How Japan’s Startup Ecosystem Grew Alongside Its Large Firms

Startup Japan

The Need for a Good Framing to Discuss Startup Ecosystems

Many discussions about startup ecosystems run into difficulty when participants turn out to be using very different mental models of what they are and how they work, leading to conflicting policy prescriptions and politically easy but ineffectual appeals to voter constituencies. As recent work in social psychology shows, how issues are framed powerfully shapes people’s mental models of cause and effect. Stories and narratives are commonly shared frames about how things came to be and where they are headed—and the existing narratives around Japan's startup ecosystem are often misleading.

This piece contributes a useful framework to analyze and evaluate startup ecosystems—a necessary first step toward meaningfully discussing Japan’s startup ecosystem.

• What are the main components of startup ecosystems?
• How are the components interrelated?
• How can each component develop self-sustaining positive feedback loops of development?
• What is the story of Japan’s startup ecosystem? How did it develop, and what is its trajectory?

It is tempting to take a snapshot comparing Japan with Silicon Valley, or perhaps startup-heavy Israel or Singapore, and immediately conclude that Japan’s startup ecosystem is woefully insufficient. However, Japan’s startup ecosystem does not exist in isolation—it is deeply influenced by Japan’s primarily large-firm-led economy and by global technological trajectories. The prominence of large firms and their historical role in Japan shape both the constraints that Japan’s startup ecosystem has faced and the opportunities now available. These are not the same constraints and opportunities that exist in the United States, Israel, Singapore, or even Germany or France, so a simple conception of where Japan ought to be given its economic size may not be the best framing.

Kenji Kushida

Kenji E. Kushida is a senior fellow for Japan studies in Carnegie’s Asia Program, directing research on Japan, including a new Japan-Silicon Valley Innovation Initiative at Carnegie.
More >

Japan’s large-firm-centered postwar economic model was highly successful for several decades in producing economic growth, from the 1960s through 1990s. Then, when a series of macroeconomic challenges—an asset bubble bursting and technology and software transforming the nature of high-value-added competition—coincided with other Asian countries beginning to develop rapidly into competitors, Japan was slow to adapt. Large firms are inherently slow to adapt and can be disrupted by new, fast-moving technologies and competitive realities, but the broader Japanese political economy with its highly established structures of business, employment, healthcare, and firm-centered social welfare and pension systems is predicated on large firms driving the economy. Japan’s enduring large firms did in fact—arguably even at the expense of growth—prevent the degree of social disruption and income inequality growth that the United States has seen in the past couple of decades.

Given this background, the question to ask is not “Why isn’t Japan’s startup ecosystem bigger?” Instead, a more appropriate starting point might be “How was Japan’s startup ecosystem able to overcome the major disadvantages it faced to grow to the degree it has?” After all, Japan’s postwar model included bank-led finance, long lead times for initial public offerings (IPOs), lifetime employment for top-tier human capital at large firms, a large-firm-led economy organized around corporate groups such as Mitsubishi or Sumitomo, a long postwar history of arms-length ties between universities and industry as a reaction to prewar and wartime co-option, and low social legitimacy of startups in a society focused on attaining stability—all changing slowly with a backdrop of little and slow economic growth. This combined to create an environment very inhospitable to startups.

Yet now a growing number of startups have undergone IPOs and reached the Tokyo Stock Exchange market once occupied mostly by incumbent large firms. The venture capital (VC) industry is maturing and driven by independent venture capitalists rather than corporate investment arms. Top-tier university graduates are founding startups and drawing elite human capital from large firms, the government, and universities. Large firms are increasingly partnering with startups to do things the large firms cannot easily accomplish by themselves. The number of university spinout startups is increasing rapidly, with a few growing into some of Japan’s largest startups. And the social legitimacy of the startup ecosystem is higher than at any point in at least the past half century; Japan’s major business federation, the University of Tokyo’s president, and the government’s economic strategy plan have all explicitly encouraged growth of the startup ecosystem. Now the question is how to take the next steps, given where Japan’s political economy started.

To understand how Japan’s economy changes over time, it is important to differentiate the traditional, new, and hybrid parts that coexist—observers who look at only the traditional areas may conclude that very little has changed, while those more familiar with the new areas see rapid and extensive change. For example, even in the same office building, traditional large firms populated with suit-clad workers in lifetime employment arrangements with strong seniority-based hierarchical cultures can coexist with startups or foreign firms whose employees are dressed far more casually in offices that may have hammocks and an abundance of green plants, with far more flat organizations and a low expectation of employees staying for more than a few years before moving on. Many observers of Japan have long-established networks with traditional parts of the economy and tend to underestimate or overlook the existence and speed of growth of the new portions.

Startup Ecosystems, a Framework

Startup ecosystems have several components. These components are interlocking, in that each depends on others to function.

Table 1: Components Underlying Startup Ecosystems

Component
Silicon Valley Model

1) Finance
Venture capital

2) Labor market
Dynamic, highly educated, top talent from around the world

3) University-government-industry ties
Multifaceted, conducive to forming and growing new companies

4) Industry organization
Symbiosis of large firms and startups, intense competition

5) Social norms surrounding entrepreneurship
Celebration of entrepreneurs and startup ecosystem

6) Professional services supporting startup ecosystem
Professional services supporting startup ecosystem, numerous venues for startups and collaboration

Source: Kenji Kushida, “Departing from Silicon Valley: Japan’s New Startup Ecosystem,” in Reinventing Japan: New Directions in Global Leadership, ed. Martin Fackler and Yoichi Funabashi (Santa Barbara: ABC-CLIO, 2018), 81–102.

The components here are derived from the “Silicon Valley model,” which emerged and consolidated in the 1980s to take center stage in the U.S. economy in the 1990s and beyond.

These components depend on one another—they are complementary. For example, the venture capital system works best when investors can draw from a flexible labor market to introduce people they already know to be a good fit for the companies they fund; think of a new startup whose CTO was introduced by the VC because they had had previous successful experiences at other companies the VC had invested in. Likewise, successful venture capitalists get most of their financial returns when companies they invest in undergo IPOs or are acquired by large companies. Therefore, VC investors do best when there is robust acquisition activity toward startups by large firms.

Each of the components also experiences virtuous cycles, or positive feedback loops—the more the component develops, the more its development accelerates. For example, as a country’s VC industry begins to succeed and yield high returns, more startups can get funded. This provides more opportunities for entrepreneurs who can successfully get their startups funded, leading to the possibility of more successful startups, yielding even higher returns for the VCs. As VCs start delivering higher returns on their investments, more large institutional investors and corporations are attracted to invest in VCs, giving VCs more capital to fund startups—thereby attracting more entrepreneurs, and so on. Likewise, if collaboration between industry and universities is well established with multifaceted linkages (such as researchers taking on advising and consulting positions at companies, technologies from their labs spun out into startups, researchers moving in both directions between industry and university labs, and a variety of industry affiliate programs giving researchers a sense of what theoretical breakthroughs can lead to in both academic and industrial value), then subsequent industry-university ties become easier because both sides have deep experience to draw upon.

The same logic of complementarities between components and positive feedback loops that reinforce each component once established also applies to all other components and their combinations.

Why the Silicon Valley Model?

The model used here is derived from Silicon Valley, with the recognition that Silicon Valley is a historically specific, singular development that is difficult or unwise to attempt replicating. That being said, since it is by far the most successful startup ecosystem, the components that make it work are worth examining closely.

Silicon Valley has been criticized over the past few years for a variety of reasons, such as the “tech giants” Apple, Google, and Meta becoming too large and potentially monopolistic; the unintended consequences of a focus on optimization leading to grave social challenges; serious gender and equity gaps; and many companies with problematic working cultures.

However, as an economic model, Silicon Valley’s importance in driving American economic growth and innovation over the past few decades cannot be overstated. Centered around venture capital to fund high-growth startups, with dense interpersonal networks of expertise and experience in a localized area near top VC firms, in an environment of flexible labor deployment of high-skilled workers from around the world, following a historical background of government-funded industry-university collaborations, and bolstered by a wide range of support businesses, the “Silicon Valley model” has become a core component of the U.S. economy. Geographically, the Silicon Valley model goes beyond the valley to cover much of the San Francisco Bay Area, since many of the VC-financed high-growth startups originate in San Francisco.

As of 2020, VC-backed firms accounted for over 40 percent of U.S. market capitalization and over 60 percent of research and development (R&D) spending by all U.S. public companies. If narrowed to only U.S. public companies founded within the past fifty years, VC-backed firms composed 50 percent of all public companies, 75 percent of total market capitalization, and over 92 percent of R&D spending and patent value.

Venture capital is overwhelmingly concentrated in California (see table 2), and the top VC firms continue to reside in Silicon Valley. The gap between VC-backed firms in the United States and in other G7 countries is stark (see figure 2).

Table 2: Top Five U.S. States by Assets Under Management in 2021

State
Assets Under Management (in Billions of U.S. Dollars)

California
$549.9

New York
$149.4

Massachusetts
$120.2

Illinois
$28.2

Washington
$19.0

Source: “NVCA 2022 Yearbook,” National Venture Capital Association, accessed June 14, 2022, https://nvca.org/wp-content/uploads/2022/03/NVCA-2022-Yearbook-Final.pdf.

Japan’s Historical Economic Model Hindered Startup Ecosystem Development

Japan’s postwar economic model was almost the antithesis of a startup ecosystem. The dominant model that took root in the 1960s and burst into global industries in the 1970s and 1980s was characterized by its distinctive features: main bank finance, long-term employment, strategic government-industry collaborations, and corporate groupings anchored in cross-shareholding known as “keiretsu” to which a majority of Japanese major companies and their suppliers belonged.

Table 3: Comparing the Silicon Valley and Traditional Japanese Models

Institution
Silicon Valley Model
Traditional Japanese Model

1) Finance
Venture capital
Main bank, later financial markets

2) Labor Market
Dynamic, highly educated, top talent from around the world
Long-term employment, hired out of university

3) University-government-industry ties
Multifaceted, conducive to forming and growing new companies
Strategic government-industry collaboration without substantial university involvement

4) Industry organization
Symbiosis of large firms and startups, intense competition
Vertical and horizontal "keiretsu"

5) Social norms surrounding entrepreneurship
Celebration of entrepreneurs and startup ecosystem
Large firms and government as most prestigious, low social status of entrepreneurs and startups

6) Professional services supporting startup ecosystem
Professional services supporting startup ecosystem, numerous venues for startups and collaboration
Limited professional services expertise, little media and event exposure for startups

Source: adapted from Kushida, “Departing from Silicon Valley.”

In-depth discussion of each component will show how Japan’s traditional postwar model was an impediment to the development of Japan’s startup ecosystem until a series of changes occurred in the late 1990s and the 2000s.

Japan’s Current Startup Ecosystem Is Maturing Alongside Its Traditional Economic Core

An important piece of framing is that Japan’s startup ecosystem is not replacing its traditional core of large firms. Instead, the ecosystem, which grew in parallel to and largely outside the traditional core, now feeds off the core more effectively, and each of the components is experiencing positive feedback loops.

• Venture Capital: Japan’s VC industry matured with the rise of independent investors leading the sector rather than the industry becoming dominated by VC arms of large corporations. The development of small-market-capitalization markets in the late 1990s made it possible for startups to IPO quickly, enabling VC to realize investment returns and function effectively. As VC returns increased, more people with relevant experience were drawn to the industry, increasing the opportunities for startups to get funded and VCs to realize further returns.
• Labor Market: The labor market in the startup ecosystem has become fluid and dynamic, with elite graduates leaving prestigious jobs at large firms and elite government ministries to join startups. The decreasing attractiveness and stability of traditional large corporations and government positions drove the rise of startups as legitimate and attractive destinations for elite talent. Japan’s growing information technology (IT) sector and the acceleration of foreign firms taking center stage in finance also led to pockets of labor market fluidity, which introduced greater flexibility into the labor market—benefiting the startup ecosystem. Greater numbers of the elite workforce joining the startup ecosystem in turn increased the attractiveness of the startup ecosystem as a place to work for top-level and elite workers, who were drawn in by their former colleagues and schoolmates, other interpersonal connections, and overall social legitimacy.
• University-government-industry ties: University-government-industry ties have transformed with regulatory reforms at national universities, enabling them to rapidly develop close ties with industry in general, with a particular focus on feeding the startup ecosystem. The government actively supports spinning out startups from university labs and research, and deepening collaboration between universities, industry, and the government. Some notable successes and learning on both sides have led to further collaborations and spinouts. Overall, through a myriad of programs, the government has contributed to the startup ecosystem’s social legitimacy.
• Industrial Organization: The way large firms relate to startups is shifting; large firms are collaborating with startups more than ever, with unprecedented efforts to create new value by working with and even purchasing startups. This was a sharp departure from previous practices relying on in-house R&D and treating small firms only as suppliers. Waves of collaborations by some major established firms led to further opportunities for subsequent firms, and better understanding of the startup ecosystem by large corporations raises the likelihood of success in subsequent collaborations.
• Social Legitimacy: The social norms surrounding the startup ecosystem and entrepreneurship transformed significantly to the point that the startup ecosystem enjoyed real social legitimacy and widespread normative support. The startup ecosystem often made front page business news, with startup pitch contests, government-sponsored awards, and other publications and frequent showcase events.
• Ecosystem Services Businesses: Finally, professional services supporting the ecosystem grew substantially, with law firms adding practice areas specializing in startups and some notable accounting firms even actively supporting the startup ecosystem by holding pitch contests and taking equity stakes in startups. Large firms as well as independent firms, even from Silicon Valley itself, rushed to set up startup accelerators, incubators, and some early-stage investment funds in Japan. Specialized startup ecosystem database firms and media that focused on the startup ecosystem thrived in helping to match large firms with startups. Once these support services and the myriad of opportunities were established, subsequent startups faced an environment in which it was easier to succeed, further bolstering the business of startup support services.

Table 4: Silicon Valley Ecosystem Characteristics Compared to Japan in the 1990s and Japan in 2018

Silicon Valley startup ecosystem characteristic
Traditional Japanese model impediments to startup ecosystem (through mid-1990s)
Japan from mid-2010s onward: changes that facilitate startup ecosystem

Financial system: venture capital
Bank-centered, traditional financial markets. Small VC industry, mostly corporate and investment via loans
New small-cap financial markets, growing VC industry, rise of independent VC investors

Labor market: fluid, diverse, highly skilled
Long-term employment with seniority ties creating illiquid labor markets. Best and brightest locked into large firms for entire career
Increasing labor mobility, especially in IT sector and foreign firms. Lower prestige and opportunity with large firms. Higher salaries, stock options at startups

University-government-industry ties
Numerous formal regulatory constraints on universities, lack of talent circulation
Active efforts by universities, private venture capital, and government to spin out successful startups with university technology

"Open" innovation with large firms and small firm symbiosis
Closed innovation with large firms conducting in-house R&D and uninterested in business with startups
Firms more interested in open innovation, participation in VC funds, business with startups

Social norms surrounding entrepreneurship
Entrepreneurship seen as low prestige vis-à-vis large firms and government, financially risky
Rising attractiveness of entrepreneurship as large firms enter competitive crises, increases cases of successful startups

Professional services supporting startup ecosystem
Small and inexperienced professional services ecosystem
Law firms and accounting firms setting up startup-focused practice areas to foster and benefit from growing startup ecosystem

Source: adapted from Kushida, “Departing from Silicon Valley.”

By the mid-2010s, the environment underlying and enabling Japan’s startup ecosystem had transformed almost beyond recognition. It far more resembled dynamics seen in Silicon Valley, though at a much, much smaller scale. Yet, since the core of Japan’s large-firm-centered industry remained mostly intact, this transformation was driven by the emergence of multiple new logics in the Japanese political economy.

Each of the components that developed as Japan’s startup ecosystem rapidly matured after the mid-2000s will be covered in the next parts in this series.

End of document