JPRI Working Paper No. 102 (August 2004)
Japan’s Quest for Entrepreneurialism: the Cluster Plan
by Kathryn Ibata-Arens

In the early 2000s, Japan was still failing to recover fully from the economic doldrums. By 1999, its unemployment rates surpassed those of the United States (Porter, Takeuchi et al. 2000). In 2003, new business creation remained paltry. A 2003 study (Global Entrepreneurship Monitor GEM) indicated that Japan is less entrepreneurial -- on a variety of firm- and individual-level measures -- than all of the advanced industrial countries, save Russia.1 Of the world’s 40 major economies, most belonged in one of three entrepreneurship levels (high, moderate, low) based on measures including new firm start-ups, innovative output and the like. Highly entrepreneurial countries included Chile, Korea, and New Zealand. Most countries were moderately innovative, such as Canada, Finland, Singapore, the United Kingdom, and the United States. The least entrepreneurial countries were France, Japan, and Russia. Japan fails at both individual and firm level innovation and entrepreneurship, according to the GEM Report, making it among the least entrepreneurial countries in the world. (Porter 1990; Porter 1998; Porter, Takeuchi et al. 2000). (See Chart 1: Trends in Global Entrepreneurship) (See Chart 2: Entrepreneurial Activity by Country, 2002-2003)

The Ministry of Economy of Trade and Industry (METI) is spearheading the Japanese government’s current efforts to fix its innovative and entrepreneurial problems. In 2001, METI launched its “Cluster Initiative” culminating the following year in a package of policies called the “Cluster Plan.” The “Plan” has become the most ambitious and comprehensive METI strategy since its 1960’s bet on heavy industry (Inoue 2003).2

The intent of the Plan is threefold: to improve productivity, spur innovation, and foster new business creation. The Plan targets 19 clusters3 across 9 major regions in Japan, incorporating 5,000 SMEs, 200 universities, and a variety of support institutions, all coordinated by the national and regional METI bureaus. Having followed METI and MOF policy since the mid-1990s, I sensed that there was a different kind of buzz around this new “Cluster Plan.” (See Chart 3: METI Cluster Plan 2003)

I set out to get the perspective of someone currently in industry who also had an insider’s historical memory of METI in its 1960s heyday. I obtained an interview with Y-san, a recently retired METI bureaucrat and current senior executive of a large Japanese corporation. In the 1960s, Y-san graduated with a law degree from Tokyo University and joined MITI. As is the custom with elite track bureaucrats, he was soon dispatched for his two year (obtain a masters degree from a prestigious American or British university) overseas rotation. Y-san received an MPA from the Kennedy School of Government at Harvard University.

He quickly moved up the ranks at MITI and was seconded to projects in energy, small and medium enterprises (SMEs), national security, R&D, and intellectual property rights. He served as commissioner of the Japanese Patent Office and as director-general of several bureaus, including the International Trade Policy Bureau and the Trade Administration Bureau during the Hashimoto and Obuchi administrations.

Nearing the end of his tenure at MITI, Y-san represented Japan at APEC and WIPO meetings in the mid-1990s and later was a chief architect of the Asian Economic Recovery Plan during the Asian Economic Crisis. Retiring from MITI in 1999, Y-san spent a year at Stanford University. He assumed his post at his current firm in September 2001. I contacted Y-san in 2003.

One July day in 2003, I was whisked up a private elevator in Y-san’s corporate headquarters in Tokyo’s Ginza. Stepping off the elevator into a wide hardwood-paneled hallway, cool and quiet, I was escorted into Y-san’s private conference room. As I sat waiting, natural sunlight poured in through the many windows, and a smartly dressed OL brought me some tea.

Y-san came in at the appointed hour and after a few pleasantries, I began my questions, which centered on his perspective on METI’s ambitions with the new “Cluster Plan.” While Y-san does not doubt the sincerity or the intentions of his former colleagues, he is doubtful that this “new” plan will be a success. Y-san’s critique of the national approach to the Cluster Plan is threefold: its insensitivity to regional history, its inconsistency, and its hubris. First, national bureaucrats tend to be insensitive to (and often ignorant of) the utility of tailoring policy to the historical context of each region. Y-san noted that Japan’s regions have a 2000-year history from which particular kinds of region-specific industries emerged. Within Japanese regions themselves, the indigenous capacity to produce products varies widely. For example, he compared Ota ward (in Tokyo) with regions outside of Tokyo. Ota was developed by national policy fiat, in a previously undeveloped physical space (during postwar Japan’s heavy-industry-based high-growth period). Because of this dependence on the national government Ota has always lacked self-reliance (jiritsu sei). Though this lack of self-initiative may be characteristic of the once numerous firms located in the environs around METI offices -- this is not the nature of industry in most places outside of Tokyo.

This notion brought Y-san to his second critique of current METI policy -- its inconsistency. In the past, the shortest implementation timeline of a given policy was 5 years -- usually it lasted more than 10. Now, METI changes policies every few years. What was once serious and patient is verging on the dilettantish.

Finally, Y-san’s most virulent critique is leveled against what he describes as national hubris. “Public officials have hubris when tackling policy issues, they think that whatever they decide at the national level can be done at the regional level.” He agrees though that the national government should play a supporting role in regional development, particularly through financial and infrastructural measures. Unfortunately, according to Y-san, national level bureaucrats have a control problem. “They often interfere with regions when regional problems should have regional level solutions.” Y-san concludes “to have central government control of the Cluster Plan means that it will certainly fail.” Having already interviewed a number of METI officials, I noted that they seem to be working so hard, and with such sincerity, there must be some possibility of success. I was dispatched by Y-san to METI’s “Cluster Plan Promotion Office” to find out for myself (Y-san Interview July 2003).

I left Y-san’s pristine corporate headquarters and made my way on a hot July day in Tokyo over to METI’s Bekkan (annex). The Japanese describe the retirement of elite ministry bureaucrats into corporations at the end of their long service as bureaucrats as amakudari, or the “descent from heaven.” I couldn’t help but think that I was not going to the “heaven” of one of the most elite ministries in Japan, rather I was descending into the bowels of some Kafkaesque hell. Such was the working environment of a mid-level bureaucrat in Japan’s elite national ministries.

The Bekkan, a dreary, grayish 1950’s style building is where scores of METI offices staffed with cadres of the best and brightest minds of Japan toil away doing the grunt work of METI policy design and implementation. On this day in the Bekkan at METI’s “Cluster Promotion Office” the air conditioning was out -- though the windows remained sealed shut. Dim fluorescent lighting illuminated the steel gray rows of desks, at each huddled a dripping-with-sweat METI functionary. Perhaps Japanese decorum kept these men from even loosening their ties or rolling up their long (no longer crisp) white shirt-sleeves.

I gratefully accepted from a uniformed OL, a mugicha (summer tea), in a small glass with a large chunk of ice. I was there to meet M-san, who immediately launched into the standard “chronology and contents of this project” speech. I dutifully took notes and waited for my opportunity to ask questions (hopefully before my allotted time ran out). I was particularly interested in how this ambitious new policy was designed, which models and scholars were the most relevant, who participated in the process, and so forth. I was also interested in learning how the project would be evaluated as it progressed.

The surprising answer to the first question was: “Michael Porter’s clusters was our inspiration.” M-san knew of no other scholar’s perspectives. Precisely as Y-san had said, M-san described a vast nationwide METI-led framework, albeit with regional variations on the national model. Maybe it was the heat, perhaps the topic, but by the end of the interview our heads were hovering inches above the coffee table, M-san was leaning close and speaking in a whisper: “If we don’t start supporting the smallest of the SMEs -- and not just the established medium sized (chuuken) which are reaping the most benefits from this project -- the Japanese economy will fail.” I asked if he thought then that the “Cluster Plan” would fail. He smiled sardonically and said nothing (M-san Interview 2003).

One example of the “regional variation” of the Plan is the way in which the national METI has gone about implementing it at the regional level. According to T-san of the regional METI Kinki office, their so-called national cluster project is actually a homegrown “Strategy Project.” T-san was adamant in stating, first of all, that his organization started the project in the mid-1990s, long before this new national plan emerged. The local METI Kinki project itself originated out of an earlier study group that looked into the micro foundations of entrepreneurial success in local cities. From this, the regional bureau created a package of policies geared at supporting existing firm-level initiatives, including the Kyoto Shisaku Net.4 T-san was also adamant that their approach was nothing like the current “mega” level cluster plan of the national METI. In fact, T-san reported that a few years ago, national level representatives suddenly “informed us that our project was now a ‘national’ project and would henceforth to be called ‘Cluster Project.’” (T Interview 2003) Perhaps the fever pitch at which the national METI is going about imposing the Cluster Plan throughout the regions is a function of how desperate the situation really is. A look back at the policy environment from which the “Cluster Plan” emerged is informative.5

In the late 1990s as the Japanese recession neared a decade in length, the government tried to implement a series of policies to jump-start innovation and foster new-business creation. Unfortunately, their efforts resulted in bad policies which were poorly implemented. For example, in October 1998, the then still Ministry of International Trade and Industry, MITI, announced a new policy ostensibly to support struggling small businesses. Forty trillion yen in government-backed loans were earmarked for SMEs and intended for investment in new equipment and technology.

Before the ink was dry on the legislative books, however, firms were finding themselves approached by bank representatives, urging them to apply for the funds. Firms noted that these were the same representatives who six to eight months earlier had abruptly cut off funds and called-in existing loans. Bank representatives were reported to have told firms that if they applied for this new program (with the bank’s assistance), they would be certain to get the funds. In return, firms were expected to use these funds to pay back old loans. Since the program was backed by the Ministry of Finance (and administered by MITI), there has been speculation that the so-called program for SMEs was actually a veiled bailout for banks.

The sheer speed (within days of the announcement of the MITI policy) with which banks at the local level all over Japan were able to obtain detailed information and applications for the program is highly suggestive in this regard. At the same time, there are reports of strong-arm practices being used by banks to collect funds of called-in loans. One bank employee, interviewed anonymously, told of a “Collections Manual” that showed how medium to smaller sized firms were specifically targeted. Mistakes in reporting and the like were identified (ochido) and used as excuses for suddenly calling in loans. “Negotiations” would then begin with firms for the return of funds. In actuality, these “negotiations” were unilateral demands by the bank for unrealistic monthly payments. Many firms have reported harassment and even threats of violence by bank operatives.

In November 1998, MITI introduced measures aimed at promoting new business formation. Loan guarantees would be offered to regional credit associations, to be used to assist new start-ups (though funds would be limited to those that already had part of the funds needed to start their businesses). In November 1999, the Japanese government announced an 18-trillion-yen stimulus package, the largest ever. The “new” measures included expanding the amount currently allocated to credit guarantees for small businesses, loans to encourage more start-ups, and the like. The year 2000 brought yet another ineffective and ill-fated stimulus package of similar ilk.

Another policy involved the Industrial Revitalization Law (effective October 1999) (sangyo katsuryoku saisei tokubetsu sochiho). The Law had a number of good components. First, it seemed a genuine attempt at patent reform. For example, firms approved under IR (Industrial Revitalization) would spend less time in the patent queue. Also, the fees for obtaining patents were reduced (in particular those potential innovations demonstrating an industry-private industry link). Second, it supported TLOs (technology licensing organizations), also for promoting university-private sector new-product R&D. Lower fees would be collected from patent applications generated from these collaborations in the hope of promoting technology transfer from universities to industry. Third, there would be life after business failure through the Civil Reorganization Law (minji saiseiho), enacted in April 2000. The CRL was modeled after the old Chapter 11 Code in the U.S. in that a business owner was not, upon declaring business bankruptcy, forced also to declare personal bankruptcy. These efforts would be coordinated under the related Industrial Revitalization Corporation, or IRC.

Unfortunately, as is often the case in Japan, the ideals of the legislation often broke down in implementation in the private sector. For example, the first “approved” firms were the usual suspects -- meaning the big industrial monoliths, the least likely places for quick and nimble new business creation: Sumitomo Metal, Fuji Heavy Industry, Tokyo Motor, and Mitsubishi. Of the ten new business models touted by METI’s JETRO, seven were Tokyo firms (plus one in Osaka, one in Saitama, and one in Akita). It seems the intended national far-reaching impact of the IRL did not fall so far from Tokyo after all.

The recipients immediately instituted mass layoffs under the guise of returning to “core competencies” (under the auspices of the IRL and IRC). “Restructuring” (resutura) was now spun as core competence. In return, these firms got tax credits and tax holidays, such as the “angel tax” of 1/4 of capital gains. The angel tax was intended to jump-start high risk/high return investment through venture capital funds (its name is a misnomer, as angel investors are individuals rather than firms). Second, the IRC also helped to soften the burden for large firms of their earlier bad investments (it also camouflaged MOF’s failure in this regard). These trends in consolidation in the biggest Japanese firms were also occurring in the context of the 1997 repeal of the law preventing pure holding companies (i.e., zaibatsu).

After suffering through these policy failures in the 1990s, the new and revamped METI also moved away from its “network creation” in the 1990s (as it likewise did from earlier “technopolis formation” in the 1980s). By the early 2000s, METI’s new catchphrase was “regional clustering.”6 How did METI come up with the most ambitious policy package since the 1960s? It outsourced.

In 2002, Harvard Business School alumna and Hitotsubashi professor Yoko Ishikura was selected to head the “Cluster Plan Research Group.” In autumn 2002, she was dispatched to Cambridge to attend courses and workshops under the tutelage of Michael Porter (who was trained originally in mechanical engineering at Princeton University). Porter’s name has become synonymous with “cluster” through the popularity of his “diamond model” of innovation (Porter 1990; Porter 1998).7 (See Chart 4: Sources of Local Competitive Advantage)

The diamond model shows the main foundations of successful innovative clusters of firms.8 For Porter, clusters are spatial agglomerations of interconnected firms and institutions in a particular field, linked by commonalities and complementarities (Porter 1999: 199).9 The availability of various factor input conditions (tangible and intangible resources), supporting industries (suppliers), and demand conditions (customers) all interact to create an environment conducive to innovation and cooperation within local firms. I address the local context at the heart of successful (diamond) clusters below.

Three influences of successful clusters on competition are, for Porter, increasing productivity, increasing capacity for innovation, and stimulating new business formation. These happen to be the three goals of METI’s new Plan. In 2000, Porter, with Hirotaka Takeuchi and Mariko Sakakibara, had published the results of an eight-year research project on Japan’s economy. (Porter, Takeuchi et al. 2000)10 Around the same time, Porter’s ideas gained salience in the policy-making groups that orbit around METI.

In March 2003 a group of about 500 Japanese bureaucrats and academics, mainly from the top ministries and universities in Tokyo, attended the first “Industrial Cluster Conference” whose keynote speaker was Porter himself. By this time, METI had itself a new plan, one that seemed quite like the ones that came before it -- although no self-respecting bureaucrat in Tokyo can be found uttering names like “Technopolis” and “Brains Location.”11 This one, however, had the Harvard name, which carries much weight among Japan’s economic and political elites. In fact, in all of the volumes of METI “Cluster Plan” documents there is a noticeable absence of references to any other perspective than Porter’s. A fact check confirms the eerie similarity of structure and content of Japan’s most ambitious national policy initiative in 30 years to a single chapter (“Clusters and Competition”) in Michael Porter’s On Competition.

In emulating Porter’s approach to clusters, however, METI has lost in translation several critical factors that energize cluster development and sustained innovation. Porter observes that even in spatial agglomerations of firms that appear to have all the requisite basics (factor, support, demand), innovation and new business creation do not necessarily follow:

"While the existence of a cluster makes such [personal] relationships more likely to develop and effective once in place, the process is far from automatic. Formal and informal organizing mechanisms and cultural norms often play a role in the development and functioning of clusters."12

The core of the diamond is the “local context,” comprised of the social relationships described above. This core is said to encourage appropriate forms of investment and sustained upgrading. This context is not purely cooperative. In fact, the most innovative communities are charged with vigorous competition among locally based rivals, according to Porter. Further, the emergence of clusters is often inextricable from particular local histories (e.g., the presence of certain natural resources) and chance (e.g., the happenstance of a particular firm having located its operations and then spurred later unanticipated spin-offs in a certain sector).

Porter describes these social and cultural factors as the socio-economy of clusters. (Porter 1998) I would add that entrepreneurs within a given ‘local context’ are embedded in a complex environment of socio-political relationships. These relationships -- and the social capital (investments in trusting, reciprocal arrangements) that abound within the most successful clusters -- are the stuff of sustained community-wide innovation. In these successful clusters, civic entrepreneurs (enterprise mavericks engaged in activities that benefit the community as a whole in the long term and often benefit their firms only indirectly) have stores of social capital and often serve as brokers of tangible and intangible resources among community stakeholders. Analysis of these complex -- and region specific -- socio-political relations is lacking in cluster research to date in Japan. While the Japanese government seems to ignore the socio-political context of successful clusters, it also fails to heed Porter’s specific admonitions about what kind of role government should play in cluster development.

According to Porter (and based on case studies in 29 countries), successful national level policies geared at developing clusters (1) have been attuned to local level perceptions (aiming for shared understandings of the need for clusters), (2) include traditional and even declining sectors, (3) are inclusive and not exclusive (even of difficult individuals), (4) are private sector-led and ideally take place through an entity independent of government, and (5) have neutral facilitators of interpersonal communications that lead to trusting personal relationships. As METI is only midway through its implementation, the full impact of the Cluster Plan cannot be judged. However, we can get a sense in 2004, of how METI has gone about emulating “Porter-style” cluster development. At this point, we can evaluate the Plan in two major areas: its divergence in practice from the prescriptions of Porter and its distinctions (if any) from METI’s earlier policy initiatives.

Japan’s METI diverges in several ways from the Porter-style approach. First, the Cluster Plan is not only a government project, but is a national government project to boot. Worse, national control is being attempted in a situation where national/regional/local government relations in Japan are sometimes contentious and cannot be described as entirely trusting. Further, the Plan is wholly focused on high tech industry, and traditional/declining sectors are negligibly incorporated, if at all. Finally, the “coordinators” of inter-firm, inter-personal networking in Japan’s Plan are largely METI bureaucrats, hardly neutral parties. Japan’s national bureaucracy also seems to miss the region-enterprise (and often wholly organic) nature of the world’s most successful clusters.

For example, of the three American regions that METI has chosen as models, Silicon Valley, Austin, and Philadelphia, only the Philadelphia area has direct links to federal government programs.13 Even the Cluster Research Group’s leader Ishikura admits that the Japanese national government is attempting to create -- by policy fiat -- clusters modeled on innovative communities in the United States that developed without one iota of federal government initiative. For example, Silicon Valley emerged from an organic process (university initiated, local firm linked, with virtually no government involvement in the early stages). Austin, although supported by the local government, was also an enterprise-initiated by the founder of Teledyne. The state took a hands-off approach.

Consequently, I would argue that Japan’s national METI clusters fall short on the most critical factors (i.e., the socio-political, region-initiated and enterprise-led “local” context) that make clusters competitive with the most successful clusters worldwide. In essence, METI’s Plan diverges in critical aspects from the bases of successful efforts elsewhere. At the same time, METI has been unable to break out of its historical top-down, centralized management style and create truly new and improved policy.

In a recent government-sponsored book on Cluster Plan 3 (Ishikura, Fujita et al. 2003), the authors outline five “differences” between the new plan and its predecessors. First, the method of targeting firms and institutions for clustering is said to be different. METI has supposedly shifted to a region-initiated policy. Until now, according to Ishikura, Japanese national government policy has been characterized as “mura okoshi, baramaki” (the central state trying to force major change on sometimes intransigent regions, while trying to please all parties by spending money recklessly).14

Second, the goals and expectations are said to be higher. In other words, process improvements to existing technology will no longer suffice. METI acknowledges that to emerge as a competitive and innovative country, Japan’s national government must provide incentives for new product innovation. Third, the boundaries (limits) of the industrial targets are broader than ever before. In making these changes, METI is expanding the “value chain” of clusters not only by emphasizing the need for technological development but also marketing, management, and finance. Fourth, METI promises that they are committed to the Cluster Plan for the long haul: twenty to thirty years. Fifth, METI is being more selective in its targeting. This selectivity has two implications. On the one hand, METI will no longer start from nothing through green or brownfield investments. Instead, it has focused on identifying existing agglomerations of firm-level potential. On the other hand, METI’s biggest bet is on high tech industry, particularly IT and bio-pharmaceuticals. In 2004, the Plan was through the second of its four stages: preparation, formation, clustering, and establishment. The budget for these activities in the first two years was 44.8 billion yen (2002) and 41.3 billion yen (2003).

Stage I (preparation) began officially in 2001, and a special budget allocation was not received until 2002. In preparation for cluster formation, METI officials searched for (new) spatial agglomerations of cooperative, complementary related firms in bio, IT, and high-tech manufacturing. The three largest of the nine regional clusters are Tokyo, Hokkaido, and Kansai (Kinki), outlined below (Inoue 2003).

Tokyo’s clusters are located in the northwest environs of Tokyo’s metropolitan area (West of Shinjuku station on the Chuo line to Hachioji City), colloquially referred to as “Tama.” By WWII, the Tama region had become home to numerous small manufacturers, many of whom had relocated from the Keihin (Southeastern Tokyo) coastal area. These manufacturers produced electrical and precision machinery as well as transportation equipment. In the mid-1990s METI’s SME Agency for high-tech based network formation targeted Tama. By the late 1990s, network policies targeting Tama firms incorporated universities, government organizations, and others. By 2003, the Tama Regional Cluster, spread between cities like Hachioji and Sagamihara, included 260 firms, 17 local governments, 28 universities, and other support institutions (public research institutions, incubators, financial firms).

Between 2001 and 2002, 1.73 billion yen of public funds were invested in 56 companies and 17 universities to promote the management consulting and technical assistance of appointed “coordinators.”15 Inter-firm networking was promoted through “order-exchange” and “technical assistance” meetings as well as through promotional web-sites. The origins of Tama’s clusters couldn’t be less like those of Hokkaido.

Hokkaido’s IT/Biotech Clusters (“Hokkaido Super Cluster Promotion Project”) includes 230 firms (50 of which are biotech), 15 local governments, and a number of support institutions. The IT cluster is situated around Sapporo Station and has high concentrations of bio-tech enterprises (e.g., sugar-chain engineering). Following the national cluster model, activities include database development (for promotion of local firms), exchange meetings (like those in Tama), and “diversified business support networks.” In fiscal year 2001-2002 2.18 billion yen was invested in projects involving 73 firms and 26 universities.

Hokkaido’s IT cluster emerged, however, in the 1970s through the efforts of university researchers at Hokkaido and other universities in conjunction with entrepreneurial firms like Hudson, Softbank, Tomcat, B.U.G. and later (in the 1990s) Cyber Trust and Soft Front. Unlike the hodge-podge of Tokyo Tama cluster firms, numerous spin-offs from these entrepreneurial start-ups stimulated further cluster development over the last twenty years. In sum, Hokkaido’s clusters have evolved out of an organic process. The origins of Kinki’s clusters are more like Hokkaido’s and less like Tama’s.

Kinki’s Bio clusters include 200 firms, 9 local governments, 36 universities, 14 public research institutions, 20 incubators, and 24 financial firms. Its two technological cores are pharmaceuticals (4 firms) and tissue engineering in firms located in and around Kobe, Kyoto, Nara, and Osaka. Cluster activities include exchange meetings, technical assistance workshops, and web/email clustering mechanisms supported by METI’s national Plan. In fiscal year 2001-2002, 3.18 billion yen of public funds were invested in 81 firms and 93 universities.

These three regions (Tokyo’s Tama, Hokkaido’s Sapporo, and Kinki) exhibit several regional variations on the national cluster model. For example, Hokkaido’s clusters appear to have emerged from an organic process (universities and firms in the 1970s), while Tama’s seems like a hodge-podge of various sized manufacturers dispersed across a number of industries and cities. As a result, Tama’s clusters seem to lack a cohesive “cluster core” around which complementary, cooperative clustered firms are said to revolve in the national model. In Kinki’s clusters there appears to be more cohesion in the technology core (bio-pharmaceuticals), and also much greater diffusion of this technology across firms. Reports (such as T-san’s above) from firms and Kinki regional government officials indicate that a number of these “new” agglomerations were well-known to local officials and had proven track records in the very areas that the new Cluster Plan is said to promote.
On the one hand, this selection process (attaching the national label to existing agglomerations and local level initiatives) guarantees the ability of national METI (and MEXT) bureaucrats to report success in their Tokyo “Cluster Seminars.” On the other hand, it has further undermined trust (and morale) in the sincerity of national level efforts to actually shift to so-called region-initiated policy. Having “prepared” the regions in Stage I, METI moved to Stage II.

Stage II (formation) involved regional METI officials conducting extensive surveys of industrial, semi-governmental, and private research institutes related to the targeted sectors in each region. For example, Kansai’s METI (METI Kinki) conducted a survey of 701 firms and 170 organizations -- of which 153 and 53 responded (response rates of 21.8% and 31.2% respectively).16 The nature of the survey questions indicate that at least the MITI Kinki office is assessing the firm level needs and then responding with appropriate policy mechanisms (even though they were already doing this before the new Cluster Plan). If the regional METIs are actually able to draw significant amounts of national level resources into enterprise-based policy (say for development of a web-based communications backbone for inter-firm exchange, collaborative manufacturing, and the like) then this might actually distinguish this Plan from its predecessors. By 2003, METI, having completed reports that outline the spatial conditions (and expressed needs) of firms in the targeted sectors (Stage II), was ready to begin clustering full force.

Stage III (clustering) focuses on the clustering of firms and institutions through facilitating (human) networking. This involves two main activities: sending emails to target firms with various industry and topical information; and encouraging firm managers and technologists to get together for symposia and workshops. These networking workshops are often moderated by academics, and occasionally a successful local entrepreneur will give a talk.

Reports from the entrepreneurs of this study indicate that the most innovative firms ignore the emails. Further, they report that they do not see much value in attending cluster workshops. That is, the productivity of their already established personal networks is just fine. If asked to attend as a keynote speaker, however, they become interested. Talking to a workshop about their own firm has PR value -- and a new client might come out of the hours spent on the activity. This interaction between established entrepreneurs and brand new start-ups might result in the establishment of truly new (and value-added) networks -- for the new entrepreneurs at least (but we already know that the main beneficiaries of the Cluster Plan so far have been established medium-sized firms). As of this writing, METI was struggling to get the right mix of attendees (and in the right numbers) at the regional workshops.

Stage IV (establishment, yet to be implemented nationally in 2004) involves activating industrial clusters through building density, encouraging international inter-cluster exchanges, and expanding the scope (Inoue 2003; Ishikura, Fujita et al. 2003). In the Plan, density is built in two ways. First, firms are expected to use support personnel, such as “coordinators,” more frequently. Perhaps this is intended as a government manufactured proxy for angel investors, consultants, and civic entrepreneurs who are abundant in successful clusters elsewhere (and who act as internal and external brokers of community resources). I would argue that it is around these “brokering” activities that innovative communities emerge and coalesce.17 It also helps if these interactions are charged with a shared vision and rest on trust. Consequently, if the national METI can manage to get their coordinators placed in the right way (selecting home-grown civic entrepreneurs with proven track records would help) in local clusters, it might work. Second, confidentiality agreements are expected to increase inter-firm cooperation (it is unclear from policy documents, but I assume that the basis for these formal agreements is that having established the prerequisite foundation of informal trusting, cooperative relationships among firms will follow). Domestic density will ideally (for METI) be complemented by international openness.

International inter-cluster exchanges are expected to accelerate technology transfer and inward FDI. The scope of Japan’s clusters, if all goes according to the Plan, should be expanded in three ways: creation of trial product-manufacturing inter-firm networks; creation of financial, personnel, management consulting service networks; increased involvement of trading companies and other demand-side entities crucial to Porter’s “feedback” mechanism to ensure ongoing competition and quality improvements. My observations on the ground on the state of METI’s Plan in 2004 echo the sentiments of METI’s own commissioned research.

METI commissioned several private sector studies currently underway by Mitsubishi Research Institute (MRI) and UFJ to assess the project mid-stream. By 2004, the 19 regions had entered the third stage, where human networks were being facilitated by getting interested business people together with government and university folks. (Ishikura, Fujita et al. 2003), (Ministry of Economy Trade and Industry METI Kinki Region 2003). Unfortunately, METI doesn’t seem to be heeding the findings of its earlier commissioned research.

One such study, done by Mitsubishi Research Institute, a private-sector research organization with perhaps the closest ties to METI, found that despite the sincere efforts by central state bureaucrats to both design and implement “network creating” policies throughout the 1990s, the state largely failed. This comprehensive study conducted in the mid-1990s found that networks sponsored by any level of government fared poorly. One reason for failure is that governments in attempting to please everyone with broad and extensive policies (sobana teki na tori kumi) end up doing poorly in each specific area.18

Attempts by Japan’s national government to create Silicon Valley-like clusters have focused on the basic ingredients for innovation (e.g., infrastructure and formal institutions). Unfortunately, METI has not been very effective in developing sufficient conditions such as a shared national-local vision among community stakeholders and civic entrepreneurship. The flip side of this problem is that the national METI lacks social capital in the regions even to plug into existing coalitions of community stakeholders. In short Japan’s Cluster Plan is heavy on fabricating the gesellschaft (formal institutions) that undergird the clusters but light on nurturing the gemeinshaft (informal social) relations that energize clusters.

Successful innovative networks (or “clusters”) are infused with intangible know-how and vision possessed by community members and socio-political savvy on the part of civic entrepreneurs in brokering resources. These informal relations provide both the glue and the electricity -- the conduit -- between institutions and people. Creating gemeinschaft is proving to be a daunting task for Japan’s METI Cluster Plan.

KATHRYN IBATA-ARENS, Ph.D. Northwestern University, is an assistant professor in the department of political science at DePaul University in Chicago. Her research interests are in Japanese political economy and innovation and technology policy. This article is excerpted from a forthcoming book on the politics of innovation in Japan. Ibata-Arens was recently a JSPS post-doctoral fellow (2002-2003) at the Center for Advanced Economic Engineering (AEE), Tokyo University, and was a fellow in the Alfred P. Sloan/Social Science Research Council Program on the Corporation as a Social Institution (2002). Her work has appeared in the Asian Wall Street Journal, Asian Business and Management, Asian Perspective, and Review of International Political Economy. She is the author of JPRI Working Paper No. 59 (July 1999), “The Business of Survival: High-Tech Small and Medium-Sized Enterprises in Japan.” Ibata-Arens is a guest co-editor and author of the theoretical introduction for an forthcoming special issue on embedded enterprises in the journal Enterprise and Society.


Broad, W. J. U.S. Is Losing Its Dominance in the Sciences. New York Times. New York.

Fumio Hayashi, E. P. (2003). The 1990s in Japan: A Lost Decade.

GEM 2003 Executive Report, Global Entrepreneurship Monitor (GEM).
The Global Competitiveness Report 2002 – 2003. New York: World Economic Forum, Oxford University Press.

Inoue, H. (2003). Activating Industrial Clusters - RIETI Cluster Seminar no. 5. RIETI Cluster Seminar, Research Institute of Economy, Trade, and Industry (RIETI).

Ibata-Arens, K. “Alternatives to Hierarchy in Japan: Business Networks and Civic Entrepreneurship,” Journal of Asian Business and Management, Volume 3, Number 3 (13 August 2004)

Ishikura, Y., M. Fujita, et al. (2003). Strategy for Cluster Initiatives in Japan. Tokyo: Yuhikaku Publishing.

Johnson, Chalmers (1986). MITI and the Japanese Miracle: the Growth of Industrial Policy, 1925-1975. Tokyo: Charles E. Tuttle Co.

M Interview (2003).

Ministry of Economy Trade and Industry METI (2002). Kinkichiiki ni okeru sangakukanrenkei no genjo ni tsuite (The State of Industry-University Relations in the Kinki Region). Kinki, Office for the Promotion of Industry-University Relations, Kinki METI.

Ministry of Economy Trade and Industry METI (2002). Sangyo kurasuta keikaku ni tsuite purojekuto gaiyo (Regarding the Industrial Cluster Plan: Project Summary), METI Regional Economy Industrial Group.

Ministry of Economy Trade and Industry METI (2002). Sangyo kurasuta keikaku ni tsuite, chiikisaisei sangyoshushukeikaku (Regarding the Industrial Cluster Plan: Project Summary), METI Regional Economy Industrial Group.

Ministry of Economy Trade and Industry METI Kinki Region (2003). Kaurasuta coa jittai chosa (Cluster Core Survey on Actual Conditions). 2002 Annual Survey on the Actual Conditions Among SMEs. Kinki, METI.

Mitsubishi Research Institute (MRI) (1996). Shuseki Jibasangyo Shinnettwaaku Kochiku, Jireichosa, (March 29, 1996) (Comprehensive Report on the Structure of Local New Business Networks: Case Study Report). Tokyo, Mitsubishi Research Institute.

Nakagawa, T. The Creation of Intellectual Clusters in Japan - RIETI Cluster Seminar no. 3. RIETI Cluster Seminar, Research Institute of Economy, Trade, and Industry (RIETI).

Porter, M. E. (1990). The Competitive Advantage of Nations. London: Macmillan.

Porter, M. E. (1998). On Competition. Boston: Harvard Business School Publishing.

Porter, M. E., H. Takeuchi, et al. (2000). Can Japan Compete? Cambridge, Massachusetts: Perseus Publishing.

T Interview (2003).

Wada, K. (2003). Development of Policies to Support Industrial Agglomerations in Japan and Future Issues. Globalization, Regional Concentration and Clustering of Industry. H. Fukushima. Tokyo: Research Institute of Economic Science, College of Economics, Nihon University.

Y Interview (2003).


1. This underscores the point that innovative capacity does not necessarily translate in automatic fashion into entrepreneurial strength. For example, according to the Global Competitiveness Report 2002-2003, Japan ranks fourth (behind U.S., U.K., Finland, and Germany) out of 58 countries in terms of national innovative capacity. This capacity is a composite of measures of scientific and engineering manpower, innovation policy, cluster innovation environment, innovation linkage, and company innovation orientation. See Michael E. Porter and Scott Stern, “The Impact of Location on Global Innovation: Findings from the National Innovative Capacity Index,” Table 1, National innovative capacity rankings, p. 229. Return to Text

2. Chalmers Johnson’s seminal MITI and the Japanese Miracle (1982) explores MITI’s role in industrial targeting behind Japan’s high growth period. Johnson, C. A. (1986). MITI and the Japanese Miracle : the Growth of Industrial Policy, 1925-1975. Tokyo: Charles E. Tuttle Co. Return to Text

3. A cluster is generally defined as a spatially agglomerated group of firms that produce similar products, while having complementary ties with one another. Return to Text

4. For an in-depth discussion of the role of collaborative manufacturing networks in various regions in Japan, see Kathryn Ibata-Arens “Alternatives to Hierarchy in Japan: Business Networks and Civic Entrepreneurship,” Journal of Asian Business and Management, 3:4 (13 August 2004). Return to Text

5. The 1990s have been described as “Japan’s lost decade” Fumio Hayashi, E. P. (2003). "The 1990s in Japan: A Lost Decade.” REITI, EWC August 2004 Tokyo Conference.Return to Text

6. Ministry of Economy Trade and Industry METI (2002). "Sangyo kurasuta keikaku ni tsuite, chiikisaisei sangyoshushukeikaku" (Regarding the Industrial Cluster Plan: Project Summary). METI Regional Economy Industrial Group. Ministry of Economy Trade and Industry METI (2002). "Sangyo kurasuta keikaku ni tsuite purojekuto gaiyo" (Regarding the Industrial Cluster Plan: Project Summary), METI Regional Economy Industrial Group. Ministry of Economy Trade and Industry METI (2002). "Kinkichiiki ni okeru sangakukanrenkei no genjo ni tsuite" (The State of Industry-University Relations in the Kinki Region). Kinki, Office for the Promotion of Industry-University Relations, Kinki METI. Return to Text

7. The Global Competitiveness Report issued by Porter’s Council on Competitiveness seems to have had a significant impact on the Cluster Research Group. The Report outlines various (particularly macro) structural factors supporting competitiveness across 74 countries. Measures included are patent figures, IPOs, opinion surveys to business executives on perceptions of competitiveness, and the like. (2003). The Global Competitiveness Report 2002 – 2003. New York: World Economic Forum, Oxford University Press. Return to Text

8. One method of measuring the potential for and/or existence of innovative and competitive regions or spatial clusters is to identify the composition of its value chain (vertical alliance of enterprises optimizing internal and external supply-chain functions) in traded clusters (trading products and services outside the region, in contrast to local clusters, whose producers serve exclusively local markets). For example, if the bulk of a region’s firms produce goods (collaboratively and competitively) that compete with goods produced elsewhere for national and international markets, then that region can be said to have a successful cluster in a particular product. A further step in analyzing a region’s relative success is to compare local firm output and sales by industry compared to the national average. Return to Text

9. Porter cites Alfred Marshall’s nineteenth century writings on the externalities of specialized industrial locations as the intellectual foundation of clusters. Porter, M. E. (1998). On Competition. Boston: Harvard Business School Publishing. p. 206. Return to Text

10. Porter and his research team found that in aggregate Japan’s postwar national targeting produced more uncompetitive industries than successes. On the other hand, industries that were not targeted (motorcycles, 1960s; audio equipment, 1970s; autos, 1980s; game software, 1980s) emerged as Japan’s most competitive industries. Return to Text

11. The laws underlying these policies were enacted in 1983 and 1985 respectively. Return to Text

12. Porter, M. E. (1998). On Competition. Boston: Harvard Business School Publishing. pp. 213-214. Return to Text

13. A few European cases are also mentioned (Finland, France, Germany). In May 2004, the New York Times reported that the United States is slipping from its position as world innovative leader. See NYT, May 8, 2004. Broad, W. J. "U.S. Is Losing Its Dominance in the Sciences." New York Times. New York. Return to Text

14. Not surprisingly with a bureaucratic project of this magnitude in Japan, there also seems to be a bit of inter-ministry rivalry. Within months of METI’s launching the “Cluster Plan,” MEXT (Ministry of Education, Science, Culture, Sports, Science and Technology) launched its own “Knowledge Clusters” project. Some interviewees note that METI’s cluster initiative, with its emphasis on greater university-industry collaboration, is intruding on territory traditionally under the purview of the former Ministry of Education. Nakagawa, T. The Creation of Intellectual Clusters in Japan - RIETI Cluster Seminar no. 3. RIETI Cluster Seminar, Research Institute of Economy, Trade, and Industry (RIETI). Return to Text

15. Coordinators appear to be culled from the ranks of METI bureaucrats, firm managers, and university professors. Ishikura, Y., M. Fujita, et al. (2003). Strategy for Cluster Initiatives in Japan. Tokyo: Yuhikaku. Return to Text

16. Ministry of Economy Trade and Industry METI Kinki Region (2003). "Kaurasuta coa jittai chosa" (Cluster Core Survey on Actual Conditions). 2002 Annual Survey on the Actual Conditions Among SMEs. Kinki, METI. Return to Text

17. Informal investment (assets of family and friends) is said to comprise the vast majority of start-up capital for entrepreneurs, far outweighing the role of angel investors and venture capitalists. See GEM 2003 Executive Report, Global Entrepreneurship Monitor (GEM). Return to Text

18. Mitsubishi Research Institute (MRI) (1996). Shuseki Jibasangyo Shinnettwaaku Kochiku, Jireichosa, (March 29, 1996), (Comprehensive Report on the Structure of Local New Business Networks: Case Study Report). Tokyo: Mitsubishi Research Institute. See also: Wada, K. (2003). "Development of Policies to Support Industrial Agglomerations in Japan and Future Issues." Globalization, Regional Concentration and Clustering of Industry. H. Fukushima. Tokyo: Research Institute of Economic Science, College of Economics, Nihon University. Return to Text

Downloaded from