From regional to global and back again? A future agenda for regional evolution and (de)globalised production networks in regional studies

ABSTRACT

This paper builds common grounds for a future research agenda in the regional studies of evolutionary economic geography and global production networks. I put forward two ‘troubling themes’ of (geo)politics and heightened risks as the most disruptive forces in today’s increasingly fragmented global economy and argue for their significance in regional studies throughout the post-pandemic 2020s. Massive global change through the reconfiguration of and strategic (de/re)coupling with global production networks will engender new path formation in regional transformation. In this analytical move from the global ‘back again’ to the regional, there are common questions on epistemology (causal explanations) and substantive issues (network/regional resilience, institutions/the state, inequalities/uneven development and new forms of regional policies) for both communities of researchers.

KEYWORDS:

• Evolutionary economic geography
• global production networks
• geopolitics
• risks
• regional economies
• research agenda

JEL:

• F23
• L22
• R58

1. INTRODUCTION

The interdisciplinary field of regional studies has witnessed some rigorous debates on theories and approaches in recent years. In his 2021 Area Development and Policy lecture, Ron Boschma (2022), the leading proponent of evolutionary economic geography, has engaged with four strands of literature in the field that share a common concern with regional development in a world of global value chains. In these broad literatures on global production networks, clusters, global innovation systems and the geography of functions, he argues that their interactions with evolutionary economic geography have been mostly implicit and underdeveloped, and he proposes several useful research agendas to bridge each of them in relation to the central tenets of evolutionary economic geography (see also Boschma, 2024, in this issue). To a certain extent, Boschma’s (2022) efforts have been productive in addressing some of the key concerns in my 2020 Regional Studies lecture (Yeung, 2021) that seeks to theorise how the strategic coupling of local and regional actors with global production networks can serve as an extra-local/regional causal mechanism for explaining regional diversification and path development, in addition to the region-specific endogenous role of related variety and economic complexity much theorised in evolutionary economic geography (Boschma, 2017; Boschma et al., 2017).

This duality of exogenous and endogenous causation of regional evolution, however, should not be viewed conceptually as a rigid dichotomy or a binary either/or choice in epistemology. In both theory and practice, strategic coupling and related variety almost always interact in dialectical ways that shape any regional formation and transformation in such interconnected worlds of global production networks. In this sense, they are mutually constitutive – the branching of regional actors into related industries is reinforced by their strategic coupling with global production networks that in turn strengthens and/or upgrades the firm-specific capabilities of these actors for further related diversification within regions. Viewed in this dialectical conception, the key concepts underpinning theories in evolutionary economic geography and global production networks are not only complementary in their conceptual orientations, but also critically necessary in explaining the geographies of industrial change from the regional to the global scale – regional changes and capability development are dynamically coupled with global processes of innovation and production. Here, I concur with Poon’s (2024) commentary in this issue that argues for theorising regional change through the analytical lens of dynamic capabilities at the firm and regional levels.

In this reflexive piece framed within the spirit of the above positive scholarly exchanges and those in this issue, I want to take a different analytical tack and thus the paper’s title ‘back again’. I argue that much of our interconnected world of regions has unfortunately changed since our exchanges in the previous two journal lectures, even though they are very recent! During this intervening period, the world has experienced massive economic disruptions and (geo)political shifts taking place since the COVID-19 pandemic at the beginning of the 2020s. Localisation and regionalisation of production networks in the name of supply chain resilience and ‘national’ security have come to the forefront in considerations by many policy makers and business practitioners throughout the world. Such directional change, and perhaps deglobalisation in the political-economic organisation of the world economy, will likely produce very different ‘regional worlds’ for evolutionary economic geography and global production networks studies in the post-pandemic world (Yeung, 2023a). As discussed further in this paper, fundamental change in the geographical scales of industrial transformation has begun and will easily last throughout the 2020s and even beyond. It is about returning ‘back again’ from the global to the regional economy or the resurgence of regions. But this regional resurgence is linked to a geo-politicised kind of ‘regional economies’ very different from those previously conceived in the pioneering work of Storper (1995, 1997), Scott (1998, 2001) and others.

In this highly (geo)politicised world of regionalisation and deglobalisation throughout the 2020s, a key set of new research questions for regional studies is about whether the regional scale might become a development trap (cf. Diemer et al., 2022; Pinheiro et al., 2024) or a new ‘spatial fix’ (cf. Peck et al., 2023) in this changing world order? This ‘critical juncture in the world economy’, coined in Lee’s (2024) introduction to this set of exchanges, might well be the time for new path formations/path breaking and regional diversification. But what sort of path dependency and relatedness/related variety will be involved in driving such regionally oriented industrial transformation and production networks? These are some of the key common grounds on which a future research agenda for regional studies in evolutionary economic geography and global production networks can be developed and exchanged. While this exchange paper offers no clear answers, I hope its more speculative and forward-looking posture can help break some old barriers through complementary work in both strands of literature in economic geography and regional studies. In this regard, I very much welcome Boschma’s (2024) contribution in this issue that attempts to theorise explicitly the role of complexity and relatedness in shaping the origin and evolution of regional upgrading and downgrading in global production networks. The next two main sections will first elaborate on the critical issues confronting the ongoing remaking of the global economy before revisiting several key common grounds in such a complementary research agenda comprising both epistemological and substantive issues.

2. REMAKING THE GLOBAL ECONOMY

In my earlier Regional Studies lecture, I argued for a relational and yet integrative view of understanding regional change in a world economy characterised by highly globalised production networks. Seeking a middle ground for the study of regions in evolutionary economic geography and global production networks approach (see also Boschma, 2022, 2024), this integrative view of regional change places significant explanatory power on the interactive effects of regional assets and extra-regional production networks in order to overcome an endogenous ‘container’ view of regional worlds and a too globalist approach common in global value chains studies sponsored by international organisations (e.g., UNCTAD, 2013; World Bank, 2020; WTO, 2021; ADB, 2023). In short,

a necessary and sufficient condition for analysing regional transformation needs to take into account both internal dynamics for change and extra-regional opportunities embedded in GPN [global production networks]. This integrative approach requires regional studies scholars to examine regional evolution through tracking related and unrelated diversification initiatives and their strategic coupling or ‘fit’ with the competitive dynamics of relevant GPN. (Yeung, 2021, p. 993; original italics)

But what if those demand conditions and competitive dynamics of relevant global production networks have now been significantly reshaped due to much wider and unexpected conjunctures of events and processes at the global scale? How will they impact on the research agenda in both evolutionary economic geography and global production networks approaches? Will this remaking of the global economy bring back again regional economies as the motors and frontiers of industrial transformation and economic development? Will it disrupt preexisting pathways of regional evolution and enable both advanced and peripheral regions to break out of path dependency and develop new growth paths? What are its implications for the evolution and adaptation of regional systems and networks? How will politics and institutions become more influential in such regional evolution and change?

Here, I consider two such ‘troubling themes’ of (geo)politics and heightened risks that are likely to be highly significant for regional studies throughout the 2020s. These two are among some of the most disruptive forces in today’s increasingly fragmented global economy, others being technological shifts, climate change, social upheavals, armed conflicts and so on.

2.1. (Geo)politics in regional studies

Inter-national politics led by big power rivalries has apparently become the forerunner steering global economic change in the post-pandemic era. Despite the claims of some pundits (see Alami and Dixon, 2020, 2023; Whiteside et al., 2023), this form of the so-called ‘state capitalism’ is not new in light of the policy failures of state-led interventions in many developing countries since the 1970s (Yeung, 2016; Hamilton-Hart and Yeung, 2021) and the previous pursuit of strategic trade policy in the United States and then European Economic Community in the 1980s (Block and Keller, 2011; Mazzucato, 2013; Weiss, 2014). But what is new and unusual is its increasingly widespread adoption, even in market-based neoliberal economies, and its extra-territorial reach and transnational policy tools. This unusual mixing of domestic/national politics for electoral gains, ‘making us great again’, and ‘rebuilding better’ with inter-national/geopolitical imperatives (e.g., decoupling from China and sanctions on Russia and, in specific industrial sectors, China) in recent years has effectively turned (geo)politics into perhaps the single most potent force in disrupting the existing pathways of regional evolution and global production arrangements in both industrialised economies and developing countries (see Witt, 2019; Witt et al., 2023).

Some empirical specificities are useful to illustrate these major (geo)political imperatives. In the United States and Western Europe, national states at the federal level and the European Union (EU) have taken a much more proactive role in economic management and policymaking. Most significantly, geopolitical and security concerns have now been firmly inscribed into new industrial and technology policies that will last for a long time to come. In terms of industrial policy such as the 2022 Bipartisan Innovation Act and Inflation Reduction Act in the US and the 2023 European Chips Act and the proposed European Green Deal Industrial Plan in the EU, massive state and EU-level subsidies have been enacted to promote high tech industries and digital innovation, such as semiconductors, artificial intelligence, telecommunications and electric vehicles, and to bring production back to home regions. While this kind of nationalistic ‘Made in Home Country’ approach in industrial policy has been in place in large developing countries such as China and India since 2014, its regulatory adoption and widespread support across the political spectrum in the US and the EU in the early 2020s will likely embolden many more nationalist states to take a similar ‘beggar-thy-neighbour’ approach to reshoring, ‘friend-shoring’, localisation of production and so on (see also De Propris, 2024).

Unlike their policy (in)effectiveness in the 1980s though, the world economy then was organised rather differently in the form of multi-national worlds of industrial and innovation systems led by ‘national champions’ or leading domestic firms that were well supported by home governments through public procurement and subsidies in research and development (see Yeung, 2022a, pp. 13–16). For example, Japan’s share of the world semiconductor market peaked at 46% in 1986 and exceeded the 43% share held by American firms such as Intel, Motorola and Texas Instruments. As the inventor of DRAM memory chips in 1970 (and microprocessors in 1971), Intel exited the very DRAM market it founded in 1986 that compelled its then CEO Andrew Grove (1990, p. 159) to complain that ‘Intel is a sizable and strong company, but we are located in the wrong country. All of the action in our industry is moving to Japan’. The US federal government promptly switched back to heavy state support through policy interventions. Offering a comprehensive analysis of the British electronics industry in their pioneering work Microcircuits of Capital, Morgan and Sayer (1988, p. 65) point out this apparent irony in relation to state interventions in the US:

The American industry began with heavy state support. It quickly progressed to a period of world leadership in the process of internationalization, during which it legitimized itself by appealing to a static conception of comparative advantage and to free markets as a realm of natural justice and order. When such arguments failed to stem the rise of a new state-backed power which refused to accept the static view of comparative advantage, it switched sides. Examples of such opportunism are of course rife in the history of capitalism, but this particular instance is notable in the light of the celebration of the alleged role of free market forces in the development of the industry in the ideology of high tech.

When such a nationalistic form of strategic trade policy in semiconductors and large-scale computer systems was implemented in the 1980s against competitors from Japan (then and now a geopolitical ally of the US!), vertical integration was still the dominant mode of organising industrial production among the largest semiconductor and mainframe computer firms in the US (e.g., IBM, Sperry Rand, DEC, Texas Instruments, Motorola, Intel, Fairchild and National), Western Europe (e.g., ICL, Bull, Olivetti, Philips, Siemens, SGS and Thomson) and Japan (e.g., Fujitsu, Hitachi, NEC and Toshiba). Up until the early 1990s, the global market for electronics had an estimated value of $100 billion, but only about 5% was being outsourced (Cassia, 2010, p. 9).

In the early 2020s, however, more than half of the world’s electronics production was outsourced and organised through intricate global production networks connecting diverse regional economies across the global economy. And yet Andrew Grove’s 1990 lament remains quite valid whereby much of the action in global semiconductor manufacturing has moved to East Asia (Yeung, 2022b; Yeung et al., 2023). In short, the same set of strategic trade and industrial policies in the late 1980s did not work well and, when fully implemented by the mid-2020s, will find it much harder to produce the intended positive outcomes for home regions in the US and the EU (e.g., more production, employment, innovation and linkages). For example, the tens of billions of public subsidies and infrastructural resources put up by state/regional governments in the US and the EU since 2021 to attract new semiconductor fabrication facilities – known as fabs in industry terms – by TSMC (Arizona, US), Samsung (Texas, US) and Intel (Dresden, Germany) have been confronted with the grim reality of overcapacity and weak demand of chips. The global semiconductor market in 2023 shrank by 11% (https://www.gartner.com, accessed on 26 January 2024). Despite their much larger fabs in home regions in South Korea (Hwaseong, Giheung and Pyeongtaek) and Taiwan (Hsinchu and Tainan) humming loudly and churning out state-of-the-art memory and logic chips, Samsung and TSMC have respectively announced significant delays in mass production in their new US fabs till 2025 and 2027/2028.

In technology policies, explicit sanctions on specific technology exports to geopolitical rivals and economic competitors such as China have been rolled out since the Trump administration and further reinforced by the Biden administration in the US (and will likely remain so, irrespective of Biden or Trump winning the next US election in November 2024). Unlike those voluntary import expansions imposed on Japanese producers of memory chips in the 1980s (e.g., the US-Japan Semiconductor Trade Agreement of 1986; see Tyson, 1993) and the government-led consortium of national firms (e.g., SEMATECH or Semiconductor Manufacturing Technology in the US in 1987 and, earlier, the European Strategic Programme for Research and Development in Information Technology in 1983; see Langlois et al., 1988), these renewed and punitive technology policies in the 2020s are highly intrusive in terms of target-specificity and extra-territorial reach in policy tools and their applications, particularly in relation China and Chinese high-tech firms (e.g., Huawei, ZTE, SMIC and so on).

In this partial approach to technological decoupling from China, the national security state in the US has even compelled its allies, such as EU member states (e.g., the Netherlands and Germany) and Japan, to follow its ‘chokepoint’ approach to limiting the growth and development of China’s semiconductor industry and, more broadly, information and communications technology (ICT) applications, such as supercomputers, artificial intelligence, military equipment and so on. In semiconductors, proprietary equipment manufacturers from the US (Applied Materials, KLA and Lam Research), the Netherlands (ASML) and Japan (Tokyo Electron, Nikon and Canon) are world-leading and dominant players in equipping wafer fabrication facilities (known as ‘fabs’). Their equipment tools are indispensable in the manufacturing of advanced chips at 10 nanometres or lower, including complicated processes of cleaning, deposition, etching, lithography and so on in clean-room conditioned fabs.

While this techno-nationalist approach to technological sovereignty in semiconductors may not make much impact on the existing geography of advanced chip production mostly centred on East Asia (Yeung, 2022b), it will impact significantly on the R&D and related innovation activities in semiconductor manufacturing, both in the home regions of these equipment makers (e.g., the US, the EU and Japan; see Huggins et al., 2023) and the new centres of innovation in other emerging regions (e.g., Shanghai and Shenzhen in China, Hsinchu and Tainan in Taiwan and Seoul and Gyeonggi-do in South Korea; see Gao et al., 2023; Yeung et al., 2023). To a large extent, the same intent and purpose of this extra-territorial export control might also be applied to technology policies in other high-tech sectors such as energy, green transport (including battery technologies and electric vehicles), biomedical sciences, digital economies and so on. As noted in De Propris’ (2024) commentary in this issue, the current race for Industry 4.0 represents a transformative historical moment for regional economies because of the irreversible, cumulative and disruptive choices made by governments, lead firms and their partners in global production networks.

2.2. Risks in regional studies

Heightened risks in remaking the global economy are clearly another major ‘troubling theme’ for regional studies in this decade and beyond. Recent events during the 2020–2023 COVID-19 pandemic have accentuated all kinds of risks associated with the overdependency of critical supplies and production, supply chain disruptions, financial market upheavals and market fragmentation. Many of these risks go well beyond the sort of (geo)political risks discussed earlier that tend to concentrate on issues of supply chain resilience and reduction of overdependency through national and macro-regional diversification. Beyond these geopolitically infested risks, critical issues for regional studies include business risks at the firm level and institutional risks at the regional scale. In my recent monograph Interconnected Worlds (Yeung, 2022a), I outline five risks that confront firms and non-firm institutions in different local and regional economies and offer some examples in the global electronics industry (see Table 1). Many globalised industries are now characterised by rapid technological shifts, massive production fragmentation and cross-regional outsourcing, and the rise of new markets and competitors. Local and regional actors are confronted with a far greater sense of uncertainty and unpredictability because poor risk mitigation can present disastrous consequences for them and their home economies (e.g., the devastating supply chain disruptions during the pandemic peaks in the 2020–2022 period and the high inflation arising from such disruptions in 2023 and possibly 2024).

Table 1. Risks in global production and implications for firms and regional economies, 2010–2023.

Form	Nature	Causal effects on firms	Examples in global electronics
Economic risk	Systemic shifts in markets – new technologies and innovations, changing demand, financial disruptions, exchange rate fluctuations, border closure and national lockdowns and so on	Loss of production capacity and competitive position in cost and/or market leadership; reduction in financial returns and profitability; lower income and structural volatility to localities and regions	Decline of Canada’s RIM (BlackBerry), Finland’s Nokia, and South Korea’s LG in smartphones since 2013; demand shift and supply disruptions due to the COVID-19 pandemic in 2020 and beyond; worldwide shortage of semiconductor chips for automotive production in 2021–2023
Product risk	Quality, safety, branding and efficiency considerations	Negative views of goods or services by consumers and customers; greater demand for corporate social responsibility	Samsung Galaxy Note 7 faulty battery and fire risks in August-September 2016
Regulatory and geopolitical risk	Political, public-to-private governance, changing standards and norms and big power conflicts	Disruption or termination of global production, existing industrial practices and organisational arrangements	US trade sanctions on China’s ZTE, Huawei, and SMIC, 2018–now; Japan’s restriction on exports of chemicals for chip making to South Korea, 2019; massive lockdowns worldwide and highly politicised push for domestic production and reshoring during the COVID-19 pandemic in 2020–2022
Labour risk	Struggles over working conditions, payment delays and employment practices	Resistance and industrial action by employees; worker rampage and violence; disruptions to global production and employment prospects; and potentially greater reputational risk	Strikes in Foxconn’s plants in China, maker of Apple’s iPhones, due to workers demanding for better terms and working conditions, 2012–2013; temporary shutdown of Wistron’s India factory due to contract worker rampage in December 2020; labour supply crunch due to travel restrictions during the COVID-19 pandemic in 2020–2022
Environmental risk	Natural hazards, public health crisis or human-made disasters	Accentuating the above four forms of risk and their causal effects	Japan’s 2011 Fukushima earthquake and production stoppage at Japanese semiconductor fabs; distortions to global chip markets from SK Hynix’s Wuxi (China) plant fire in 2013, winter storm-induced shutdown of Samsung and NXP fabs in Austin (Texas) in February 2021, and Renesas’ Naka (Japan) fab fire in March 2021; the COVID-19 pandemic in 2020–2023 and unprecedented global production disruptions thereafter

Source: Yeung (2022a, Table 3.2, pp. 104–105; reworked from Yeung and Coe, 2015, Table 2, p. 41).

Some of these risks in Table 1 are known as ‘GPN risks’ by virtue of their origin in and transmission through the interconnected worlds of global production networks (see Coe and Yeung, 2015; Yeung, 2022a, p. 106). Their nature and effects in the 2020s and beyond have a far broader geographical scope and faster temporal diffusion. The 2020–2023 pandemic has thrown into sharp relief such extensive geographical reach and rapid effects of global supply chain disruptions. Geographically, these risks can spread across the entire range of spatially dispersed actors and their home regions that are well articulated into different global production networks. Mitigating this kind of global production network (GPN) risk can have profound causal impact on firm strategies, capability-building, and regional resilience in the post-pandemic world economy (see next section on research agendas).

Let me give one specific empirical example. In late 2020 and throughout the 2021–2023 period, significant shortages in semiconductors that are necessary to ICT end products and automotive assemblers contributed to the crippling of industrial production in many of their factories worldwide. These chips, such as Wi-Fi and Bluetooth chips, analogue power chips and microcontrollers, are specific to end products from iPads and notebooks to passenger cars. Chip factories in Europe and Southeast Asia were impacted severely by the unexpectedly prolonged lockdowns and movement restrictions between late 2020 and early 2022. To a certain extent, the practice of just-in-time supply chain management and the opaque demand planning in the automotive industry worried leading chip makers such as Taiwan’s TSMC and its chipmaker customers (top auto-chipmakers such as NXP, STMicroelectronics and Infineon outsource advanced logic/digital chips to TSMC). All these demand and supply risks led to their fab capacity reallocation to other industrial customers and thus the chain reaction of chip shortages and lower production output among all major car makers in 2022 and 2023. This process of chip shortages, in turn, impacted severely on local and regional economies hosting major car makers (e.g., GM and Ford plants in the US, Canada and Mexico shutting down for periods of time and Toyota cutting total production by up to 20% in 2022).

In existing studies of evolutionary economic geography and global production networks, core research topics tend to focus on socio-technical systems (e.g., transition studies), regional growth trajectories (e.g., the relatedness, economic complexity and regional diversification literature), change agency and institutions (e.g., studies of regional development paths) and strategic coupling of local firms and regional upgrading in global networks. Few researchers are seemingly concerned with the above kinds of ‘GPN risks’ that may fundamentally reshape local and regional change because of their unpredictable nature and unknown conditioning effects. This blind spot of missing out heightened risks and uncertainties can be detrimental to the future relevance of both literature strands in a more geopoliticised world of technological decoupling, market fragmentation and ultimately economic de/reglobalisation. Though a small number of studies have examined the risks inherent in production network governance (e.g., Billing and Bryson, 2019; Bryson and Vanchan, 2020; Yeung, 2023a, pp. 674–675), these risks are not yet well theorised in evolutionary economic geography, and this brings me to some common grounds for future research in the next section.

3. IN SEARCH OF KEY COMMON GROUNDS: EPISTEMOLOGICAL AND SUBSTANTIVE ISSUES

Given the above two ‘troubling themes’ of (geo)politics and heightened risks – over and above other disruptive forces in today’s increasingly fragmented global economy – I think there is much work to do among researchers interested in regional evolution and (de)globalised production networks. Lee’s (2024) introduction has already identified that together, both fields of evolutionary economic geography and global production networks studies have produced nearly 2000 papers indexed in the Scopus database, an indication of the tremendous potential for future breakthrough work on the basis of shared common grounds. Here, I believe two broad common grounds can be forged in these interrelated fields: (1) epistemology and (2) substantive issues.

3.1. Epistemology in regional studies

I think the two literature strands indeed have much in common – the search for causal explanations and theories in their historical-regional/geographical contexts. Interestingly, causality and explanatory theories have recently returned to the forefront of epistemological debates in economic geography and regional studies (Bathelt et al., 2017; Gong and Hassink, 2020; Martin, 2021; Peck, 2023a; 2023b; Yeung, 2023b). Constructing causal explanations can be a productive way forward to rebuild the analytical rigour, responsible practice and public relevance of these interrelated disciplines. In my new work Theory and Explanation in Geography (Yeung, 2024), I make a strong case for ‘our theories’ to be explanatory in order for critical human geography to develop practically adequate mid-range theories and to address challenging societal and planetary changes. Earlier in Global Production Networks (Coe and Yeung, 2015), we have outlined such a causal theory – nicknamed ‘GPN 2.0’ – in which we theorise the competitive dynamics (‘why’) and the causal mechanisms (‘how’) underpinning firm-specific choices, their geographical configurations (‘where’) of production activities and, more importantly for critical scholars in regional studies, the uneven outcomes in local and regional development (‘so what’). Through different strategic coupling pathways and value capture trajectories, local and regional economies experience differentiated costs, benefits and incentives in such variegated configurations of global production networks. Such diverse regional outcomes include technological change, new firm formation, industrial transformation, social and economic upgrading and so on.

Similarly, and despite the possible conflation of causal mechanism with process, my earlier work in Yeung (2019, pp. 244–246) has demonstrated how evolutionary economic geography has provided causal explanations of regional evolution and economic change through the concept of path dependence. In this highly influential geographical analysis of regional evolution, the explanatory focus is placed on the adaptive transformation of the economic landscape and the uneven process of regional development (Boschma and Martin, 2007; MacKinnon et al., 2009, 2019; Martin and Sunley, 2015). As the central plank in the call for evolutionary theory as ‘a truly new and promising paradigm in economic geography’ (Boschma and Frenken, 2006, p. 295; also Boschma, 2022, p. 125) and as ‘the main explananda in economic geography’ (Boschma and Frenken, 2006, p. 295; original italics), path dependence has become the core organising concept to comprehend place-based processes that in turn require evolutionary-geographical explanations (see also Martin and Sunley, 2006, 2015; Martin, 2010). Conceptually, path dependence is viewed as a process in which the past influences the present (and the future). Causal explanation is clearly the raison d'être of this kind of evolutionary economic geography. In particular, path dependence has been used to explain causal outcomes in the location behaviour of firms, the spatial evolution of industries and networks, the co-evolution of firms, technologies and territorial institutions and the convergence or divergence in spatial systems (see also Boschma and Frenken, 2018; Boschma, 2022).

3.2. Substantive issues in regional studies

This common ground in search for causal explanations of regional change, whether from an ‘outside-in’ viewpoint (global production networks) or from an ‘inside-out’ perspective (evolutionary economic geography), can be highly productive in understanding post-pandemic industrial transformation in local and regional economies. In Table 2, I portray four such evolutionary-geographical scenarios in the case of global electronics, but clearly their relevance can be found in many other industries such as aerospace, automotive, machineries, pharmaceuticals and so on. These diverse scenarios for regional evolution bring me to several key substantive issues arising from geopolitical policy shifts towards, and heightened risks in, the resurgence of regions and regionalisation in the post-pandemic world economy – (1) network and regional resilience; (2) institutions and the state; (3) inequalities and uneven development and (4) new forms of regional policies. Together, they constitute the second common ground for the regional studies of evolutionary economic geography and global production networks.

Table 2. Evolutionary-geographical scenarios of regional change: some examples from global electronics.

Regional scenario	Evolutionary process	Regional change	Examples in global electronics
Inside-out	The internationalisation of a domestic firm to other localities outside its home region	This inside-out process of foreign investment enables a domestic firm to embed into global production networks coordinated by lead firms based elsewhere or to develop into a global lead firm in itself by incorporating other firms and non-firm actors into its expanding production network.	Taiwan’s Wistron or Foxconn setting up assembly factories in regions in China and India. Samsung in Northern Vietnam or Xiaomi in India’s Andhra Pradesh and Tamil Nadu.
Inside-out	The export of intermediate goods or services by a domestic firm to other firms outside its home region	This inside-out form of international trade in industrial goods and software services allows a domestic firm to remain in situ geographically and, yet, embed in global production networks coordinated by lead firms and their network partners based elsewhere.	A semiconductor foundry in Singapore exporting chipsets to assembly factories in Malaysia’s Penang and China’s Shanghai American firms exporting software products to be installed on personal computers and smartphones assembled in China’s Chongqing and Chengdu and Mexico’s Juarez
Outside-in	The inward investment of foreign firms in the home region of a domestic firm	This outside-in process by either lead firms or platform/technology leaders from elsewhere facilitates the embedding of local and regional firms into global networks mediated through these foreign firms. With the participation of these foreign firms, domestic firms and production networks become progressively globalised over time through outsourcing or strategic partnership arrangements with the local subsidiaries of lead firms and/or technology leaders.	Hewlett-Packard or Qualcomm setting up regional headquarter and/or R&D operations in Taiwan’s Taipei or Hsinchu in order to be close to their strategic partners there
Outside-in	The import of intermediate goods or services by a domestic firm from dedicated suppliers outside its home region	This outside-in mode of network embeddedness takes place when foreign suppliers and platform or technology leaders become articulated into the globalising production network governed by a locally embedded domestic lead firm.	Qualcomm and Broadcom from California supplying wireless chipsets to Oppo’s in-house smartphone factories in China’s Dongguan and Shenzhen The dependence of Chinese personal computer and smartphone lead firms on operating systems from Microsoft and Google respectively in Seattle and Silicon Valley of the US

Source: Reworked from Yeung (2018, p. 386, 2022a, pp. 91–92).

First and foremost, the post-pandemic scalar shift from the global ‘back again’ to the regional in relation to geopolitical imperatives and risk mitigation will necessitate the rethinking of how production networks can be regionalised in ways that make them more robust and resilient to external shocks and risks. As such, network and regional resiliences are critical agendas for both literature strands. Here, the robustness of production networks within macro-regions such as the Americas, Europe and Asia refers to their capacity to recover during sudden and unexpected disruptions such as pandemic lockdowns. And yet network resilience is more about such recovery after major disruptions and ruptures and thus takes a longer time frame.

In both literature strands of evolutionary economic geography and global production networks, the study of networks as a meso-level unit of analysis for explaining regional evolution has been very well established (see Yeung, 2023b). Evolutionary economic geography, for example, views networks as critical organisational infrastructure for knowledge diffusion and innovative activities within industrial clusters and regional economies (Boschma and Frenken, 2010, 2018; Boschma et al., 2017; Grillitsch, 2019; Hassink et al., 2019). As well argued in Boschma’s (2022, p. 127) lecture, there is no reason why both literature strands cannot embrace each other’s approach to networks and their role in regional economies. In doing so, they can collectively contribute to the identification of new growth paths and upgrading possibilities for different regional economies. Overall, I concur with his observation. Surely unpacking such networks at the global scale and within regional economies can contribute much to our understanding of post-pandemic restructuring in the world economy, irrespective of whether it is about politically charged notions of deglobalising, decoupling or de-risking or about overdependency on such national economies as China in manufactured goods and the US in technology products/services, and/or the excessive concentration of industrial power in specific macro-regions (e.g., Northeast Asia in semiconductors and the EU in advanced manufacturing).

Second, both literature strands have already emphasised the importance of institutions and the state in regional evolution (e.g., Rodríguez-Pose, 2013, 2021; Pike et al., 2017; Boschma, 2022, 2024) and the governance of global production networks (e.g., Henderson et al., 2002; Gereffi et al., 2005; Smith, 2015; Coe and Yeung, 2019). In fact, regional institutions are highly valued and broadly discussed in Coe et al.’s (2004, pp. 471–473; Table 1) original paper on the strategic coupling of regional actors with global production networks. This line of thought on the role of institutions and the state in the development and evolution of global production networks remains highly visible in my later work on East Asian industrial transformation published in Regional Studies (Yeung, 2009, pp. 332–335) and, in a monograph form, Strategic Coupling (Yeung, 2016). Here, as a call for a common ground in both literature strands, I concur again with Boschma’s (2022, p. 128) observation of a ‘lack of geographical wisdom’ in some regional studies applying uncritically the relatedness framework in evolutionary economic geography,

without considering too much the specific social, political and institutional context in which processes of regional diversification take place … Diversification in China is likely to be very different from diversification in Latin America because of prevailing policies and institutions, but there is still little understanding of how it differs. Studies in EEG have started to explore how national and regional institutions enable or impede diversification in certain places, and how varieties of capitalism and institutions affect the type of diversification in regions.

My two ‘troubling themes’ of (geo)politics and heightened risks discussed earlier seem to compel us to bring such ‘geographical wisdom’ back into both literature strands! Yes, institutions and the state matter. But arguing and providing evidence for ‘institution matters’ is one thing in these literature strands (e.g., Farole et al., 2011; Rodríguez-Pose, 2013, 2021; Smith, 2015; Werner, 2021); theorising and demonstrating how these institutions and the state co-evolve and transform together with the competitive dynamics of global production networks is quite another matter and will perhaps take on a more (geo)political orientation (e.g., Yeung, 2014, 2016, 2023a). In the post-pandemic 2020s, there is no question at all that the (geo)political has come to shape powerfully the competitive logics of innovation and production networks within diverse capitalist regions in the world economy. Much of the current (geo)political pressures and national policy priority has been placed on (re)building supply chain resilience through the localisation and regionalisation of production networks within and among advanced industrialised countries and their friendly allies in different macro-regions.

Ironically though, this reorientation of national security state-led/regulated approach to governing global production networks in different industries, from trade and investment policies to technology and national security concerns, will likely punctuate the previously dominant neoliberal world order established and reinforced through international organisations such as the World Trade Organization, the World Bank and the International Monetary Fund. And yet sub-national regional institutions may also resist such national state-imposed industrial and technology policies. Taken together, I believe such divergent interests and multi-scalar tension in this nexus of national states vs. international organisations and regional institutions in governing the post-pandemic deglobalised world order will present researchers in both evolutionary economic geography and global production networks studies new opportunities for theorising and demonstrating empirically why and how ‘institutions matter’ – and at what geographical scale(s) – in regional evolution and change (see my earlier examples of semiconductors in the US and the EU).

Third, rebuilding network/regional resilience through institutional and state-led efforts will invariably accentuate spatial inequalities and uneven development in the post-pandemic 2020s. This common ground for both literature strands can almost be taken for granted since geographical differentiation in regional evolution and change has always been couched as the ‘dependent variable’ in evolutionary economic geography and global production networks studies. In fact, the two literature strands have shared much research interest in the evolutionary change of regional clusters, innovation systems, economic complexity and functional specialisation (see Boschma, 2022, 2024). In the 2020s and beyond, this change might not be evolutionary in nature due to greater geopolitical interventions and risk mitigation initiatives that may push for revolutionary and path-breaking transformations in regions and regional economies. These uncertainties in the future pathways of regional development and their consequential uneven outcomes for local and regional actors ought to be the core common concern in both communities of researchers.

Last, but not least, the post-pandemic changing world order has led to new forms of regional policy agendas that seemingly break away from the so-called ‘Washington consensus’ set in place since the early 1990s. In this sense, globalisation in the form of unfettered marketisation and neoliberalisation of the world is dead. Long live globalisation, as a more strategic, regionalised and resilient kind! But what sort of regional policies might help achieve such an interconnected world without the high risks and highly uneven and exploitative outcomes? The current vogue of using geopolitically infested industrial and technology policies to rebuild better and to force the reshoring or ‘friend-shoring’ of productive activities can only go so far to address domestic political and security concerns (cf. the same kind of policy intent and purpose against Japanese producers in the 1980s and yet perverse and unintended outcomes of promoting growth in other East Asian locations by the 2000s).

The EU, for example, is much more concerned with de-risking its supply chain dependency on China rather than decoupling from China. South Korea and Taiwan are reluctant partners in joining US-led technology alliances such as ‘Chip 4’ since late 2022 (including Japan). Specific technology firms such as Intel and Dutch-based ASML questioned the institutional logic of restricting semiconductor manufacturing in China, their biggest market. In brief, the geographical implications of these techno-nationalist policies for specific regions in the US and the EU can be daunting. In the US, for example, it is almost certain that the $52 billion incentives in the 2022 CHIPS Act (Creating Helpful Incentives for Producing Semiconductors for America and Foundries Act) will bring highly uneven benefits to different regions within the US – existing centres of chip fabs will be highly advantaged. The same uneven ‘regional problem’ will also arise from the €43 billion European Chips Act passed in April 2023.

In the US case, the Commerce Department has asked for specific requirements on potential recipients of CHIPS Act incentives to provide childcare facilities that benefit those regions hosting such recipients (e.g., Arizona, California, New York, Oregon and Texas). Despite its strategic policy intention, the grounded reality is that few or none of those American semiconductor firms (e.g., Intel and Micron) or foreign fabs (e.g., TSMC, Samsung and GlobalFoundries) might take up any of the CHIPS Act incentives because of two highly restrictive conditions of refraining from making excessive profits and investing in China, Russia, Iran and North Korea for 10 years. The former condition requires recipients to reveal their highly secretive corporate accounts in terms of all costs and revenue streams. With these restrictive conditions, even Intel is mulling over the option of applying for the Act’s incentives. In their unusual gestures in March 2023, the trade and economic ministers in South Korea and Taiwan even spoke out openly on behalf of their national champions (i.e., Samsung/SK Hynix and TSMC) on the unacceptability of these restrictive conditions under the CHIPS Act.

4. CONCLUSION

Taken together and summing up this exchange paper, I believe (geo)politically expedient national policies must be built on a far better understanding of regional economic complexity (well examined in evolutionary economic geography) and the competitive dynamics of industry-specific production networks (grounded in global production networks studies). Otherwise, and often lacking in empirical evidence, such wishful state-sanctioned policies in trade, investment and technology arenas may not only be ineffective in their implementation, but also create unintended consequences of greater regional inequalities and counter-productive outcomes that impact perversely on the people and places in these regions. Unprecedented events in the early 2020s have led to massive global change and thus post-pandemic reconfigurations of and strategic decoupling/recoupling with global production networks in relation to new path formation in regional evolution and change. Whether we are able to capitalise on this once-in-a-lifetime change and opportunity in post-pandemic regional studies depends on our willingness to engage with each other and advance key common grounds in a future agenda for researching into regional evolution and (de)globalised production networks.

ACKNOWLEDGEMENTS

This is a revised paper first written for, and presented at, the Regional Studies Exchanges Plenary at the Regional Studies Association 2023 Annual Conference held at Ljubljana, Slovenia, 16 June 2023. I thank Jessie Poon for inviting me to participate in this very important event, Ron Boschma for his gracious engagement with my work and the global production networks literature and Neil Lee, Lisa De Propris and Jessie (again!) for their encouraging and thoughtful commentaries. My excitement from the huge turnout of over 500 participants in person at the plenary session must be fully acknowledged here; so are the many constructive suggestions and questions there. Finally, I am grateful to helpful comments for revisions from John Harrison at Regional Studies. All errors and/or mistakes are mine.

DISCLOSURE STATEMENT

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

Additional information

Funding

This work was supported by the National University of Singapore: [Grant Number E-109-00-0008-01].