives of tourism businesses, non-profit tourism serviceproviders, and intermediary or destination management organizations, as well as experts familiar with tourism development in Israel. Interviewee selection was in part based on other interviewees’ recommendations, thus pursuing ‘snowball sampling’ (Goodman 1961), but where necessary to find further interviewees and to find entry points for snowballing also drew on Israeli media coverage (e.g. podcast or television reports on recent developments in tourism in Eilat and the Negev). Interviews loosely followed a short interview guideline that included largely open questions to direct and structure the interview course when needed (Helfferich 2019). As the interview series progressed, the guideline was modified and extended where appropriate, and in its final version (see Annex 1) included, inter alia, questions related to recent tendencies in tourism in the region and challenges to it, the effect of the aviation agreement and the new airport, the development of new tourism models, patterns of cooperation between tourism agents, and interviewees’ outlook on tourism development in Eilat and the Negev. The total interview time was more than 13 h and 45 min. Since all interviewees consented to being recorded, all interviews were taped and transcribed.⁹ Coding was based on a deductive coding structure (Mayring and Fenzl 2019) that drew on insights from the literature on Eilat’s tourism history (see Annex 2). Coding was carried out with qualitative data analysis software (Kuckartz and Rädiker 2019).

Since the interviews allow primarily for analysing more recent developments, to redraw the sector’s long-term history the empirical vignette relies primarily on available literature on the development of Eilat and its tourism sector. While the number of studies about Eilat is limited, those that are available still provide a consistent picture that allows for checking the plausibility and robustness of the insights gained from the interviews, and they complete the picture of the sector’s long-term history throughout the decades at least to a sufficient degree to illustrate the application of the model. Together, information from both the interviews and available literature on the history of Eilat’s tourism sector allow for an approximate analysis of a six-decade sequence. The identification of practices relies on the interview data and focuses on the recent episode since 2013.

The vignette is meant only to illustrate the logic of the model as an inspiration for further, comprehensive empirical research and to elucidate methodological challenges related to the model’s application. By describing separate episodes of path development, the vignette draws inspiration from the idea of temporal bracketing which views episodes as analytical abstractions replicating each other in a single-case research design over time (Langley 1999).

Eilat’s tourism history¹⁰

Eilat is a young city with about 50,000 inhabitants that came into existence only after 1949 on the site of an outpost taken by the Israeli army at the end of the 1948–1949 war after the founding of the State of Israel (Azaryahu 2005; Ergas and Felsenstein 2012; Gradus 2001; Kliot 1997; Mansfeld 2001; Stylidis, Belhassen, and Shani 2015, 2017). In the early 1950s, Eilat became a port city, driven in part by geopolitical developments such as the Sinai war of 1956 and the subsequent lifting of the blockade by Egypt (Azaryahu 2005; Gradus 2001; Kliot 1997; Zerubavel 2019). Later, the city became an oil pipeline terminus and came to rely on mines in the region (Gradus 2001).

1958 to late 1960s: path creation of liminal domestic tourism

Eilat’s location on the Gulf of Aqaba offers a variety of advantages for tourism such as a year-long warm climate, proximity to the Negev desert and its natural and archeological sites, and the coral reefs off the shore (Azaryahu 2005; Cohen-Hattab and Shoval 2004; Gradus 2001; Kliot 1997; Mansfeld 2001; Schmidt and Altshuler 2021; Shaari 1973; Stylidis, Belhassen, and Shani 2015; Zerubavel 2019). Eilat’s era as a domestic tourist destination started in 1958 when the new road through the Negev made it accessible from the densely populated centre of the country (Azaryahu 2005; Zerubavel 2019) but tourism development remained limited into the 1960s (Mansfeld 2001). Soon after its foundation and up until the 1960s, Eilat became engrained in Israeli popular imagination as a site for relaxation with a natural beauty and a liminal character and came to represent ‘a substitute to travel abroad’ (Azaryahu 2005, 121) and a liminal place attractive to international non-conformist hippie youths (Azaryahu 2005; Belhassen 2012; Kaplan 2020; Zerubavel 2019).

1970s: path upgrading towards international tourism

Eilat’s status in Israeli imagination as ‘the beach at the end of the world’ (Azaryahu 2005) was lost to a certain degree after the 1967 war to Sinai (Azaryahu 2005; Hazbun 2008; Noy and Cohen 2005; Shaari 1973; Zerubavel 2019). Notably during the 1970s, Israel’s tourism policy redirected its attention from the Mediterranean coastal resorts towards making Eilat ‘a major winter-sun destination for European tourists’ (Hazbun 2008, 94) and promoting international investment there, as well as towards tourism in the Dead Sea area and the Sea of Galilee (Cohen-Hattab and Shoval 2004; Givton 1973; Hazbun 2008; Mansfeld 2001; Shaari 1973). Eilat’s growth into a sun, sand, and sea destination for European tourists was underpinned by the setup of charter flights that started in 1975 (Azaryahu 2005; Hazbun 2008; Mansfeld 2001; Zerubavel 2019).

The policy focus on promoting Eilat as a sun, sand, and sea destination for European tourists was driven by objectives of diversifying the country’s export base and generating foreign exchange revenues, seizing Eilat’s climatic advantages as one of few winter destinations close to European markets, and softening the local economic impact of the closure of regional mines (Achituv 1973; Blizovsky 1973; Givton 1973; Gradus 2001; Mansfeld 2001; Shaari 1973). The government promoted tourism development in the city through infrastructure investments, allocation of state-owned land to national and international investors, and financial incentives (Blizovsky 1973; Federmann 1973; Krakover 2004; Mansfeld 2001; Shaari 1973).

1980s: path importation of large-scale mass tourism

On the domestic market, Eilat’s transformation into a mass tourism destination largely replaced the 1960s’ to 1970s’ beach camping of nonconformist youth which reflected the changing preferences and rising purchasing power of Israeli tourists (Azaryahu 2005; Belhassen 2012; Cohen-Hattab and Shoval 2004), although the city kept a certain romantic image up to the 1980s (Kaplan 2020). Mansfeld (2001) describes how this policy led to an expansion of supply, notably through the emergence of the national Isrotel chain backed by a British investor with experience in the Spanish hotel market which since the early 1980s built a series of large-scale hotels on the marina that became the core of Eilat’s tourism zone (Isrotel n.d.).

After 1982 when Israel’s pull-out from Sinai was completed, Eilat again turned into the country’s liminal ‘frontier town’ (Azaryahu 2005, 119), although neighbouring Taba was disputed for several more years and handed over to Egypt after international arbitration in 1989 (Hazbun 2008; Kemp and Ben-Eliezer 2000). In 1985, the Israeli government granted Eilat a special status as a free zone that brought with it exemption from value-added tax (Gradus 2001) and customs which further reinforced Eilat’s liminal image (Azaryahu 2005). Legislation passed in 1988 provided for financial incentives for young people discharged from military service when working in hotels for a limited time which led to an influx of workers to Eilat (Belhassen 2012; Belhassen and Shani 2012; Mansfeld 2001; Stylidis, Belhassen, and Shani 2015).

1990s and 2000s: path downgrading of international tourism and extension of domestic tourism

During the 1990s, the Fordist model of standardized sun, sand, and sea tourism (Hazbun 2008) in Eilat was further entrenched by large-scale tourism development and investment as the city’s hotel room capacity doubled (Mansfeld 2001; Zerubavel 2019; see also Krakover 2004). Mansfeld (2001, 172) considers the massive expansion in Eilat’s hotel capacity during the 1990s ‘an uncontrolled development process’ that resulted in low unemployment at the time but also in disadvantageous long-term effects such as massive dependence on tourism and particularly on low-skilled and low-paid jobs, social problems, inward and outward migration, environmental damage, a certain degree of antagonism, and a reputation of a lower quality of life (Azaryahu 2005; Belhassen 2012; Belhassen and Shani 2012; Kaplan 2020; Kliot 1997; Mansfeld 2001; Schmidt and Altshuler 2021; Stylidis, Belhassen, and Shani 2015, 2017). The consequences of uncontrolled development included a loss of attractiveness to international tourists due to dissatisfaction with service quality, the limited choice of attractions, and the perception ‘that Eilat looks more like a building site than a tourist destination’ (Mansfeld 2001, 176), a problem an interviewee mentioned even about present-day Eilat. Further, environmental degradation caused the exit of German tour operators during the 1990s (Mansfeld 2001).

In the wake of the 1994 peace agreement between Israel and Jordan, ambitious plans for regional cross-border tourism cooperation between Eilat and neighbouring Aqaba including a joint airport, theme parks, or a casino emerged but did not materialize (Gradus 2001; Hazbun 2008; Kliot 1997; Mansfeld 2001). Eilat’s 1990s boom in large-scale sun, sand, and sea tourism lasted until the breakdown of international tourism to Israel caused by the second ‘intifada’ in 2000 (Cohen-Hattab and Shoval 2004; Hazbun 2008; Israeli and Reichel 2003; Zerubavel 2019).

While Eilat’s tourism sector generally relied both on domestic and international tourism during the 1980s and 1990s (Ergas and Felsenstein 2012; Krakover 2004; Mansfeld 2001), in the early 2000s Eilat turned into a primarily domestic destination with domestic tourists accounting for about 85 percent of total visitors in 2010 (Ergas and Felsenstein 2012; see also Belhassen and Shani 2012; Stylidis, Belhassen, and Shani 2015, 2017). When international tourism broke down and international hotel chains left Eilat, Israel’s tourism policy turned away from supporting international package and charter tourism to Eilat. Among Israelis, Eilat kept its position as a holiday ‘counter-place’ (Zerubavel 2019, 113) far away from daily life in the densely populated centre of the country (Azaryahu 2005; Belhassen 2012; Belhassen and Shani 2012; Kaplan 2020; Stylidis, Belhassen, and Shani 2015).

After 2013: path diversification and branching towards niche tourism and extension of domestic mass tourism

The period after 2010 is marked by the impact of the aviation agreement between the EU and Israel liberalizing air traffic that was signed in 2013 and implemented until 2018 (European Commission 2021; Reich 2015) and the recovery of international tourism to Israel after the 2014 Gaza war. The liberalization lowered airfares not only for incoming European tourists but also for outgoing Israeli tourists. To promote international tourism to Eilat, the government started subsidizing flights to Eilat’s airports, an incentive that was initially co-funded by the local hotel association, and began widening its focus in international tourism marketing from Tel Aviv and Jerusalem to include Eilat and the Negev. However, while the subsidization policy did attract a number of European airlines, the impact on Eilat’s tourism sector is generally regarded by interviewees as disappointing because a significant part of arriving tourists are believed to move on to other destinations in Israel or neighbouring countries or to resort to low-price apartments or other accommodation options.

Eilat’s small inner-city airport used mainly for domestic flights due to its limited capacity and inadequacy for international aviation and the civilian part of the Ovda airbase used mainly for international flights were replaced by the new Ramon airport located north of Eilat designed to accommodate larger aircraft and to increase the capacity for arrivals by air (Ergas and Felsenstein 2012; Gradus 2001; Stylidis, Belhassen, and Shani 2015, 2017). Both the subsidization of flights and the opening of the Ramon airport can be understood as vesting practices by policymakers.

To increase Eilat’s appeal to international tourists, tourism policymakers over time initiated a wide range of festivals and hosted sports events as an expression of enabling practices. Special-interest niches targeted include bird-watching tourism driven by an international research centre located in the region and Eilat’s favourable location along the routes of migrating birds. Interviewees mentioned embedding Eilat into a wider regional tourism product that includes the Sinai, Aqaba and Petra, and the Negev, although regional tours have been offered before (Gradus 2001; Hazbun 2008; Mansfeld 2001; Shaari 1973; Zerubavel 2019). The recent market entry of a national hostel chain and tour operator to Eilat that offers tours to the wider region seems to have given an impetus to this path. In contrast to the package tour, charter flight model dominant from the 1970s to the 1990s, this path targets young, international independent tourists and backpackers likely to use low-cost carriers and accommodation and to combine a beach vacation with small-scale, community-based tours (see also Noy and Cohen 2005). According to one interviewee, ‘the concept was again to create a hub for independent travellers in the city centre of Eilat, to have a lot of activities, a lot of day tours, multi-day tours, packages that will depart from [the hostel]’ (interview #17, 2020), echoing a practice of theorizing that centres on active travellers who benefit a wider range of tourism businesses in Eilat and the region. Consequently, network construction is a major practice behind the constructive change agency in this model:

The [hostel/tour operator] did the tour in the kibbutzim and they had a tour that started in the Red Canyon and then kibbutz Neot Smadar, Kibbutz Yotvata, and Kibbutz Ketura. (…) It’s good for everybody. (interview #8, 2020)

Another interviewee referred to the help the municipality provided in finding a suitable property for the hostel and highlighted the difference to the institutionalized large-scale hotel model:

It was very difficult to find [a] hostel or such a building which is not in the Northern beach because they didn’t want to be there with all these big hotels. (interview #1, 2020)

The latter aspect hints at a practice of moral disassociation from the established mass tourism model as well as identity construction, both of which are associated with change agency. An interviewee expected the appearance of further hostel brands in Israel which, if it materializes, could eventually lead to the acceleration of this path through a practice of mimicry.

The network construction practice behind this form of constructive agency stands in stark contrast to the deterring of cooperation involved in the integrated large-scale mass tourism model. According to one interviewee, in this model cooperation ‘wouldn’t happen so because when the hotel is sending someone out of the hotel, it’s going to lose the beer money, the lounge money, and the foods money’ (interview #13, 2020).

Still, the largely domestic sun, sand, and sea mass tourism path prevails. While the COVID-19 pandemic was not in the scope of the present study, an interviewee mentioned that limited possibilities for international travel in 2020 strengthened domestic tourism to Eilat (see also Schmidt and Altshuler 2021). Further, interviewees indicated that despite some signs of downgrading, domestic demand for Eilat has not suffered much from the increased international competition in the wake of the aviation agreement because, first, lower airfares enabled Israelis to travel more, second, domestic tourism offers advantages hard to find abroad such as the availability of kosher food, and third, visits to Eilat remain in part linked to trade union vacation offers, corporate incentives, or conferences. The latter two reasons offer cases of embedding of the mass tourism model into wider social practices that account in part for its path extension.

Discussion

While a precise historical account of events contributing to the long-term development of Eilat’s tourism sector is beyond the scope of this article, Figure 2 represents this sequence as a stylized long-term path. Because of difficulties in precisely distinguishing and delimiting patterns of agency, path development forms, and path transformation stages empirically due to the idealized nature of these analytical categories (Baumgartinger-Seiringer, Miörner, and Trippl 2021b; Blažek et al. 2020; Jolly, Grillitsch, and Hansen 2020; Grillitsch and Sotarauta 2020; Miörner and Trippl 2019), Figure 2 necessarily involves stylization and interpretation that depends on judgments about the degree of path development processes and agency patterns and limits the influence of structural conditions to major changes for the sake of simplicity.

Figure 2. Eilat tourism long-term tourism path. Source: author’s elaboration.

Under the changing structural conditions of air-traffic liberalization and airport relocation, constructive change agency is particularly visible through practices of enabling, identity construction, network construction, theorizing, and vesting for path diversification and branching towards niche tourism. As for the path extension of domestic sun, sand, and sea tourism, the deterring of cooperation inherent to the integrated, large-scale mass tourism model as well as the embedding of the model in wider social practices represent practices associated with stability agency. Hence, the propositions about the role of agency patterns and practices included in Table 1 are generally confirmed, although the examples show that borders between agency patterns and between changing or stable structural conditions are fluid and the heuristic categorization in Table 1 primarily serves as an approximation. Further, not all practices are easily distinguishable and some of them overlap.

While it seems too early to evaluate interpath relationships between the paths evolving since 2013, a plausible presumption is that the branching path towards special-interest tourism and the diversifying path towards a small-scale collaborative regional tourism model support each other, but how they relate to the extending path of large-scale domestic tourism over time remains to be seen.

Methodologically, the path development episodes described suggest how temporal bracketing (Langley 1999) might be useful in more comprehensive future long-term path development case studies. In particular, the method’s focus on agency makes it suited for evolutionary empirical research on the role of agency in critical junctures (Langley 1999; Stephens and Sandberg 2020). Due to its methodological limitations, the Eilat case as presented here can be seen primarily as an illustration of the discontinuity-development model and highlights the difficulty in pursuing a long-term analytical perspective towards regional evolution. Reconstructing changing patterns of agency during past episodes in more detail than could be done in this article is fraught with challenges (Grillitsch, Rekers, and Sotarauta 2021b) and requires adding methodological competences in economic history to the toolbox of evolutionary economic geography (Henning 2019). For example, two studies drawn on (Azaryahu 2005; Zerubavel 2019) offer excellent examples for reconstructing historical processes by drawing on long-term media discourses (see also Grillitsch, Rekers, and Sotarauta 2021b), and Henn and Laureys (2010) provide an instructive example of a historical cluster case study by drawing on archival materials. However, qualitatively redrawing the long-term history of regional industrials paths risks rationalizing events ex-post and selectively interpreting them (Grillitsch, Rekers, and Sotarauta 2021b; Henning 2019; Jolly, Grillitsch, and Hansen 2020; Steen 2016). This is another reason why understanding changing patterns of agency and related practices in path development and transformation is important. In particular, if we can sharpen our understanding of which changes in agency patterns and which practices are likely to induce path transformation, our methods in observing path development as it happens (Steen 2016) could significantly improve. Further, combining quantitative and qualitative methods in a long-term perspective appears promising (Grillitsch, Rekers, and Sotarauta 2021b).

Outlook

This article sought to demonstrate the merit of understanding path development and transformation as a long-term sequence of discontinuity and development. By showing that shifts in agency do not occur automatically at any stage but are related to the dynamics of the agent scene and to the precise practices employed under particular structural conditions, the article contributes a more nuanced perspective to pre-existing path-as-process conceptualizations (e.g. Bækkelund 2021; Martin 2010).

However, more in-depth empirical work is necessary to refine the changing patterns of agency and related practices illustrated in Table 1 to distinguish in more detail how and under which conditions path transformation leads to a specific subsequent form of path development and not to others, and how they drive the various stages of transformation. In particular, why changes of agency patterns occur and how their occurrence varies merits further research that could benefit from combining agency approaches with other explanations. While much needs to be done to fully understand the role of agency in regional evolution, the discontinuity-development model provides a framework for embedding specific patterns of agency and practices more closely into conceptualizations of regional industrial path development and can help refine empirical research designs.

Acknowledgements

Empirical research reported in this paper began while the author was employed at Heidelberg University. The author is grateful to Michaela Trippl and Simon Baumgartinger-Seiringer for valuable discussions and suggestions. Research for this paper was presented at the online Modul University research colloquium in March 2021 and at the online International Conference on Cluster Research ‘Rethinking Clusters’ in September 2021. The author is grateful for suggestions and comments given by participants and in particular to Robert Hassink for drawing his attention to the importance of interpath relationships. Last but not least, the author is grateful for constructive comments given by anonymous reviewers, and to the editor, Philip Cooke. Of course, all remaining errors and omissions are the author’s alone, as are the opinions expressed.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This research was funded in part by the Austrian Science Fund (FWF): M 2992-G.

Notes

1 I am grateful to an anonymous reviewer for drawing my attention to the points mentioned in this paragraph.

2 I am grateful to an anonymous reviewer for drawing my attention to these ontological differences.

3 I am grateful to an anonymous reviewer for drawing my attention to the points mentioned in this paragraph.

4 The example of a path given in Figure 1 does not imply any determinism. The path is the result of evolutionary dynamics as they unfold that can be analyzed and redrawn only later.

5 For examples of path development studies that similarly draw on institutional work practices, see Fuenfschilling and Truffer (2016) and Binz and Gong (2021).

6 For the original definitions of each practice of institutional work, see Lawrence and Suddaby (2006), and for a further detailed discussion, see Fuenfschilling and Truffer (2016). For Table 1, Lawrence and Suddaby’s (2006) enumeration of institutional work practices has been slightly adapted.

7 I am grateful to anonymous reviewers for drawing my attention to various points mentioned in this paragraph.

8 This sample is part of a slightly larger sample that covers a wider region and that was analyzed with a different research question in Benner (2022).

9 Where interviewees are quoted directly, the readability has been enhanced through minor language corrections.

10 This sub-section in part draws on Benner (2022) and Benner et al. (2017).