Innovation platforms as a tool for anchoring non-local knowledge: smart specialisation strategies in Guangdong, China

ABSTRACT

Innovation platforms can be useful for promoting the diversification of regional industrial paths. Until recently, the literature had depicted such platforms primarily as a tool for enabling local knowledge recombination – and not for anchoring non-local knowledge. In many regions, however, ‘smart specialisation strategies’ for modernising and transforming industrial structures are difficult to implement without transplantation of non-local knowledge. This paper explores innovation platforms as a tool for anchoring non-local knowledge. We elaborate on recent Chinese experiences by studying diversification-oriented regional innovation policy in Guangdong province. We employ an embedded single-case study design, involving a regionally typical diversification strategy, which we substantiate by tracing platform development with two non-local actors, a university and a technology firm. The paper reveals that non-local actors can play an important role in unlocking regional industrial development potential, especially if platforms enable these actors to relate to local industry by performing desired intermediary functions.

KEYWORDS:

• Smart specialisation strategy
• regional innovation policy
• innovation platform
• relatedness
• industrial diversification
• innovation intermediary

JEL:

• O21
• O25
• O38
• R11

1. Introduction

Capacities for new path development and diversification differ between regions (Tödtling and Trippl 2018). Whereas new development paths might emerge endogenously in regions characterised by a wide variety of knowledge bases and capabilities, lagging regions are less able to link diverse regional paths and hence more reliant on non-local knowledge (Asheim, Boschma, and Cooke 2011; Isaksen 2015; Boschma 2017; Tödtling and Trippl 2018; Trippl, Grillitsch, and Isaksen 2018; Hassink, Isaksen, and Trippl 2019). The inward transplantation of knowledge from elsewhere – by attracting research institutes, firms and skilled labour – can be an important way of importing knowledge for regional path diversification (Martin and Sunley 2006; Isaksen and Trippl 2017). However, profiting from transplanted knowledge in the medium to long term requires anchoring it in the regional context (Crevoisier and Jeannerat 2009). While attracting firms and research institutions as ‘anchor tenants’ (Feldman 2003) may open channels for regions to access external knowledge, the important question is how this knowledge diffuses, and how it is solidified and absorbed within the region (Vale and Carvalho 2013; Binz, Truffer, and Coenen 2016). Without generating relatedness between the transplanted knowledge and regionally existing knowledge bases, the transplantation of firms and research institutions runs the risk of creating ‘cathedrals in the desert’ (Hardy 1998; Cooke 2001).

Policies aimed at attracting non-local knowledge actors can be beneficial for the development of new growth paths (Dawley 2014). The challenge is to involve these actors in strategic, place-based diversification strategies. While the recent regional innovation policy literature takes up the issue of path diversification in connection with the concept of ‘smart specialisation’ (Foray 2014; McCann and Ortega-Argilés 2015), the perspective is primarily inward looking, emphasising local knowledge and prevailing structures (Giustolisi, Benner, and Trippl 2022). Empirical research on smart specialisation strategies in the European Union demonstrates a number of problems with existing strategies. These problems are associated with a region’s capacity to introduce bottom-up policy processes in order to exploit the distributed knowledge of regional growth potentials (Estensoro and Larrea 2016). They specifically concern the identification of priority areas for coordinated investment and the implementation of transformative activities in lagging regions (Kroll 2015; Capello and Kroll 2016; Hassink and Gong 2019). Drawing on non-local actors might alleviate some of these problems, while potentially exacerbating others. How to include non-local actors effectively in these participatory processes at regional level and secure their initiative is one question the literature has yet to tackle.

This paper responds to the call for a further elaboration on outward-looking smart specialisation (Giustolisi, Benner, and Trippl 2022). Our focus is on the centrepiece of smart specialisation – the ‘entrepreneurial discovery’ process that combines bottom-up with top-down elements of priority identification and the implementation of transformative activities (Foray 2014, 2019). We do so by studying diversification-oriented regional innovation policy in Guangdong, China. Regional innovation policy in China resembles the European Union’s smart specialisation strategy, as it aims at modernising traditional sectors by diffusing general-purpose technology (Foray 2014; Barzotto et al. 2020). However, previous research has shown that regional policy, especially in Guangdong, is particularly outward looking (Conlé et al. 2021a).

In our case study, which covers the practice prevalent among (district-level) governments in Guangdong of infusing non-local knowledge to transform and upgrade regional industry, we specifically focus on the use of ‘innovation platforms’ (chuangxin pingtai) in the regional entrepreneurial discovery process. ‘Innovation platform thinking’ (Harmaakorpi, Tura, and Melkas 2011), involving a shift of perspective from developing narrowly defined industrial sectors and clusters towards bridging structural holes and connecting disparate but related knowledge fields and actors (Uyarra and Flanagan 2016), has come to inspire European regional innovation policy debates (Cooke 2007; Asheim, Boschma, and Cooke 2011; Bailey, Pitelis, and Tomlinson 2020). However, the platform concept has primarily been discussed in the context of local knowledge recombination – and not as a tool for anchoring non-local knowledge (Tödtling and Trippl 2018; Giustolisi, Benner, and Trippl 2022). We turn to the latter, based on the assumption that innovation platforms serve this very purpose in Guangdong. We specifically ask the following questions: What role do top-down and bottom-up processes play in Guangdong’s outward-looking smart specialisation approach? How do innovation platforms relate to the entrepreneurial discovery process?

The remaining paper is organised as follows. We first review the literature on smart specialisation and related variety, focusing on entrepreneurial discovery and the role of regional innovation platforms. Based on recent studies, we argue that the smart specialisation literature has not sufficiently accounted for the relevance of exogenous knowledge for new path development. We take up this issue by studying the smart specialisation strategy in Guangdong. To do so, we employ an embedded single-case study design, using the diversification strategy of Nanhai, a typical district within Guangdong’s Pearl River Delta, as the unit of analysis and two of the Nanhai government’s collaborations with non-local actors as subunits. Following a description of our research context, case selection and method, we present our findings, discuss the characteristics of the innovation platform model, and conclude.

2. Literature review

2.1. The rationale for smart specialisation strategies

In view of the increasing mobility and combinatorial nature of knowledge (Crevoisier and Jeannerat 2009), scholars and policymakers have turned their attention from capabilities sustaining and reinforcing specialisations at the regional level towards regional capacities driving diversification and the development of new growth paths (Uyarra and Flanagan 2016; Boschma 2017; but see Baumgartinger-Seiringer, Miörner, and Trippl 2021). This shift is undergirded by the theoretical insight that existing proximities do not only permit cumulative knowledge generation and incremental innovation but may also, in fact, lead to cognitive and technological lock-in and thus an exhaustion of innovation possibilities (Boschma 2005; Martin and Sunley 2006). Leveraging growth potentials requires an interplay between proximity and distance (Parjanen and Hyypiä 2018). Distance between actors and knowledge bases is essential for creating the variety that sustains innovation in the long term. Proximity, on the other hand, is necessary to exploit the innovative potential of distance.

Regional industrial evolution is more than just the reproduction of existing growth paths (Martin 2010). Firms tend to extend, and even renew, their growth paths as they – and start-ups established by their employees (Klepper 2001) – diversify into related activities by exploiting and broadening their existing resources and capabilities (Piscitello 2004). Knowledge spillovers among firms and sectors generate related diversification on a regional scale (Frenken, Van Ort, and Verburg 2007; Boschma and Frenken 2011; Hassink, Isaksen, and Trippl 2019). Related diversification is an endogenous process that regularly occurs without policy intervention (Breschi, Lissoni, and Malerba 2003; Neffke et al. 2018). However, spillovers among paths can fail to occur endogenously, especially when it comes to transformative change, which requires the combination of more distant knowledge (Foray 2014).

Relatedness is an ambiguous concept in this regard (Tanner 2014), because it actually covers both, activities requiring similar knowledge as well as complementary activities based on different knowledge and technology (Breschi, Lissoni, and Malerba 2003; Boschma 2017). Knowledge may only be unrelated as long as it has not been successfully connected (Boschma 2017). Yet regional actors might narrow the scope of their search for new knowledge combinations too much, failing to renew or create regional growth paths. Therefore, Harmaakorpi (2006, 1093) holds that ‘learning and knowledge creation are too important to be left to occur spontaneously’ – a point echoed by Foray (2014).

Tying in with the literature on related variety (Asheim, Boschma, and Cooke 2011), smart specialisation is a policy concept that seeks to facilitate the construction of new regional comparative advantages without presuming the knowledge of an omniscient planner. According to Foray (2019), three basic principles characterise the smart specialisation approach to new path creation. First, smart specialisation builds on the assumption that sector-neutral, horizontal policies aimed at general framework conditions are not sufficient to promote growth in lagging regions (Foray 2018). Therefore, regional innovation policy needs to be vertical, concentrating on selected priorities to achieve sufficient economies of scale. To create synergies, these priorities should relate to the region’s existing capacities and structures. Second, the policy focus is on the transformation of existing structures through the upgrading and diversification of growth paths. The transformation of structures – the ‘transformative activity’ – is brought about by funding and implementing a variety of related projects addressing problems such as R&D, skill development, and technology services (Foray, Eichler, and Keller 2021). Third, the identification of priorities and the implementation of the transformative activity should follow an ‘entrepreneurial discovery’ logic. Choices should be based on public-private cooperation and involve considerable bottom-up initiative.

2.2. Entrepreneurial discovery and innovation platforms

Smart specialisation strategies can forge relatedness and thereby prevent regional path exhaustion and failures in spontaneous new path development (Harmaakorpi 2006; Foray 2014, 2018; Uyarra and Flanagan 2016). Regions are important spaces for the emergence of new combinations of more distant knowledge as geographical proximity can help bridge distances, especially cognitive distances (Boschma 2005; Menzel 2015). At the regional level, the ‘platform approach to regional innovation policy’ (Asheim, Boschma, and Cooke 2011) can serve to unlock the region’s growth potential by connecting disparate but related knowledge fields and actors. Platforms provide an organised context for a shared definition of thematic priority areas and the coordination of regional capability development (Pekkarinen and Harmaakorpi 2006), involving the creation of innovation networks and the pursuit of transformative activities. To utilise platforms as a tool for promoting entrepreneurial discovery in regional smart specialisation strategies requires a consideration of several governance aspects. Two of them are particularly relevant. First, this concerns the determination of the responsible platform actor who coordinates platform processes and activities. Possible actors include government, specialised innovation intermediaries, and actors from various backgrounds who carry out selected intermediary functions (Howells 2006). Second, it involves top-down and bottom-up modes of decision-making with regard to priority identification and the coordination of the transformative activity (Foray 2019).

While ‘innovation platform thinking’ (Harmaakorpi, Tura, and Melkas 2011) has entered regional policy debates, empirical platform models are few. One most frequently cited (e.g. Cooke 2007, 2012; Asheim, Boschma, and Cooke 2011; Uyarra and Flanagan 2016) is the Lahti innovation platform model elaborated by Harmaakorpi (2006) and collaborators. Based on experiences of Scandinavian countries, particularly Finland’s Lahti region, Harmaakorpi (2006) formulates an innovation platform method for discovering priority areas and coordinating implementation. The method involves a broad range of stakeholders and experts from the private and public sectors who assess regional resources and capabilities with the aim of defining, in a bottom-up fashion, priority areas connecting industries, areas of expertise or megatrends (Harmaakorpi, Tura, and Melkas 2011; Uotila, Harmaakorpi, and Hermans 2012). Private or (quasi-) public ‘systemic intermediaries’ (van Lente et al. 2003; Parjanen and Hyypiä 2018; Janssen, Bogers, and Wanzenböck 2020) manage discovery and implementation by acting as brokers and coordinators of interactions between regional innovation actors. Regional (platform) policies accompany the entrepreneurial discovery process, supporting the formation and alignment of complementary resources around cross-sectoral innovation platforms (Asheim, Boschma, and Cooke 2011).

Smart specialisation strategies are supposed to facilitate diversification in various regions, especially in less advanced regions. However, the preconditions for new industrial path development among regions differ along two dimensions (Grillitsch and Asheim 2018). First, regions differ in terms of the breadth of existing industrial paths. It can generally be assumed that the higher the number of regional industries, the higher the potential for new knowledge combinations and path branching (Asheim, Boschma, and Cooke 2011; Boschma and Frenken 2011). Lagging regions thus tend to have relatively limited opportunities for endogenous new path creation. Second, regions differ in terms of their system’s capacity, which concerns the region’s knowledge density as well as organisational and institutional thickness (Isaksen 2001; Tödtling and Trippl 2005). Regional system capacity sets limits to local platform development. Recent evidence suggests that lagging regions face substantial difficulties in implementing bottom-up entrepreneurial discovery processes and integrating private and public stakeholders (Estensoro and Larrea 2016). Lagging regions tend to have a shortage of industrial and innovative capabilities (Capello and Kroll 2016), public sector entrepreneurs (Estensoro and Larrea 2016), innovation intermediaries (Pinto, Fernandez-Esquinas, and Uyarra 2015), and regional governance capacities (Kroll 2015).

While these challenges prevent lagging regions from pursuing an innovation platform model aimed at local knowledge recombination, the literature has yet to come up with an alternative model.

2.3. External knowledge

The relevance of interregional knowledge flows, through the establishment of ‘pipelines’ (Bathelt, Malmberg, and Maskell 2004), has long been recognised in the literature. Despite this, it is fair to say that regional development concepts, from ‘regional innovation systems’ (Cooke 2001) to ‘constructed advantage’ (Asheim, Boschma, and Cooke 2011) and ‘smart specialisation strategies’ (Foray 2014) predominately focus on internal structures and processes (Giustolisi, Benner, and Trippl 2022). Recent studies, in contrast, highlight the importance of non-local knowledge and emphasise that a comprehensive understanding of new regional path development requires the inclusion of external knowledge sources (Hassink, Isaksen, and Trippl 2019). Following Trippl, Grillitsch, and Isaksen (2018), Giustolosi et al. (2022) distinguish between two kinds of knowledge flows: interregional knowledge linkages and the arrival of non-local actors.

In the European Union, strengthening ties among regions has become an important part of the smart specialisation framework. The recent literature underpins this endeavour by demonstrating the relevance of interregional and global linkages for regional industrial diversification. Balland and Boschma (2021), for instance, conclude that regions can increase their diversification potential beyond the potential already existing there from intraregional linkages by additionally accessing complementary capabilities from relevant other regions. However, they also point out that complementary interregional linkages tend to strengthen existing capabilities, whereas they do not compensate for the region’s missing capabilities. Santoalha (2019), in turn, finds that cooperation within and between regions needs to go together to promote diversification. Interregional knowledge linkages thus do not help in overcoming existing structural constraints of lagging regions, although individual firms with strong in-house capabilities may compensate for the lack of local knowledge spillovers by having more interregional cooperation (Grillitsch and Nilsson 2015).

For lagging regions, the attraction of non-local actors can be a particularly useful policy approach. It can be an important means for promoting linkages to external knowledge sources as well as adding momentum to indigenous knowledge dynamics (Conlé et al. 2021b). However, the previous literature shows that embedding non-local actors into regional knowledge networks involves problems. Academic institutions, for instance, can strengthen regional knowledge flows (Isaksen and Trippl 2017) but local universities (or their branch campuses) often fail to adapt to local demand conditions, favouring interregional academic interactions instead (Hewitt-Dundas and Roper 2011). Partly because of a lack of strategic vision, earlier regional development polices, which focused on the importation of technology via foreign direct investment, were frequently unsuccessful, as imported technologies lacked relatedness to regionally existing capacities and structures (McCann and Ortega-Argilés 2019). Inward investors – in Central and Eastern Europe, for instance (Radosevic and Stancova 2018; Potter and Lawton Smith 2019) – tend to be poorly integrated into regional innovation networks, preventing the diffusion and ‘anchoring’ (Crevoisier and Jeannerat 2009) of external knowledge.

Moving away from an inward-looking perspective on smart specialisation therefore raises the question of how to involve actors that have recently located to a region in regional entrepreneurial discovery processes and innovation platforms to create new path development (Giustolisi, Benner, and Trippl 2022).

3. Methodology

3.1. Research context

Instead of focusing on local knowledge recombination, smart specialisation strategies in lagging regions are supposed to have a more outward-looking orientation. This requires a different kind of ‘innovation platform thinking’ than the one that is described in the extant literature. To explore an alternative approach, we look beyond European experiences. Especially in China, the concept ‘innovation platform’ (chuangxin pingtai) pervades innovation policy at several government levels. While the central government has launched innovation platform policies at the national level to support China’s drive for indigenous innovation in strategic techno-industrial sectors (Li, Deng, and Sorensen 2011), innovation platforms more commonly relate to the regional level, usually appearing in the context of innovation infrastructure building (Barbieri, Di Tommaso, and Huang 2010). In its 2010 ‘Guiding Opinions on Strengthening the Construction of Regional industrial Innovation Basic Capacity’, the National Development and Reform Commission (NDRC) mentions several types of innovation platforms existing with national and regional government support. The types mentioned in the national document can also be found in the usually more extensive lists of platforms mentioned in the respective lower-level regulations – the Guangdong provincial government’s comprehensive 2016 regulations, for instance.

Innovation platforms are primarily established as collaborations between the government and non-government actors. To study these collaborations, we adopt the following terminology:

• Platform projects are particular instances of platform construction. Local governments seek to establish platforms together with selected partners.

• Platform actors (or initiators) refer to the government’s partners in platform construction. They include universities, public research institutes, technology firms, and knowledge-intensive business services firms. While platform actors can be from the same region, non-local actors form a particular salient group of platform actors.

• Platform policies concern preferential policies in support of platform development that differ with regard to platform type. They can include, inter alia, favourable tax rates and generous subsidies as well as preferential access to land, research funding, and skilled professionals under China’s multiple talent recruitment programs. Typically, platform development is supported by a wide combination of complementary preferential platform policies.

We assume that local governments in China pursue platform projects with non-local project actors to initiate transformational activities in strategic priority areas, seeking to generate knowledge spillovers and anchoring knowledge within the region.

3.2. Case selection

This paper is part of a larger research project on public-private collaboration in China’s development of regional innovative capabilities. Our empirical focus is on Guangdong, a coastal province in the southeastern part of China, whose Pearl River Delta is one of the three economically most dynamic areas on the mainland. The Pearl River Delta’s nine prefecture-level cities are Guangdong’s industrial heartland. While that area was mostly rural until well into the 1980s, the massive inflow of foreign direct investment, especially by Hong Kong investors, turned it into a global hub for labour-intensive manufacturing and a domestic pole of attraction for unskilled migrant workers. Spearheaded by Shenzhen, the Pearl River Delta has yet experienced a further transition starting in the early 2000s, when innovative activities increased and the regional economy became more strongly orientated towards the domestic market. This shift, which involves an increase of R&D activities, the emergence of new industries, and a rise of domestic migration of skilled knowledge workers, was accompanied by strong policy support from the local-level governments who came to see technological innovation as the main driver of industrial upgrading and transformation (Barbieri, Di Tommaso, and Huang 2010).

In the context of our larger project, we conducted altogether four months of field research during 2018 and 2019, covering five of the most dynamic Pearl River Delta cities in terms of innovative development over the previous decade (Guangzhou, Shenzhen, Dongguan, Foshan, Huizhou). During our research stays, which involved numerous visits of platform projects in all five cities, we came to recognise the policy relevance of innovation platforms for the inward transplantation of knowledge actors from targeted industrial technology sectors. In this paper, we explore the policy approach using a qualitative case-study design. To allow a detailed analysis, we limit ourselves to the in-depth study of knowledge anchoring within a single region. Based on our parallel research in the other field research locations, evidence exists that similar processes and practices of innovation platform development occur in Pearl River Delta cities more generally (Conlé et al. 2021a). We therefore contend that the case, which we present in this paper, is a typical case for the Pearl River Delta.

The selected case concerns knowledge anchoring in Nanhai, one of Foshan’s five county-level districts. Our choice of the district rather than municipal level as the unit of analysis stems from the nature of the country’s policy process. In China, ‘policies are typically designed at the central, provincial and municipal levels, [whereas] counties and districts are responsible for their implementation on the ground’ (Schubert and Alpermann 2019, 212). As we are interested in top-down and bottom-up processes in the implementation of outward-looking smart specialisation strategies, we consider the district to be the administrative level at which the relevant developments take place.

We focus on a district in Foshan city to show that such outward-looking strategies are at work even in a city that is more limitedly orientated towards foreign trade and global value chains than other Pearl River Delta cities (Rubini and Barbieri 2013; Zhang and Kloosterman 2016). According to Deloitte (2018), Foshan has yet made the largest scale of investment in ‘industrial technology reform’ in all of Guangdong in 2017 to facilitate the city’s industrial transformation. Within Foshan, we chose Nanhai, one of Foshan’s industrially most vibrant regions, as our case study site, although we also visited the local sites of several platform actors in two further districts. Nanhai has a permanent population of 2.9 m inhabitants (Deloitte 2018) and a total area of 1074 sq. km, administratively divided into one urban subdistrict and six townships.

3.3. Data collection and method

We adopt an embedded case study design to explain the regional implementation of innovation platform policies (Yin 2018). While our unit of analysis is the district-level specialisation strategy, we pay particular attention to the innovation platform projects that were established by the government with two non-local platform actors – a university and a technology firm. The particular collaboration with these two non-local actors form the subunits of our analysis, which we selected based on our sectoral focus on advanced manufacturing equipment. Our research relies on the data that we collected during our fieldtrips in 2018/19 and by means of (mainly Chinese-language) database and online searches.

At the unit level, we analyse priority identification by studying the government strategies at national, provincial (Guangdong), city (Foshan), and district (Nanhai) level. The relevant strategic documents include the five-year plans (FYPs) and other major-related plans. We concentrate on the 11^th (2006–2010), 12^th (2011–2015), and 13^th (2016–2020) FYP periods.¹ Our inclusion of the higher levels follows the existing literature that characterises China’s economic governance as strictly hierarchical and call for an analysis of regional policymaking under the ‘system of nested authority’ (Heilmann and Melton 2013). The analysis of the plans allow us to assess the top-down nature of strategy development.

To reconstruct Nanhai’s actual industrial change, we particularly consult two yearbooks. In addition to the Foshan Statistical Yearbook, which provides industrial output data for each of Foshan’s districts, the Nanhai Yearbook (NY) proved an invaluable source. NY offers a wealth of annual information on the district’s industrial development, providing data relating to the size of each sector as well as giving information on a sector’s major firms, new entrants, and new infrastructures. We triangulate this information with information from other sources, including articles from journals (e.g. Guangdong Science & Technology) and other media. Even more importantly, we had three formal and several informal meetings with city, district, and high-tech zone government officials that helped us gain an understanding of relevant policy practice.

The latter two sources were also instrumental in developing our two cases at the subunit level. Concerning the collaborations with the two platform actors, we personally visited their local sites to conduct semi-structured interviews with key staff (recording time >2 hours each). The respective meeting-room interviews were complemented by extended guided tours of the local sites that helped us gain a concrete understanding of how the platforms fit in with the actor’s local subsidiaries. Moreover, we gathered and analysed documentary information from the organisations’ brochures, exhibition-hall presentations and websites, and media representations. During data analysis, we compared collaborations within Nanhai – and also drew on evidence from more than a dozen other collaborations with non-local actors in other Pearl River Delta regions – to check the validity of our inferences.

4. Smart specialisation in nanhai

4.1. Determination of priority areas

Nanhai’s specialisation strategies are adaptations of higher-level strategies outlined in the relevant plan documents. Local governments seek to implement the strategic priorities of the centre as good as possible, given regional conditions. In China’s 11^th FYP for the years 2006 to 2010, the central government had developed five different categories of industries, with different tasks determined for each of them. As sketched in Table 1, the plan expected a structural change of industry from traditional sectors involving raw materials towards equipment manufacturing and high-tech industries, as well as an improvement in the country’s position in light industry market segments and an expansion of new energy industries.

Table 1. Industrial development strategy of National 11th Five-Year Plan.

Development task	Relevant industries
Prioritise path development	Vigorously develop the equipment manufacturing industries: 1) Major technical equipment 2) Automobiles 3) Shipbuilding
	Accelerate the development of high-tech industries: 1) Electronic information 2) Bioindustry 3) New materials 4) Aerospace
Upgrade and diversify paths	Optimize the development of the energy industries: 1) Coal 2) Electric power 3) Petroleum and natural gas 4) Renewable energy
Upgrade existing paths	Promote the level of light and textile industries: 1) Textiles and clothes 2) Household appliances 3) Plastics
Restructure and confine paths	Adjust the structure of the raw material industries: 1) Building materials industry 2) Metallurgical industry 3) Chemical industry

Source: China 11^th Five-Year Plan.

With a view to the five categories of industries, the central government addressed both path extension, involving a move into higher value-added activities and segments within existing industries, and path diversification, comprising a move towards more technology-intensive sectors, the development of core technologies, components, and equipment, and the expansion of the service industry, especially knowledge-intensive business services.

The plans at the lower levels related to the national plan. However, to increase the chances of successful plan completion, the lower-level governments devised their strategies against the background of their regional ‘pillar’ industries. In Guangdong, which has developed rapidly into one of the main beneficiaries of the reform and opening-up policies since the early 1980s, electronic information manufacturing and household appliances have become major industries. Accordingly, the province’s own 11^th FYP emphasised tasks and measures to extend the paths of these pillar industries. The provincial plan also covered other industries from the national plan, focusing in particular on the further encouragement and support of those sectors where regional paths had already emerged, such as automobiles in Guangzhou or additional high-tech industries beyond electronic information in Guangzhou and Shenzhen. At the same time, the provincial plan accommodated national-level demands for rationalising traditional industries.

The policy environment allowed a few leading regions with advanced pillar industries to continue along their paths while developing their fledgling high-tech sectors (using similar strategies as the lagging regions). For regions like Foshan’s Nanhai district, path extension was more difficult. In the early 2000s, Nanhai had already developed manufacturing capacity in multiple sectors, but its pillar industries were two raw materials sectors, i.e. building materials (predominantly ceramics) and metal processing (mainly aluminium profiles and hardware), as well as two light industry sectors, i.e. textiles and clothing as well as household appliances (NY 2005). Although Nanhai firms and products had gained substantial domestic market shares in these sectors, the pressure to reduce overcapacity, resource consumption and environmental pollution imposed strict limits on path extension (Nanhai 11^th FYP). In addition to supporting the upgrading and transformation of its pillar industries,² Nanhai therefore focused its efforts on path diversification.

The definition of new priority areas was couched in terms of higher-level plans but actual developments in Nanhai were based on existing capacities and emerging, often unplanned, opportunities. On the one hand, the Nanhai government followed the Foshan municipal government in declaring equipment manufacturing a ‘strategic industry’ (Foshan 11^th FYP). On the other hand, Nanhai’s 11^th FYP heeded the call from higher-level governments to promote high-tech industries.

Nanhai sought to jumpstart its equipment manufacturing industry from an initial base of roughly 150 firms already manufacturing a diverse range of products, including machinery, motorcycles, automobiles and automobile parts (NY 2004). Sectoral growth was to feed from two sources. First, equipment manufacturing was to be more strongly linked to the technological needs of local traditional industries by branching into special-purpose (e.g. ceramics machinery and equipment) and industrial automation equipment. Second, based on its vehicle part firms, Nanhai had identified additional opportunities provided by its spatial proximity to Guangzhou with its expanding automobile sector. Capitalising on these opportunities, Nanhai hoped to locate supplier firms and possibly even auto manufacturers (Nanhai 11^th FYP).

Concerning high-tech sectors, the Nanhai government tried to build on its existing traditional Chinese medicine firms as an entry point for biomedicine and on its LED lighting firms as an entry point for optoelectronics. However, it was clear that local capabilities were not sufficient to develop these priority areas. Nevertheless, as in Nanhai’s emergent automobile sector, the district sought to draw on non-local actors to develop its high-tech sectors. The outward-looking strategy persisted as path branching extended into the subsequent 12^th and 13^th FYP periods, when Nanhai, consistent with the higher-level government emphasis on ‘strategic emerging industries’ (China 12^th FYP), continued to pursue related diversification, inter alia, from automobiles towards new energy vehicles and from basic technical equipment towards robotics and precision engineering.

4.2. Relevance of non-local actors

In the years following the start of the 11^th FYP, Nanhai’s industrial development moved in the direction envisioned by the government plans. Although the district’s traditional industries continued to expand in absolute terms, their share in Nanhai’s industrial output declined considerably. At the same time, the manufacturing equipment sector’s share has increased strongly, surpassing the traditional industry output share in 2019 (Figure 1). The share of high-tech industries also increased notably, albeit on a much smaller scale.

Figure 1. Industrial output shares in Nanhai (2006–2019, in percent). Note: Traditional industries include textiles and clothing, non-metallic minerals, and metals. Equipment manufacturing industries include general and special purpose machinery, transport equipment, electronic equipment, and measuring instruments.

The district’s industrial performance builds on its success in attracting external actors to the region. The automobile sector is a case in point. While Nanhai had already originated a number of auto and auto parts manufacturers, the district’s rise to a new domestic centre of automobile production resulted from the successful localisation of firms in Nanhai during the 2000s. This includes, inter alia, the establishment of subsidiaries of numerous Japanese parts manufacturers in the context of the creation of the Honda, Nissan, and Toyota joint ventures in neighbouring Guangzhou, and, even more importantly, the construction of the South China manufacturing base of the First Automobile Works (FAW)-Volkswagen (VW) joint venture in Nanhai in 2010 (NY 2006, 2007, 2011).

The role of non-local actors is also significant in other sectors. While Nanhai had ventured into electronic information, for instance, the localisation of a Taiwanese producer of TFT-LCD panels created the momentum for Nanhai to become China’s leading LCD TV module production base (NY 2012). Just as for other governments in the Pearl River Delta, the establishment of subsidiaries of such knowledge actors in the region was to play a ‘radiating and leading role’ (Foshan 11^th FYP).

Three aspects of government locational policy are notable in this respect. First, attracting firms is focused strategically on the selected priority areas. However, the government redefines the boundaries of these priority areas flexibly based on the opportunities that it receives with locating particular actors as well as the opportunities arising from its existing industrial base. In this context, the government scans for path branching possibilities, based on expert analyses of technological relatedness, involving the mapping of ‘industry chains’ and ‘technology chains’ (authors’ interviews).

Second, the local government targets not only industrial firms but also academic institutions. Much like other Guangdong regions, Nanhai has invested significant effort into attracting universities and public research institutes (like the Chinese Academy of Sciences) from other parts of the Chinese mainland, from Hong Kong, and from outside of China. The subsidiaries of technology firms and academic institutions are expected to contribute to clustering relevant knowledge in the region and propelling path branching through the local subsidiaries’ applied focus on related strategic sectors, such as new energy vehicles and fuel cells, robotics, and semiconductor technologies. Table 2 demonstrates the extent of the transplantation of non-local knowledge through the attraction of actors from outside the region and how it is related to Nanhai’s strategic industrial development.

Table 2. Path diversification and platform projects in Nanhai.

Prioritized Industry	Branching direction	Attracted firms and research institutions (Examples)
Equipment manufacturing industries
Major technical equipment	Robotics and precision engineering	Firms: Subsidiaries of several domestic robotics firms Academic Institutions: At least three robotics R&D institutes jointly established with domestic technology firms and/or universities including Guangdong University of Technology
Automobiles	New energy vehicles	Firms: Subsidiaries of several domestic car companies and the FAW-VW joint venture. Subsidiaries of numerous supplier firms for the Guangzhou-based Japanese auto joint ventures and for FAW-VW Academic Institutions: At least two new hydrogen and fuel cell R&D institutes jointly established with domestic labs and universities, including Wuhan University of Technology
High-tech Industries
Electronic information	Optoelectronic devices and chips	Firms: Subsidiaries of several domestic and Greater China firms Academic Institutions: Several new optoelectronics and semiconductors R&D institutes jointly established, inter alia, with universities from the province (e.g. Sun Yat-sen University) and neighbouring Hong Kong (Hong Kong University of Science & Technology) as well as various institutes of the Chinese Academy of Sciences
Biomedicine	Dental medical equipment	Firms: Subsidiaries of domestic pharmaceutical companies Academic Institutions: At least one biomedicine R&D institute, jointly established with the Chinese Academy of Sciences
New materials	Nonwoven fabrics for medical use, Energy-saving materials	Firms: Subsidiaries of domestic materials firms Academic Institutions: Materials R&D institutes jointly established with labs and universities from the province

Source: Nanhai FYPs; NY (2002–2019)

Third, non-local actors enjoy a broad range of preferential policies if they locate a subsidiary within the region. However, with a few exceptions, the local government does not provide its policy support simply in the hope of achieving industrial growth and employment but tends to condition support on the non-local actor’s willingness to contribute to regional industrial transformation by implementing platform projects together with the government (authors’ interviews). These projects pertain to the establishment of platforms undertaking one or more functional activities that are specified in the relevant government plans. The functions, include, inter alia, applied R&D, technology transfer, business incubation, technology consulting, technology training or standard testing. Preferential access to land, government funds and highly skilled employees is subject to the actor’s collaboration with the local government on providing such functions.

In this case, innovation platforms are founded on collaborative projects between government and non-local (or capable local) actors with a view to performing specific intermediary functions. As joint projects, platforms adopt a (quasi-) public organisational form, even if the localised platform initiator is a private entity. The government’s objective of establishing such platforms is to anchor knowledge and capabilities related to the prioritised areas within the region. Over time, project actors can collaborate with the government on several projects. They do so because they are allowed to profit from operating the state-owned assets and funds brought into the project directly or indirectly. What functions to invest in und how to integrate them in business models depends on the platform actors’ entrepreneurial decisions. While the relevant priority areas are chosen in a top-down fashion, the transformation is tackled in a bottom-up, trial-and-error fashion by multiple market-oriented project actors. The government knows what it wants and the project actors are there to discover how to get it.

4.3. Collaborative platform development in equipment manufacturing

Collaboration between government and non-local actors is also a key element of Nanhai’s branching out towards robotics and precision engineering within equipment manufacturing. Two collaborations, one with a university and one with a technology firm, particularly stand out. The subsidiaries of both platform partners are recognised as provincial-level ‘New R&D Institutes’ (Conlé et al.2021a; NY 2019). The two New R&D Institutes include the Foshan Nanhai Guangdong Technology University CNC Equipment Cooperative Innovation Institute (hereafter Foshan GDUT Institute) and the Foshan Institute of Intelligent Equipment Technology (hereafter Foshan IET Institute).³

4.3.1. Cooperation with a non-local university

The Foshan GDUT Institute is one of the increasingly numerous university satellite institutes that are established in Guangdong cities to improve university-industry knowledge transfer (Conlé et al. 2021a, 2021b). Creating platforms for strengthening university-industry linkages has been a major thrust of Guangdong’s regional innovation policy since the early 2000s. The Foshan GDUT Institute was established in the context of Guangdong’s 2011 Plan, a provincial-level platform policy launched in 2013 based on the respective national plan, which promotes university-industry cooperative innovation centres in strategic sectors. The institute was jointly established in 2013 by the government at provincial, municipal, and district level, as well as Guangdong University of Technology (GDUT), the province’s major engineering university and one of the best engineering universities of the country, to promote the regional equipment manufacturing industry.

According to the Foshan innovative city plan, GDUT’s Nanhai-based satellite institute is supposed to leverage the Guangzhou-based parent university’s expertise to contribute to the region’s industrial upgrading and diversification objectives. Since its inception, the satellite institute engages in four intermediary functions, including the provision of innovation services for regional firms, the incubation of technology firms within the government-prioritised sectors, the R&D of key enabling technologies to boost Nanhai’s pillar industries, and regional skill development. To engage in these functions, the satellite institute has received generous land allocation, buildings, and initial start-up investment (authors’ interviews). However, it does not enjoy the government’s permanent funding commitment. Instead, the government ties its funding and provision of state assets to the various platform projects, while it expects the institute to generate revenue from its various activities.

Each of the Foshan GDUT Institute’s intermediary functions is associated with one or more platform projects (Table 3). Regarding applied R&D, the satellite institute has established two platforms. The two (provincial-level) ‘research engineering centres’ are state research units that connect non-local expertise in intelligent manufacturing systems and industry robotics with the specific regional demand for industrial automation. More concretely, the engineering centres are platforms to which research staff from the parent university and/or (inter-) national experts from the government recruitment programs (e.g. the national-level ‘Thousand Talents Program’ or the district-level Nanhai ‘Blue Ocean Talent Program’) can attach in order to work on technological solutions for regional industry. Regarding incubation, the Foshan GDUT Institute’s development is geared towards the government’s increased focus on ‘mass entrepreneurship and innovation’. In line with national and provincial-level plans (13^th FYP period), the institute has established an ‘incubation chain’ that covers two national-level makerspaces, a national-level business incubator, and a business accelerator (that was under construction at the time of the authors’ visit). The platforms particularly support the formation of businesses by innovation teams from the parent university and joint ventures of the GDUT Institute with technology firms from outside the region.

Table 3. Innovation platform projects by two platform partners.

Type of innovation platforms	Foshan GDUT Institute	Foshan IET Institute
Entrepreneurship platforms	▪▪	▫
Applied R&D platforms	▪▪	▪
Technology service platforms	▪▪	▪
Technology training platforms	▫	▪▪

Source: Own compilation based on material provided by institutes.

Note: ▪ denotes number and scope of platforms related to given intermediary function.

While the local government provides guidance on the region’s industrial priorities, it is up to the institute’s management to decide how to fit the various functional activities and platforms into a consistent business model. In the case of the Foshan GDUT Institute, the institute has gradually taken over the role of a technology broker that collects and articulates the demand of regional industry and then jointly develops solutions through its interregional innovation networks (Foshan GDUT Institute presentation; Conlé et al. 2021b). As a public technology service platform, the institute has received government support in developing solutions like automated demonstration lines for upgrading traditional industry manufacturing processes that it can sell to local firms at government-subsidised prices (authors’ interviews). The Foshan GDUT Institute incubates some of its solutions itself, as it benefits from its entrepreneurship activities because of its venture capital function. At the same time, Nanhai’s industry benefits from the increase of equipment manufacturing firms and from the technological solutions geared towards the demands of regional traditional industry (authors’ interviews).

4.3.2. Cooperation with a non-local technology firm

The Foshan IET Institute has its roots in robotics research at one of China’s most prestigious manufacturing-related R&D platforms – the National Numerical Control Systems Engineering Research Centre in the capital of Hubei Province, Wuhan. The R&D platform was co-established by Huazhong University of Science & Technology and its own spin-off, Wuhan Huazhong Numerical Control (HCNC), a university-owned diversified machine tool company established in 1994 and listed in Shenzhen in 2011. When HCNC began developing robotics technology applications for automobile parts manufacturing and precision engineering, it gradually launched related subsidiaries in various cities. HCNC located a subsidiary in Foshan in 2015. The company came to Foshan to seize the opportunities arising from Guangdong’s plans to prioritise robotics and advanced manufacturing equipment development. Offering its cooperation to the local government, HCNC’s original plan was to confine its regional engagement to the establishment of a testing centre and an entrepreneurship platform. However, seeing the comparative advantage of the Foshan GDUT Institute’s robotics incubation facilities, HCNC reassessed its strategy (authors’ interviews).

After consultation with the government at municipal and district level, the company refocused towards applied industrial robot R&D and commercialisation. While HCNC would not be able to receive direct funding for setting up its individual for-profit manufacturing activities, the company could expect government support by establishing its R&D centre as a platform project (authors’ interviews). HCNC thus co-established the Foshan IET Institute together with the government as a public institution without permanent funding guarantee. As in the above case, the government provided land and a lump-sum start-up investment to launch the institute. Follow-on support has been contingent on the institute’s securing competitive research funding from government and industry.

The Foshan IET Institute used its entrepreneurial flexibility to develop a suitable business model (Table 3). It limited its incubation activity to establishing two robotics companies. The two companies, both recognised as national high-tech companies, are majority-owned HCNC subsidiaries, with the Foshan government holding a little stake in them in return for its policy support (authors’ interviews). Both firms’ products are geared towards the regional market. While one of the firms engages in the development and manufacturing of small-sized and lightweight industrial robots as well as related components, robot software, and system integration solutions, the other one is a specialised servomotor producer for intelligent manufacturing applications.

The Foshan IET Institute primarily supports the HCNC subsidiaries by focusing on robotic core technology R&D, the development of system integration solutions, and talent training (Foshan IET Institute webpage and exhibition hall displays). For this, the institute reaches out to branches of national-level engineering research centres and key labs from Wuhan and Shenzhen. By linking up to these extra-regional platforms, the Foshan IET Institute – and Nanhai – has gained talent and experts from outside the region, especially from Huazhong University of Science & Technology. More importantly, the institute has initiated the Foshan Robotics Training Center which responds to the government’s call to overcome the shortage of robotics technicians by cooperating with vocational schools to provide comprehensive technology training. In this connection, the institute’s target audience consists of both, enterprise workers and school educators. While increasing the regional skill level, the institute’s training programs connect well with the industrial and educational robots manufactured and offered by HCNC’s Foshan subsidiary.

5. Discussion

Our analysis of Nanhai’s smart specialisation strategy demonstrates how innovation platforms in Guangdong – and possibly in other parts of China – are utilised to facilitate the transformation of regional industrial structures. Since regions like Nanhai lack the capabilities for endogenous upgrading and diversification, local government seeks to develop platform projects in cooperation with non-local actors from both, industry and academia. In its stylised form, the Guangdong innovation platform model differs notably from the most elaborate and salient model of innovation platforms discussed in the context of smart specialisation (and constructed advantage) in Europe – the Lahti innovation platform model elaborated by Harmaakorpi (2006) and collaborators. Table 4 summarises the differences with regard to priority identification, entrepreneurial discovery, and policy support.

Table 4. Characteristics of two innovation platform models.

	Lahti Model	Guangdong Model
Purpose	Local knowledge recombination	Non-local knowledge anchoring
Priority identification	Bottom-up identification based on inter-sectoral stakeholder conversations	Top-down identification of priority areas and bottlenecks
Entrepreneurial discovery	Collective discovery process around a composite theme (e.g. megatrend), facilitated by a systemic intermediary	Engaging non-local actors to perform specific intermediary functions; discovery of profitable technology transfer business models
Policy support	Policies to support resource formation along the new growth path	Policies to support the platform’s importation and anchoring of growth paths

5.1. Priority identification

Nanhai’s diversification strategy is embedded in the Chinese government’s sustained effort at promoting overall industrial upgrading and transformation to maintain the country’s growth momentum and continue catching up with the advanced economies. The overall layout is thus strictly hierarchical. On the one hand, this circumvents problems with bottom-up consultation processes in formulating regional strategies emphasised in the smart specialisation literature (Kroll 2015; Capello and Kroll 2016; Estensoro and Larrea 2016). On the other hand, this seems to go against a central tenet of the smart specialisation approach, as it denies the participatory, bottom-up approach, which is supposed to help adapt strategy to the regional context. However, smart specialisation, with its mission orientation (Foray 2018), does not preclude top-down elements. Foray (2019), moreover, argues that the fear for duplicative strategies among regions, which is associated with top-down prioritisation, is unwanted as entrepreneurial discovery typically occurs in later phases. Top-down exploration of policy alternatives still needs to go hand in hand with bottom-up entrepreneurial activities (Gifford, McKelvey, and Saemundsson 2021).

In China, the industrial focus is on strategic emerging industries, many of which generate general-purpose technology in need for adaptation to concrete regional demand conditions. Industrial diversification, specifically concerning equipment manufacturing, information technology, and new materials, are to facilitate the further task of upgrading traditional industries (Nanhai 11th FYP). By looking for synergies, local governments promote ‘continuity and change’ (Moodysson, Trippl, and Zukauskaite 2017), proactively forging relatedness and facilitating the industry’s development of new growth paths. Nanhai’s support of advanced robotics and precision engineering seeks to achieve both, to promote local equipment sector development and provide custom-fit technological solutions for the upgrading of local traditional manufacturing sectors.

The local government’s pragmatic policy practice adds a bottom-up element to this. Just like other local governments, the Nanhai government accepts the tasks outlined in the higher-level plans but adapts them pragmatically to local realities, seeking synergies between existing and new sectors as well as redefining sectoral boundaries in ways that combine current and new paths. Examples include Nanhai’s targeting of ‘Chinese medicine biotechnology’ to cater to existing pharmaceutical firms and capabilities (Nanhai 12th FYP) or the emerging new materials sector’s focus on nonwovens for medical use, based on initial activities by local firms from Nanhai’s traditional textile industry (Nanhai 13th FYP; NY 2019). As a by-product, the practice grants local governments the opportunity to report higher levels of regional activity in those (broadly defined) sectors than would be possible by relying on new ventures only.

5.2. Entrepreneurial discovery

The literature considers entrepreneurial discovery as the essential process for integrating ‘dispersed, decentralised and divided’ local knowledge into regional new path development (Foray 2016: 1433). Contrary to the Lahti model, the Guangdong model does not emphasise regional participatory processes. This does not mean that the latter model lacks a discovery mode, however, because entrepreneurial discovery and bottom-up participation need to be distinguished (Foray 2019).

In the Lahti model, the participatory process sets in at the very beginning of new path development, when the priority areas are identified based on inter-sectoral stakeholder conversations, utilising ‘expert panels’ (Harmaakorpi 2006) and ‘innovation sessions’ (Melkas, Uotila, and Tura 2016). The platform is defined based on the identified priority, giving direction to the ‘core process’ (Pekkarinen and Harmaakorpi 2006) of collective discovery of innovative opportunities. Private or (quasi-) public ‘systemic intermediaries’ (Parjanen and Hyypiä 2018) are the pivotal platform actors who manage the core process by acting as brokers and coordinators of interactions among the regional stakeholders. While the bottom-up logic can help overcome many shortcomings of top-down planning (Foray 2016), empirical research demonstrates two important weaknesses that prevent lagging regions from implementing this model. They include the shortage of capabilities and governance capacities (Kroll 2015; Capello and Kroll 2016; Estensoro and Larrea 2016), including a lack of competent intermediaries, as well as the problem of incorporating non-local actors and knowledge (Hassink, Isaksen, and Trippl 2019; Giustolisi, Benner, and Trippl 2022).

More recently, the smart specialisation literature argues in favour of including relevant stakeholders at a later stage, when complementary projects and activities are developed and implemented within the priority areas that were chosen in a top-down fashion (Foray 2019; Foray, Eichler, and Keller 2021). The Guangdong model has a unique take on the latter approach. In that model, government plans do not only determine priority areas but also include an outline of a transformational map defining the main tasks to be undertaken and general tools to be utilised to achieve the transformative objective. Based on these guidelines and on existing opportunity conditions, local governments seek out capable industrial or academic collaborators from inside and, more importantly, outside the region to participate in the joint development of multiple platform projects. A platform project pertains to the provision of an intermediary function, or a combination of functions, such as business incubation, applied R&D services or technology consulting. By taking on these projects, localised actors like the Foshan GDUT Institute or the Foshan IET Institute connect regional industry with external knowledge and technology suppliers, contributing to anchoring the relevant knowledge within the region.

Bottom-up participation is limited in this process, although plan making in China more generally involves a broad range of experts and a profound knowledge of regional industrial structures. Nonetheless, the projects between government and various platform actors generate substantial opportunities for differentiation and entrepreneurial discovery. Regarding further differences between the Lahti and the Guangdong model, the nature of platforms and discoveries stands out. In the Lahti model, the platform actors use platforms to coordinate regional collective learning in a designated priority area. In the Guangdong model, in contrast, platform actors use platforms to introduce R&D and non-R&D technological services that are in short supply within the region. They do so by discovering viable business models that combine the local government’s demand for knowledge anchoring with the platform actor’s profit motive.

New business models can reach a remarkable degree of hybridisation, as they combine private with state assets, profit with non-profit activities, closed with open boundaries, private with collective goods, and services with industry. In our cases, the Foshan GDUT Institute connects its spinoff activities to its technology market platforms, incubating spin-offs that, as technology suppliers, address major technological needs prevailing on the regional industry’s demand side, while the Foshan IET Institute develops robots that it deploys for its regional technology training activities. Other cases we have researched in the Pearl River Delta have developed further varieties based on their platform projects.

5.3. Policy support

Non-local actors enjoy a broad range of preferential policies when they bring their knowledge to the region. The policies are tied to contributing to the tasks and performing the intermediary functions defined in the relevant plans. Notably, the tasks and policy tools are defined at the national and provincial level, although they are applied at the regional level. Local governments perform their tasks by developing platform projects together with competent partners, who, in the case of lagging regions, primarily come from outside the region. The notion of innovation platform allows the government to limit collaboration with any partner to the desired intermediary functions.

For entrepreneurial actors from the industry or academic sector, collaboration on platform development can be an interesting option. As joint projects between government and non-government actors, the platforms themselves tend to be treated as (quasi-) public organisations, even if they are integrated into private ventures. Innovation platforms are legal entities, although they are often not clearly separated physically from the platform actor’s local R&D, entrepreneurship, training or other facilities. Having a government-recognised innovation platform is a signal of the organisation’s eligibility for accessing complementary innovation policies. R&D platforms, for instance, enjoy priority access to state research funding programs, preferential tax rates, import tariff rates for technological equipment, and can serve as affiliations for highly trained staff from government talent programs. Some New R&D Institutes, which had established entrepreneurship platforms such as business incubators, were allowed to manage state venture capital funds. For the platform actors, it is yet also a commitment to use its facility in ways that help implement the transformative activity.

6. Conclusion

In many lagging regions, strategies for modernising and transforming regional industrial structures are difficult to realise without a transplantation of external knowledge. Involving non-local actors in new path development can be beneficial for a region if these actors’ activities create externalities and facilitate access and diffusion of knowledge that is new to the region. Chinese policy experiences suggest that innovation platforms can serve as a tool for regional knowledge anchoring. In China’s Guangdong province, local governments cooperate with non-local knowledge actors in platform development, incentivising and capacitating their non-government platform partners to perform activities deemed relevant for the regional industry’s transformation.

The use of innovation platforms is distinct from the one featured in European smart specialisation debates. Instead of facilitating a collective search for new growth paths based on recombinations of regionally available knowledge, innovation platforms in the Chinese setting target the performance of particular transformational activities and intermediary functions. Top-down processes play a fundamental role in this approach. The choice of platform partners follows the definition of priority areas determined within the hierarchically structured planning system. Bottom-up processes are more difficult to identify but are actually of utmost importance to tie activities to existing regional industry. Local governments facilitate entrepreneurial discovery by contributing their knowledge of regional demand structures and by motivating localised knowledge actors to forge relatedness between existing and emergent paths. Entrepreneurial discovery is primarily associated with the localised actor’s market-driven integration of platforms into a profitable business model.

Non-local actors are attracted to the region because of the region’s entrepreneurial opportunities. This also applies to academic institutions, which are not requested to reproduce themselves on smaller scale (e.g. as branch campuses) within the region. Instead, universities and public research institutes send entrepreneurial staff and alumni as managers to the region to establish applied research and innovation service ventures, drawing on and linking up to the parent institution’s innovation resources.

In this paper, we have alluded to two factors for the attractiveness of the region. First, under China’s system of nested authority, priority definition in regional industrial plans signals the state’s commitment that investments will flow into these areas. A large chunk of state investments flows into capacity development through innovation platform construction. Teaming up with local governments allows actors to join these growth opportunities. Second, regional demand for related products and services is important. Our study relates to one of the industrially most developed parts of China, the Pearl River Delta. Still, despite their manufacturing strength, innovative capabilities and path diversification capacities have been relatively weak. The regions are, or have been, (knowledge-) peripheral, though they are not marginal economically.

Previous literature demonstrates that the characteristics of peripheral regions differ strongly (Eder 2019). Therefore, more research is needed to find out in which ways the two key factors – the institutional system of top-down planning and regional demand conditions – restrict the applicability of similar platform models within lagging regions elsewhere. Moreover, concerning locational choices, there are two further relevant points that we were unable to address here. First, in the Chinese context, local government cooperation with non-local universities and public research institutes partly rests on agreements among higher-level government bodies. Examples include the Guangdong provincial government’s ‘Plan on Developing Industry-Academia Linkages in Guangdong (2007–2011)’ with the national-level Ministry of Education and the 2009 ‘Comprehensive Strategic Cooperation Agreement between Guangdong Provincial Government and CAS’. The implementation of these agreements deserves more attention. Second, innovation platform construction constitutes but the most basic element of multi-level infrastructure building in the Pearl River Delta. Megaregion construction likely adds to the overall attractiveness of the whole ‘Greater Bay Area’, while it will affect smart specialisation strategies within this megaregion. Industrial change in megaregion development is thus another topic that we think merits further study.

Acknowledgments

The authors wish to thank an anonymous researcher for the helpful comments and suggestions that helped us strengthen our argument considerably. They also thank their colleagues Haixiong Qiu, Qing Qiu, and Yutu Yang from the Institute for Reform and Development of the Pearl River Delta at Sun Yat-sen University for their cooperation and support during field research as well as Cornelia Storz and Henning Kroll for comments and advice. The authors gratefully acknowledge the assistance of Fei Wang and Sibei Lin.

Disclosure statement

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

Additional information

Funding

This work was supported by the German Research Foundation DFG [Grant Numbers: TE 1069/5-1; TE 1069/5-2].

Notes

¹ Hierarchical coordination through plans has been playing a prominent role in China’s drive for industrial modernisation and catch-up, although that role has evolved over time. According to Heilmann and Melton (2013), the transition towards a ‘new mode of plan making’, which involves planning with and for markets (instead of substituting for them), has been completed in the early 2000s. We thus concentrate on the time since the 11^th FYP, starting in 2006.

² The provincial-level ‘Specialised Town’ policy has been instrumental in facilitating upgrading and transformation (Barbieri, Di Tommaso, and Huang 2010). Concerning rationalisation, in the ceramics industry, for instance, all but ten firms survived the clean-up (NY 2013).

³ As the platform projects enjoy the support of the higher-level governments, they are included in higher-level development plans. Both the Foshan GDUT Institute and the Foshan IET Institute are included in Foshan’s 13th FYP as important ‘public innovation service platforms’. Moreover, the Foshan GDUT Institute, which is also referred to as a ‘large public innovation platform’ in the aforementioned plan, is listed as a major ‘technological innovation project’ in the Foshan Implementation Plan for Constructing an Innovative City (2013–2020).