Introduction
In June 2017, at the Seoul Centre for the Creative Economy and Innovation located right next to the South Korean capital’s most important former royal palace, our interviewee explained why the government was bankrolling his center: “The government wants conglomerates to gain innovative DNA by working with startups.” What we heard was that the center—which assists startups by providing coworking space, access to coaching, and mentorship and networking—was not necessarily designed to bolster startups. Instead, investment in startups was a means of injecting “innovative DNA” into Korea’s large firms. This was the first time we grasped that startups were considered to be resources to boost, not challenge, the competitiveness of big business.
This interview stands out as a crystalizing moment. Until then, we had understood that government efforts like the center were fundamentally for startups that, in this book, represent new, high-growth, and often technologically oriented firms. It was surprising to hear that this flagship startup initiative positioned startups as a resource for other firms. We had thought that headlines of East Asian governments’ investment in startups represented a break with their developmental past, which was centered on relations with, and investment in, large firms. Our intention was to write a book about how even quintessential developmental states were succumbing to the global trend of adopting a Silicon Valley–styled approach.
That moment came to symbolize our finding that government policy often conceives of startups as innovation resources for big businesses, not as the ultimate beneficiaries. The finding that startup largesse is not necessarily serving startups is important because economic growth, national security, and public health are dependent on a state’s innovative capacity. The contemporary innovation imperative has been evident in the US-China trade war (Zhen 2023) and the race to develop a vaccine against COVID-19 (Neate 2021). There is no question that we are living in an epoch that can be called startup capitalism: startups play a central role in market economies’ techno-industrial competitiveness. However, many questions about the how and why of startup capitalism remain unanswered. Businesses, governments, and wider society need to understand what drives today’s cutting-edge innovation, and what role startups and corporations play in advancing national innovation capacity. Does startup capitalism involve startups as disruptive engines of innovation or resources fuelling established firms’ innovativeness?
In this book, we answer this question by debunking two myths. The first myth is that startup promotion is intended for startups. A Silicon Valley–styled version of startup-led innovation where new firms disrupt industries is considered the gold standard. We instead find that government policy often conceives of startups as resources for—not challengers to—conglomerates. Startups inject new ideas, talent, and ways of working to enable the future competitiveness of lead firms that must compete with foreign rivals. This is a model that directly challenges the stylized Silicon Valley approach.
The second myth is that startup promotion is merely a means of entrepreneurship or employment policy. While startup policies do aim to boost employment, especially among the youth, they are a contemporary form of industrial policy. Targeting specific sectoral and technological capacities, industrial policy endeavors to bolster national economic and industrial competitiveness. Historically, industrial policy emphasized the use of financing to enable the productive capabilities and export activities of established firms operating in strategic industries. Today, we argue that startup promotion is a growing tack that sees startups (David), working with (or even for) large firms (Goliath), fueling national capabilities at the world’s technological frontier. As a result, high-technology entrepreneurs and accompanying venture capital markets are part of today’s high politics; they fuel countries’ prowess in critical technologies, informing national economic competitiveness and security. Thus, the nature of the industrial-military complex in the twenty-first century is not only about big-scale government contractors; it also hinges on national supplies of startups’ cutting-edge technologies and disruptive thinking.
Our finding that startup policy is not necessarily for startups is something of a puzzle. For comparative capitalism (Hall and Soskice 2001; Schneider 2013; Witt et al. 2018), startup-centric approaches should indicate a liberal market economy (LME) model in which disruptive entrepreneurs, with their accompanying equity financiers and flexible labor markets, fuel radical innovation. From this perspective, East Asian governments’ startup capitalism would signify convergence on the Anglo-American LME, and as such, movement away from a coordinated market economy (CME).
In a similar vein, for developmental state scholars, startup largesse should infer the death of the developmental state that was characterized by its coordination with and directing of credit to large firms. In the classic developmental state, big businesses held central positions as employers and innovators.1 On the face of it, startup-centric initiatives, such as Korea’s CCEI and Japan’s J-Startup Initiative, should represent a break with that large firm–centric model. These policies should instead signify that these countries are converging on operating as entrepreneurial states (Tiberghien 2007; Mazzucato 2013) and startup nations (Senor and Singer 2009). For evolutionary economists, policy to encourage innovative startups should indicate an adherence to a paradigm in which economic growth and revolutionary innovation are driven by startups disrupting existing technologies and the positions of existing firms (Aghion et al. 2021; Akcigit and Van Reenen 2023). So, policy to boost startups should indicate that startup capitalism manifests as startups posing existential threats to incumbent firms’ competitive positioning.
In contrast to these expectations, state initiatives for startups do not necessarily constitute a break in governments’ close relationships with big business. In this book, we show how and why startup capitalism, which we define as an economic and political system in which startups contribute to employment, innovation, and growth, can take multiple forms. Startup capitalism can manifest according to contrasting logics—some that align with the notion that startups disrupt existing industries and some that fit better with open innovation models in which established firms leverage startups to benefit their competitive positioning. We argue that policy makers simultaneously draw on opposing Schumpeterian logics; they see startups as both disruptors and resources for big businesses.
With this book, we advance debates in a few important ways. First, we contribute to political economy scholarship by showing that startup capitalism does not necessarily signify the death of the developmental state nor convergence on a neoliberal paradigm. Instead, startup capitalism can exemplify the persistence of CME complementarities and the developmental status apparatus. We show how startup policies often strive for both creative destruction and the augmenting of incumbent firms’ innovation capacity. Second, we offer new evidence that may help explain why government policies targeting startups are not delivering results. Economic research has found that since the start of the twenty-first century, productivity and business dynamism have been declining in major advanced economies, including in the United States (Decker et al. 2016a; Philippon 2019). The reason being that government efforts that are ostensibly striving to cultivate disruptive innovation are, in fact, oriented toward expanding the innovation capacity and competitive positioning of established firms. We contribute to evolutionary economics by extending Schumpeterian understandings of the patterns of innovation into the context of startup capitalism.
This book’s focus is on East Asia. Japan, Korea, Taiwan, and China offer archetypal cases of state-led development and a mix of market economy types, which is helpful in challenging the notion that embracing startup capitalism is synonymous with a declining developmental state or convergence on an LME. In addition to the region’s analytical significance, there is its empirical importance. China is the second largest VC market in the world and is second to only the US in terms of its number of unicorns (Chen 2023);2 Japan ranks highest globally for the number of innovation outposts in Silicon Valley (JETRO 2019); Korean corporations (Kakao and Samsung) are among the world’s most active VC investors (CB Insights 2019); and Taiwan—long construed as the small firm–centric developmental state—has a single firm (Taiwan Semiconductor Manufacturing Company, or TSMC) that leads the world’s semiconductor manufacturing (Ryan 2022) and has been said to be the world’s best example of a state-engineered VC market (Gulinello 2005). In other words, East Asia is an ideal region for studying continuity and change in governmental efforts to boost startup-fuelled innovation.
We find that there are varieties of startup capitalism in East Asia. Japan and Korea’s startup capitalism is best seen as a continuation of their developmental state models, which are persistent in efforts to network startups with big businesses. However, this is now done in the guise of an “open innovation” mode (Chesbrough 2003; Weiblen and Chesbrough 2015). In these cases, provision for startups can best be understood as a key tool in a broader arsenal of means for bolstering the technological innovation of established firms. While remaining mostly in line with its early depiction as a small firm–oriented developmental state dubbed the Silicon Valley of Asia, Taiwanese policy continues to be dedicated to widening pools of entrepreneurs in emerging technologies. As the China chapter reveals, the country’s engagement with startups is a complex one, with a mix of support for disruptive startups and efforts to challenge the dominant positions of select industry giants.
In all cases, startups are construed as resources for big businesses in areas considered critical, or emerging, technologies, such as artificial intelligence (AI), quantum computing, robotics, and semiconductors. This aligns with notions of a techno-nationalist state (Plantin and de Seta 2019), a techno-security state (Cheung 2022), and a national security state (Weiss 2014) and the role of creative insecurity (Taylor 2016) in marshaling investment into startups as part of military-industrial complexes around the world (see Nicholas 2019). Building a venture capital state has been motivated under the guise of boosting national security (Klingler-Vidra 2018). Startups—the recipients of venture capital funding—are engines for critical technology prowess.
Startups are often engaged in a means of boosting large firms’ ability to compete with other national champions. Even markets considered quintessentially neoliberal—like the United States—include entrenched companies as crucial innovation partners. For instance, the US National Science Foundation partnered with Amazon Web Services, IBM, and Microsoft in 2022 to boost quantum computing capabilities, a field in which the United States is locked in competition with China and other leading economic powers. The NSF also partnered with Microsoft to offer cloud computing credits to startups participating in its Big Data Regional Innovation Hubs in 2016. The incumbents benefit directly from this integration into public startup policies, as they lock in swaths of startups as users.
Big (tech) companies are key partners in startup policies in Europe, too. In 2021, for instance, Nesta, an innovation agency headquartered in the United Kingdom, and the European Commission’s Startup Europe Partnership launched a new ranking for Europe’s 25 Corporate Startup Stars, which included BMW, Microsoft, and Telefonica, to further celebrate corporation-startup interactions (Dunsby 2021). German efforts to boost Dresden and the surrounding area as Silicon Saxony include the US$11 billion spent on attracting a TSMC chip plant, a key element in the European Union (EU) Chips Act, in order to link industry leaders such as TSMC to growing European producers (Blanchard and Escritt 2023) while limiting unwanted foreign investments in the technology sector (Chazan 2019). Meanwhile, the Macron administration’s 2019 aim for France to produce 25 unicorns was reached in 2022, with tailwinds coming from clusters built around the French government’s handouts to Taiwanese firms such as ProLogium in Dunkirk (Economist 2023a). By partnering with startups, these companies are praised by policy makers for not perishing as Blockbuster and Kodak, two former multinationals whose respective business models and technologies became obsolete, did.
Is this approach best for the economy and society as a whole? Seminal academic theory, in the form of Joseph Schumpeter’s work on patterns of innovation, suggests no. Over the course of the first decades of the twentieth century, he distinguished two industrial paradigms—one in which startups served as essential disruptors and one in which large companies fostered innovation and growth, owing to their ability to mobilize their significant resources. While both modes of innovation were separately valid, he posited that they did not occur simultaneously. He studied which mode could best drive innovation in a particular industry at a given time.
In this book, we hope to reinvigorate this exercise to ascertain how startup capitalism can effectively drive business dynamism and innovation. This is an important debate, as research has revealed a dearth of competition as big businesses increasingly dominate many industries in the twenty-first century (Tepper 2023). Studies show that there has been a decline in business dynamism and entrepreneurship in the United States since 2000 (Decker et al. 2016b). Thus, it has never been more important to understand how countries can best engage startups as engines of economic growth and national security.
Basis for the Puzzle
The existence of institutional similarities and differences across economies is now well established. Many scholars agree that the global economy is not converging on the so-called US style of neoliberalism and that different economic models coexist, even in the context of innovation-oriented activities (Boyer 2005). Among the different strands of literature discussing these similarities and differences from a comparative perspective, the varieties of capitalism (VoC) approach, first established by Peter Hall and David Soskice (2001), has become dominant. Using an institutionalist approach, Hall and Soskice’s edited volume theorized a dichotomy between LMEs, such as the United States and the United Kingdom, and CMEs, typified by Germany and Japan.
VoC delineates why LMEs and CMEs differ in the way that firms behave. Both market economy types possess complementarities across the institutional arenas of labor organization, the nature of interfirm relationships, corporate governance, the availability of information, and models of financing. These institutional complementarities are internally coherent and self-reinforcing and, according to the authors, shape how their firms compete in global markets.3 In the case of LMEs, market mechanisms define the arm’s-length relationship between firms accompanied by equity finance and fluid labor markets. As for CMEs, nonmarket mechanisms in the form of strategic interactions among firms underpin this relationship, with reinforcing debt-based patient capital and rigid labor markets.4 LMEs, owing to their institutional logics, are expected to excel in radical innovation, while CMEs are expected to lead in incremental innovation. While the LME and CME varieties are “ideal” models that, in practice, operate differently from country to country, they have become popular anchors to describe and analyze how market economies operate.
Underpinning much of the research on the East Asian region is the developmental state concept, which refers to late industrializing economies, with the state at the center of capacity upgrading and economic growth.5 The developmental state can be summarized as a focus on long-term (decades-long) strategic goals driving socioeconomic development, the existence of a quasi-autonomous bureaucratic apparatus staffed by talented officials seeking to maximize benefits for society as a whole, and the use of institutionalized mechanisms for public- and private-sector cooperation (Amsden 1989; Wade 1990; Evans 1995). The developmental state offered support to specific sectors or firms in a bid to advance technological capabilities, move up global value chains, and thus provide economic growth, steady employment, and national security (Woo-Cumings 1999). Largesse included the mobilization of citizens’ banking or postal savings accounts, which were used to bolster corporations’ productive capacity through access to credit (Calder 1990, 2017; Vogel 2018). In return, these large firms would create steady and relatively well-paid jobs to employ the country’s abundant labor force (Johnson 1982; Okimoto 1989). The developmental state, in these different constellations, encouraged domestic firms to manufacture and export goods that relied on Western technology, at least in the initial phases (Westney 1987), acting as a guardian for the economy and society (Tate 1995).
Synthesizing this depiction, we understand the four key defining characteristics of the developmental state as: (1) central or strong long-term state intervention in economic policy planning and implementation; (2) a perceived importance of quality employment for socioeconomic purposes; (3) a preference for bank-based financing conducive to the existence of patient capital; and (4) relatively high levels of ownership concentration, whether family- or state-dominated. A crucial aim was to accumulate technical capacity, initially in catch-up technologies but ultimately at the world’s technological frontier, thus escaping the middle-income trap (Lee 2013). This depiction has been reiterated across the economics literature, which finds that countries initially achieve growth through technology transfer and adoption and later move toward innovating at the technological frontier (Acemoglu et al. 2006; Peters and Zilibotti 2023). The literature thus draws our attention to systemic coordination and institutional complementarity often centered on large firms as drivers of technical capacity advances and suppliers of well-paid, stable employment.6
Whether the developmental state is dead or alive has been much debated. Some argue that crises such as Japan’s Heisei Era bubble burst, the East Asian Financial Crisis (EAFC), political shifts such as democratization, and the states’ economic successes, have led to a death or decline of state-led orchestrating of techno-industrial output (Lim and Jang 2006; Pirie 2018). Others point at the structural changes in the global economy that have been underway since at least the 1980s as having diminished the role for the region’s developmental states in fostering industrial capabilities (Wong 2004; Yeung 2016). Globalization and the neoliberal Washington Consensus have also been said to have eroded the role of the state in economic policy making (Wade 2018).
Others contend that the developmental state is very much alive—or evolved—given relatively unchanged normative contexts and institutional structures. Scholars argue that the underlying mindset that drove the developmental state has persisted (Thurbon and Weiss 2006; Stubbs 2009; Thurbon 2016). They assert that policies continue to be implemented in the context of a deeper entrenching of state interventionism in the economy (Lee 2024). In fact, beyond the East Asian setting, there’s growing awareness of, and acceptance for, states’ pursuit of industrial policy aimed at boosting techno-industrial capabilities (Aiginger and Rodrik 2020; Juhász et al. 2023). In other words, state interventionism has not disappeared in East Asia; rather, like a chameleon, its characteristics and modalities have changed, and have grown in global popularity.
To explore the ways in which startup capitalism constitutes continuity and change, we draw on Schumpeterian understandings of industrial dynamics that foment technological innovation. Neo-Schumpeterian economics posits that innovation is the main driving force behind economic growth (Hausman and Johnston 2014; Perez 2002, 2016). In the wider literature on evolutionary economics, scholars examine the relationship between firm size, market dynamics, and systemic technological change (see Nelson and Winter 1982). This tradition has conceptualized two distinct modes of how firms, and the characteristics of accompanying industrial systems, shape innovation activities. These two modes are referred to as Mark I and Mark II.
In what is referred to as Mark I, the emphasis is on Schumpeter’s (1934) espousing of the benefits of new entrants—startups, in today’s parlance—in instigating processes of creative destruction. Creative destruction is understood as the dynamic in which startups revolutionize “the economic structure from within, incessantly destroying the old one, incessantly creating a new one” (1942 [2010], 73). Startups “continually arrive to compete with existing firms,” and their new technologies “render existing technologies obsolete” (Aghion et al. 2021, 1; Akcigit and Van Reenen 2023, 1). In this mode, creative destruction poses a threat to established companies to revolutionize a way for new firms and technologies to boost productivity and, as a result, drive economic growth. The Mark I policy implication is that public policy needs to remove constraints to innovation for high-potential entrepreneurs, including access to market entry and growth inhibitors imposed by large firms and their nonproductive strategies (Hanusch and Pyka 2007; Baslandze 2023). Said differently, Mark I startup policies should encourage the widening of entrepreneurial pools so that would-be disruptors are able to form and grow.
In contrast, Mark II, emanating from Schumpeter’s Capitalism, Socialism and Democracy (1942), emphasizes the productive role of large firms as innovators in the first decades of the twentieth century. Oligopolistic competition dynamics endow big businesses with the financial means to invest in new products and technologies. To compete with the other dominant firms, these lead firms invest in research and development (R&D) at scale. In this context, industry giants are the engines of transformative innovation. The public policy imperative is to enable established firms’ access to resources. This aligns, in some ways, with the developmental state paradigm in which industrial policies ensured access to credit, organized consortia, and more.
By considering startup capitalism in both Mark I and II terms, this book represents a departure from seminal work on the developmental state. Rather than seeing startup policy align with either Mark I or II ideals, we find hybridity. Startup capitalism can strive for widening pools of entrepreneurs and startups can be perceived as a means of enabling innovation capacity among oligopolistic firms (see Sandulli et al. 2012; Dahlander et al. 2021). In this way, policies encourage oligopolistic firms to leverage startups as open innovation system resources (Pacheco Pardo and Klingler-Vidra 2024). This differs from the Mark II ideal type in which large firms did innovation in-house.
Our findings also challenge expectations that the rise of startup capitalism is evidence of the region subscribing to a universal convergence on a LME logic. Rather than seeing startup capitalism strive for Mark I’s creative destruction dynamics, we find that startup policies that do align more with Mark I tend to emphasize emerging industries and technologies, but not disruption. They are fostered as argonauts advancing a new frontier rather than being seen as a challenge to existing businesses.
Startup capitalism either avoids confrontation with big businesses or pursues collaboration with incumbents whose innovation capacity will benefit from interactions with startups. To analyze these variations systematically, we map startup capitalism in terms of the institutional logics of VoC (employment, finance, innovation) and the developmental state (firm relations and social purpose). In doing so, we strive to understand startup capitalism in wider political economy terms and to investigate which means of engaging startups is optimal for society, given industry patterns and life cycles in the twenty-first century.
Argument: The Rise of Startup Capitalism
We argue that startup capitalism strives to both harness the disruptive power of startups for creating new markets and benefit big businesses. Policy makers even sometimes publicly speak about their intention to concurrently create cohorts of unicorns while simultaneously injecting innovative ideas and talent into incumbents. This thinking constitutes a best-of-both-worlds logic.
If startup capitalism aligned with Schumpeter’s Mark I creative destruction paradigm, established firms would go into bankruptcy, get acquired, or shift to other industries. In the Mark I understanding, it is the fear of new entrants posing an existential threat, not only their actual activities, that is essential to boosting big companies’ innovative prowess. But the startup policies closest to the Mark I ideal aim to enable startups competing in emerging technologies, not to incentivize startups that are challenging incumbents.
This existential threat is certainly absent in the Korean Center for Creative Economy and Innovation (CCEI) example we shared at the beginning of this book. Each CCEI has a chaebol designated as the core corporate partner (Klingler-Vidra and Pacheco Pardo 2020). These centers provide entrepreneurs with consulting services, marketing assistance, prototyping help, and access to investors. In return for their contribution, the chaebol gain access to the startups’ ideas, talent, and technologies. Rather than threatening the destruction of the big businesses, the CCEIs fuse startups together with the chaebol to deliver mutual benefits. This offers innovative DNA to large firms.
Korea’s CCEI and chaebol are not alone in this. Corporate accelerators around the world speak of their efforts to foster “disruptive startups,” as Intel gushed in the September 2023 announcement of their London-based Ignite program (Intel 2023). The startups in the accelerator programs are perceived as customers, operating based on the firm’s services and technologies, be it Intel’s chips or Amazon’s cloud computing. Startups are complementary to big business, not disruptors. This does not align with Schumpeter’s Mark I paradigm in which startups are propelling creative destruction.
Likewise, our observation of startup capitalism closer to the Mark II ideal type does not align with Schumpeter’s conventional understanding of oligopolistic firms as standalone innovation leaders. Instead, we contend that today’s startup capitalism activities closest to Mark II are consistent with open innovation practices, as epitomized by the CCEI example. Startup policy, in this twenty-first-century variety of Mark II, conceives of startups as external resources for established companies to leverage in a bid to advance their competitive positioning. Incumbent firms benefit from open innovation through the injection of new ideas and access to talent and by acquiring new sales distribution channels and direct customers (Klingler-Vidra and Pacheco Pardo 2022). Corporations engage with startups for the purpose of boosting their own innovation (Weiblen and Chesbrough 2015). Patents and temporary secrecy do not serve as “protecting devices” (Schumpeter 1942, 77); in fact, open innovation systems may endow significant information and talent advantages to entrenched businesses and thus act as a barrier to startups’ ability to achieve scale. Thus, we see a variation of Mark II in which oligopolistic firms engage startups to benefit their innovation prowess instead of relying on internal R&D and consortium links with other lead firms.
There are good reasons why startup capitalism often takes this open innovation variety. Direct state largesse for incumbents has become increasingly unsalable to a public that scrutinizes government spending ever more closely and worries about dominant firms’ shaping of their daily lives. Examples of the prevalence of public sentiment against big (tech) business include a June 2023 New York Times essay whose title begins “Big Tech Is Bad” (Acemoglu and Johnson 2023). Jonathan Tepper (2023) asserts that the Myth of Capitalism in the United States is propagated by the oligopolistic firms that benefit from their dominant positions in their respective markets. This sentiment aligns with research that recognizes that “vested incumbents are incentivized to do what they can to slow down the process of creative destruction” (Baslandze 2023, 559). The Economist (2023b) covered the June 2023 World Bank arbitration ruling that describes the Korean state’s meddling in a Samsung merger as “cozy relations between government and business” and the chaebol as “too close for comfort.” There are challenges to this prevailing negative framing of big business, however. Notably, Robert Atkinson and Michael Lind’s (2018) best-selling book, Big Is Beautiful, pleads for a reassessment of the (positive) role of large firms as employers and innovators.
Policy Implications
Should the aim of public policy vis-à-vis startups today be one of instigating unfettered capitalism, with the expectations that creative destruction will propagate society’s technological advances? Or should big businesses build protective moats—including the accelerators and corporate venture capital funds that underpin their open innovation—to boost society’s innovation prowess and corresponding improvements in quality of life? Can startup policies that enable the continuance of incumbent dominance help explain declining business dynamism trends?
In contrast to Philippe Aghion and Peter Howitt’s (1992) modeling of creative destruction in endogenous growth theory, we contend that startup capitalism operates according to variations on conventional understandings of both modes. On the one hand, policies closer to Mark I aim to foster new cohorts of unicorns in emerging industries. On the other hand, startup policies engage big businesses in part to enshrine their innovation capacity. What is more, some startup policies strive to do both simultaneously—create unicorns and boost the competitive position of established companies. This conflates—and, we argue, undermines—the distinct logics of Mark I and II patterns of innovation. According to Schumpeter, both are superior at a given place or point in time. For instance, Franco Malerba and Luigi Orsenigo (1996) argued that alignment with Mark I or II is technology specific.
In the context of startups as an engine for incumbent-centered open innovation, there are several questions. For one, the ontology of open innovation is one of firm-level or microeconomic performance. Open innovation, according to Henry Chesbrough (2003), offers an updated counter to the five forces that Michael Porter and other management scholars depicted as essential to firms’ competitive strategy. Startup policy that strives to benefit big businesses should be clear that its premise is that open innovation is optimal for the economy, not only for the big corporations.
There are several reasons why the open innovation variety of Mark II could be perceived as superior. System benefits from corporate venture capital, for example, show that value is created for the corporation as well as the new venture firms (Bugl et al. 2022). Research shows that as investors in startups, incumbent firms offer access to their networks, market knowledge, and sales and distribution channels (Alvarez-Garrido and Dushnitsky 2016), which can help startups grow. In addition, startup policy acknowledges that large firms can offer an important means of exit by trade sale or acquisition for many startups. These win-win scenarios likely motivate the pursuit of startup policy enabling open innovation systems.
Yet, it is essential to also distill the potential risks involved in bringing big businesses into startup policy. For one, the ultimate winner of such startup backing may be the corporations rather than the wider economy or society. Incumbents engage in open innovation systems for both inbound and outbound purposes, with startups serving in a range of capacities, from external R&D provision to sales opportunities (Dahlander and Gann 2010). In addition, the response to open innovation startup policies may be tepid, as startups may limit how much they engage with large firms due to concerns that their efforts and ideas will be appropriated. Incumbents may “try to win the market with productive strategies” (such as accelerators) in which they integrate startup ideas and technologies or “rely on nonproductive strategies” that enable them to maintain their market position through patent portfolios and anticompetitive acquisitions (Baslandze 2023, 558). Corporations may also leverage their political connections and power, employing “different forms of lobbying” and thus integrating themselves as essential partners and benefactors of startup-led innovation (Bombardini et al. 2023, 538).
Many also worry that high-growth startups and burgeoning unicorns fall victim to the national champions that can either copy their technologies and mass produce them more cheaply or use their market dominance to prevent the distribution of new products by smaller firms. In the United States, big companies, including big tech firms, have been accused of exploiting startups’ ideas, forcing early acquisitions, and applying predatory pressures that constitute kill zones (Economist 2018a). This concern has been realized in East Asia. In Korea, for example, some analysts believe that, “immensely wealthy and historically backed by government support, the chaebol have a tradition of establishing their own affiliates to compete and undercut startups instead of acquiring the new technology through M&A [mergers and acquisitions]” (Harris 2019). As an example of the alleged downside to open innovation, there are “delivery wars” in Korea, as the chaebol are looking to “snatch market share from the delivery startups that cultivated the industry” (J. Kim 2020). Some analysts worry that they wield so much power that they can acquire teams—or whole businesses—on the cheap, allowing them to hoover up talent, ideas, and products without having to hire or build such capacity in-house. While this characterization of chaebol-startup relations is contested, policy makers’ intention is certainly not to exacerbate this predatory potential.
It is not only startups that risk negative outcomes from open innovation activities. Research has shown that incumbents may fail to capture value from engagement—across accelerators, investments, and more—due to ineffective integration, IP management, and knowledge spillovers (Dabic et al. 2023). There is also the risk of open innovation engagement serving what Steve Blank (2019) calls “innovation theater” rather than big businesses conducting substantive innovation.
From a societal perspective, there are concerns about the equity of gains and societal welfare resulting from variations of startup capitalism. Startups are high-risk, high-failure businesses that are—if dynamism and disruption forces exist—unlikely to be in existence long enough to provide long-term employment and social benefits (Klingler-Vidra and Pacheco Pardo 2019). The firms targeted by startup policy are epitomized by VC-backed companies. These companies raise early-stage equity investments due to their potential to achieve remarkable scale, but they often face binary fail-or-succeed outcomes. A sobering statistic comes from a study of two thousand VC-backed companies in the United States between 2004 and 2010 that had each received at least US$1 million in funding; Shikhar Ghosh from Harvard Business School found that 75 percent failed completely, never returning cash to their investors (Gage 2012). Similarly, Bob Zider (1998) found a high potential for failure, with only one out of ten companies in a VC portfolio likely to succeed and half failing. These failure rates hardly instill confidence that high-growth entrepreneurship can provide steady jobs, let alone employment for life, as the chaebol and keiretsu offered in the post–Korean War and post–World War II eras, respectively. In addition, job creation through tech-intensive startups that pursue scale by automating human tasks—thus allegedly reducing the need for human labor—has a fundamental tension with the provision of employment opportunities.
As Schumpeter (1942, 119) acknowledged in his exploration of innovation paradigms, access to high-performing entrepreneurship is not available equally to all members of society. Certain demographics, based on disability, ethnicity, gender, and socioeconomic attributes, may struggle to participate (Zehavi and Breznitz 2017; Klingler-Vidra and Liu 2020). What is more, technology-fueled automation can cause certain groups to suffer more acutely from job losses. Automation also has the potential to deliver greater returns to the capital owners than labor, benefiting from lower labor costs and higher productivity on account of scalable technologies and as a result accentuating inequality (Klingler-Vidra et al. 2022). The gains reaped from high-growth startups are amassed by entrepreneurial founders who may or may not make investments that benefit local communities or their wider society. And on top of that, companies can domicile in tax-efficient locales and operate online, enabling them to pay little tax and hire few (local) employees.
In this way, startup capitalism’s focus on job creation is wrought with challenges and contradictions. Policy that encourages startups as job creators will necessarily transfer more risk onto citizens than policy that favors large firms providing permanent employment. The financial reward of startup employment may be greater for the individual than a steady salary, but the risk of losing one’s job as well as social benefits is also substantial (Acs et al. 2016).
There is also the question of which of society’s challenges are being targeted by startups. In Big Is Beautiful, Atkinson and Lind (2018) make the Mark II argument that large, well-resourced companies are more likely to take on big societal challenges and thus invest substantial resources that could drive widespread impact. According to this rationale, oligopolies, with their significant balance sheets and workforces, can best take on and advance on technologies that could address society’s more pressing problems. As a contemporary example, in a May 2016 TED talk, Astro Teller, the head of Google X—which Alphabet (Google’s parent company) describes as a moonshot factory—outlines the team’s approach to solving complex challenges that affect all countries, such as the need for sustainable food and transportation.7 The X team was working on vertical farming as a way of producing (more) food without needing more agricultural land and on a lighter-than-air, variable-buoyancy ship. X, as Teller explained, invested many man-hours and tens of millions of dollars to develop and test prototypes before deciding how, or whether, to continue. The X example underscores Schumpeter’s contention that it is large firms (in his day, US Steel, for instance) that can take on these major challenges.
In the East Asian developmental state context, national champion firms often worked in tune with societal needs and at the behest of government bureaucrats. State-society priorities, in the (Japanese and Korean) developmental state’s heyday, were channeled into consortia initiatives, with the state setting the agenda and funding the technologies that were to be developed. In a decentralized system, in which each entrepreneur works on different opportunities, the ability of startups to serve the social purpose of wider society may be muted. With that said, the state’s use of tax incentives and direct funding for startups can instigate activity in line with state priorities. China’s startup competitions, for instance, offer a good example, identifying certain societal issues and emerging technologies that contending entrants should focus on (China Innovation and Entrepreneurship Competition 2020).
Collectively, startup capitalism entails trade-offs and challenges. Each variety aligns better with different aims and at distinct points in technological cycles. Policy makers are now pursuing a hybrid approach in which they strive for both unicorn births and big business successes. This combination may be optimal in the twenty-first century. If Schumpeterian logic does hold, though, then this hybrid approach would not deliver optimal results, as it does not wield the power of either paradigm. Future innovation performance of the region will reveal whether hybridity is producing results or is inadvertently undermining the power that either mode could achieve.
Research Design
In our analytical framework, we conceive of startup capitalism in terms of degrees of continuity and change from antecedent settings. We draw on theories of gradual institutional change (see Mahoney and Thelen 2010; Acemoglu et al. 2021). Our framework delineates five institutional components that constitute our synthesis of the developmental state and varieties of capitalism frameworks. To analytically account for the continuity and change underpinning the countries’ evolving approaches, we explore: (1) the size of firms expected to drive this (capability at the technological frontier); (2) the employment market’s character, as a continuum of permanent employment to more fluid; (3) sources of finance, spanning (main) banks distributing lines of credit to capital markets that issue equity; (4) the nature of the innovation sought, from catch-up toward the technological frontier; and (5) the primary social purpose underpinning policy efforts.
The book draws on a novel assessment of startup policies and the media coverage of those policies across four East Asian countries: Japan, Korea, Taiwan, and China. Policy information was collected in English as well as in Mandarin, Japanese, and Korean; when we translated materials into English, we strove for consistency with the language used by the government (in that era). In total, we analyzed startup policies employed from the beginning of each country’s developmental state era (1948 for Japan, 1961 for Korea, 1949 for Taiwan, and 1979 for China) through 2023.
We study startup policy within the wider rubric of science, technology, and innovation (STI) policies. Broadly speaking, innovation policies are those that strive to advance “new combinations,” as Schumpeter (1934) would say. There are numerous classifications of STI policies according to the intention and the instruments used (e.g., see Nee and Opper 2012; Edler and Fagerberg 2017; Schot and Steinmueller 2018). State-of-the-art classifications of innovation policy employ a tripartite classification according to the policies’ intentionality as either “mission-oriented” or transformative; “invention-oriented”; or national innovation system (NIS)–oriented (Edler and Fagerberg 2017; Schot and Steinmueller 2018).8
As a specific variety of NIS policy, the aims of startup policy are often defined as striving to advance the quantity and quality of high-technology oriented entrepreneurial ecosystem actors, such as VCs and accelerators, as well as the informal or intangible factors, such as a risk-taking and creative culture (Engel 2014; Autio et al. 2016; Klingler-Vidra and Wade 2020). The focus on startups as a target of NIS policy stems from the truism that new firms expand “the technology frontier,” a crucial aim of innovation policy (IMF 2016, 39).
Existing research on startup policy has conceptualized the policy aims and forms in several ways. The first is in terms of different life-cycle stages (e.g., antecedent, founding, behavior, and outputs) and the interventions needed at those different points (Audretsch et al. 2020). A second way of categorizing startup policies is by delineating the types of entrepreneurship targeted, delineating efforts focused on innovative new entrants from the broader category of entrepreneurship policy enabling new, small firms (Acs et al. 2016). A third way is based on the nature of the policy instruments used—such as funding, tax incentives, or regulatory changes—as a means of boosting startup activity (Pacheco Pardo and Klingler-Vidra 2019). In addition, scholars have aligned startup promotion expectations with levels of economic development, contending that creative destruction is more likely to be present and more impactful as economies develop. The argument is that in developing contexts, innovation initially takes the form of adopting technologies from abroad (Acemoglu et al. 2006; Peters and Zilibotti 2023).
In this book, we coded startup policies according to the instruments used rather than other taxonomies.9 We did this to be comprehensive in our understanding of a startup, distinguishing it from the broader category of small- and medium-size enterprises (SME) but not focusing on any one point in a startup’s life cycle or on specific innovation or employment strategies. The eight startup-centric policy types are as follows: (1) Funding; (2) Taxation; (3) Regulation; (4) Clusters, Networks, Institutes; (5) Attracting Talent and Investment; (6) Stock Market Access; (7) Technology Infrastructure and Government Procurement; and (8) Education and Training. Table 1.1 summarizes each policy type.
To identify the startup policies enacted in each country, we used two means. First, we searched for startup policies on governmental websites and search engines. We canvassed government sources, international policy databases, and web sources to identify policies implemented across the period in each country. We began with government sources such as Japan’s Ministry of Economy, Trade and Industry (METI) website and then covered international policy databases, namely the Global Entrepreneurship Network Atlas as well as the Startup Genome’s Global Startup Ecosystems Reports. Our next step was to cover wider web sources, which we did by conducting a Google search consisting of the country name (e.g., Japan) and startup policy language (see Klingler-Vidra and Chalmers 2023). Second, we canvassed the body of developmental state, comparative capitalism, and Asian business systems scholarship to ensure we captured the most comprehensive picture of startup policies across the eight instrument types. This second step helped us check for discrepancies and missing policies.
Specific policy tools | |
|---|---|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Note: Details on each policy type are sketched in terms of their alignment with conventional understandings of Mark I and II modes in chapter 1. Further explanations of policy instruments can be found in Klingler-Vidra 2014, Pacheco Pardo and Klingler-Vidra 2019.
Once we collated the startup policies for each country, we analyzed them in terms of the size of firms targeted as the participants and intended beneficiaries, the equity/debt nature of finance, lifetime or fluid employment strategies, the type of innovation targeted (radical or incremental), and the underpinning social purpose. To assess the social purpose, we examined the objectives as laid out in the policy documents themselves. We then conducted a Google search of each policy to find media communications around its launch. We coded media statements made by senior policy makers and politicians with respect to why they were pursuing the policy and what aims they hoped to achieve with the initiative. Finally, we complemented the policy analyses with semistructured interviews with national, provincial, and local policy makers as well as innovation system actors, including startup founders and venture capitalists, conducted between 2016 and 2023.
Collectively, we present a political economy account of startup capitalism across three archetypal developmental states—Japan, Korea, and Taiwan—and the juggernaut of China, which combines state-owned enterprises, established private firms, and entrepreneurs quickly advancing to the world’s technological frontier. Invariably, despite the extensive research published on East Asian industrial policies, to date, there has been insufficient investigation of each state’s startup promotion efforts. We hope to address this gap in a way that establishes startup policy as an industrial policy arena within comparative capitalism, developmental state, and neo-Schumpeterian literature.
Plan for the Book
The book is organized as follows. Chapter 1 advances the analytical framework. It conceptualizes the institutional arenas that are fundamental to studying startup capitalism. The five components are: (1) the size of firms (large or small) targeted to drive innovation prowess; (2) the fluidity of labor markets; (3) financing for innovation, as credit- or equity-based; (4) the type of innovation as either catch-up or striving to compete at the technological frontier; and (5) the social purpose. We examine how these five institutional arenas have changed in step with startup policies in terms of a continuum framed by Mark I and Mark II ideal types. Chapters 2 through 5 offer analytical narratives for each country by providing a brief distilling of the antecedent period through to a more extended analysis of the contemporary period. Each chapter closes with a discussion of the extent to which continuity or change is observed and delineates the case’s variety of startup capitalism in terms of the Mark I and II modes.
We begin the case analyses by focusing on Japan in chapter 2, the quintessential developmental state that serves as a least likely case for policy makers pursuing Silicon Valley style, given the widespread depiction of the state as coordinating large-firm consortia and working with main banks to offer the keiretsu steady credit lines (Jackson 2003; Pempel 1998). We find that the Japanese developmental state apparatus has, since the onset of the Heisei recession that began in 1990, supported startups as a means of fueling open innovation systems. This has been in line with an open innovation variety of a Mark II pattern, as the keiretsu and main banks are key partners, investors, and beneficiaries of Japan’s startup capitalism. In fact, even high-profile startup programs, such as the J-Startup Initiative and Startup Ecosystem Consortium, involve the keiretsu and main banks in selecting startups for participation. Then–Prime Minister Fumio Kishida’s 2022 “Startup Development Five-Year Plan” explains that “large, existing companies . . . can stay in business if they engage in open innovation” (Cabinet Secretariat 2022, 1). The open innovation aim is evident in the Japanese government’s articulation of social purposes of encouraging startups, which we see as threefold: first, to provide high-quality jobs and thus increase creative talent for the keiretsu; second, to boost technological capacity of the established firms; and third, to collectively drive technological prowess and economic growth.
Chapter 3, Korea, also offers a state-led case in which development was fostered in partnership with national champions: the chaebol. Our research reveals how high-profile, startup-centric initiatives have been pursued, such as President Park Geun-hye’s creation of the Creative Economy Action Plan and former President Moon Jae-in’s establishment of the Ministry of SMEs and Startups (MSS) in July 2017. Both efforts have seen funding, tax incentives, and a variety of government-led programs aimed at startups as well as startup-chaebol partnerships. We reveal how startups are often perceived as advancing large firms’ technological capabilities, consistent with a Mark II open innovation variety of startup capitalism. While startups are publicized as key targets of innovation policy in Korea, the chaebol are often essential partners and beneficiaries. Startup policies strive to provide innovative ideas and talent to the chaebol while also aiming to diversify the economy away from the giant firms and encourage an entrepreneurial mindset.
Chapter 4 examines the Taiwan case, which begins from a different origin in many ways. In political economy research, Taiwan stands out for its early orientation toward high-growth small firms. From the 1970s onward, the Taiwanese state increased its high-growth startup initiatives such as Hsinchu Park. Our research reveals that Taiwan has broadly remained consistent with a Mark I approach in the sense that it is focused on widening pools of entrepreneurship in emerging technologies. Startup policies do not conceive of the fledgling firms as resources for large firms. Dominant companies, like TSMC, have pursued their own open innovation strategies, though not in collaboration with the government, let alone as part of the startup policy mix. However, Taiwan’s brand of startup capitalism does not align with a conventional understanding of Mark I, as startups are not creative destruction engines; instead of disruption, they are only construed as advancing emerging technologies.
Chapter 5 focuses on China as the last case study, given the later timing of its developmental advance and reflecting its position as distinct from the “classic” developmental states. The chapter establishes how China, from 1979 to 2000, exhibited antecedents of its assistance for startups as an alternative growth and employment mechanism. China’s political economy has often been depicted as antithetical to the neoliberal United States, with the state marshaling resources into national champions operating in select sectors and critical technologies. However, we find that China’s brand of startup capitalism began with and continues to be characterized as having noticeable contradictions consistent with versions of both Mark I and II patterns of innovation. In some ways, only Taiwan exhibits more characteristics that align most closely with Mark I in the sense that startups are not construed as resources for big businesses. Also, unlike policy makers in the other cases, those in China have demonstrated some willingness to threaten the oligopolistic positioning of giants such as Alibaba (Deng 2022).10 In addition, the Ministry of Industry and Information Technology (MIIT), National Development and Reform Commission (NDRC), and other Chinese policy makers have mostly steered clear of closely involving incumbents, which is a signature of the open innovation–style Mark II tactics taken in Japan and Korea. Overall, the Chinese case experienced moves toward Mark I but remains a unique mix of the two modes.
The conclusion chapter first compares the four countries. Across the cases, there are movements toward the Mark I end of the spectrum but also degrees of path dependence. The earliest two startup capitalism cases—Japan and Korea—have mostly continued their Mark II developmental state approach that centers on big business largesse as a means of driving economic competitiveness and employment. The third case, Taiwan, in contrast, has mostly retained its version of a Mark I model in which policies aim to encourage a wide pool of startups in emerging technologies. Finally, we distill how China’s evolution is again different from the others, as it combines elements of both modes. In China, startups are positioned as a means of competing at the frontier in critical technologies and as a provider of (mass) employment. While there is evidence of policy to support incumbents in critical technology areas, such as around the Semiconductor Manufacturing International Corporation (SMIC), its national champion technology firm, in other sectors, the Chinese case also resembles a pared-back version of a Mark I understanding of startup backing—widening entrepreneurial pools but not necessarily disrupting markets.
The conclusion then brings these insights back to the developmental state dead-or-alive debate. Developmental states were defined according to their dedication to upgrading industrial capacity, fostering firm relations, and providing necessary access to finance. In this respect, we argue that developmental states persist, as startup capitalism can take the form of an open innovation paradigm, with startups enabling big businesses’ innovation capabilities. Developmental states, even when using the language of Silicon Valley ambitions, can still be essential enablers for established firms, which continue to serve a crucial socioeconomic role. Thus, startup capitalism does not constitute the death of the developmental state but an open innovation version of Mark II.
We show that startup capitalism has varieties and, as such, can approximate both opposing ideal types of VoC: LME and CME. The LME version is more consistent with creative destructive logic, in which new entrants are viewed as crucial growth engines. However, we argue that comparative capitalism and entrepreneurial state research often assume that this Mark I variant is the rationale and aim, and so evidence of it has shown an inevitable convergence on a Silicon Valley–styled LME. We contend that startup capitalism can also comprise open innovation approaches in which public policy strives to benefit both incumbents and startups by fostering their interactions, which necessarily subverts creative destruction. This Mark II variant aligns with CME logics in terms of close large-firm relations and relatively permanent or inflexible labor markets. Thus, the persistence of a CME-styled variety of startup capitalism suggests that rather than convergence on the Anglo-American LME occurring, different models persist.
Finally, we close the book by returning to our opening question: Are startup policies clearly designed to foster either creative destruction or oligopolistic leadership of innovation? We argue that startup capitalism does not align with either ideal type and that the pursuit of hybrid logics could help explain declining business dynamism in the twenty-first century. We reveal the implications of governments engaging startups as resources, not (only) disruptors. We underscore why policies often posit startups are enabling the growth and survival of East Asia’s large firms that must compete with American and European rivals. Finally, we weigh in on how the open innovation version of startup capitalism challenges the Silicon Valley approach and may be allowing East Asia to be at the forefront of the global technology competition.