JPRI Working Paper No. 59: July 1999
The Business of Survival: High-Tech Small and Medium-Sized Enterprises in Japan
by Kathryn C. Ibata-Arens

Small and medium sized enterprises in Japan are forced to do business in an environment dominated by the state and big business. A myriad of institutions--what I shall call "intermediating hierarchies"--serve the interests of the state and big business. These big business-biased, often state-led, institutions have become the root cause of barriers to innovation in small and medium-sized enterprises (SMEs). They also threaten the health of the economy as a whole. Barriers include the Ministry of Finance (MOF)-controlled financial system, keiretsification of the economy, and the patent system.

Despite these overwhelming barriers to success, a handful of SMEs have survived and prospered. These innovative, high-tech SMEs have avoided being assimilated into vertical production structures (that is, avoided being "keiretsified") by not doing business with traditional intermediary links and intermediating agents. They have employed four major strategies to retain their independence. They have a) controlled sales ratios among customers, b) self-funded new product research and development, c) forged connections with other SMEs internationally, and d) created internet-based networks that subvert traditional channels of information.

In the course of my research, I surveyed 43 small and medium-sized firms in three industrial areas: Tokyo's Ota ward, the southern technology corridor of Kyoto, and Higashi Osaka. I was able to interview the presidents and technology managers of 22 high tech manufacturing SMEs in Tokyo's Ota ward and of 21 firms in Kansai (twelve in Kyoto's southern corridor and nine in Higashi Osaka). In both regions, I also interviewed a variety of local, regional, and central government (e.g., MITI) officials; directors and managers of private and semi-public research institutes (e.g., the SME Finance Corporation); and directors of technology parks and industrial promotion centers (such as Kanagawa Science Park, Kyoto Research Park, and Gakkentoshi). I also interviewed some leading Japanese SME scholars (including Toshio Aida, Hisayoshi Hashimoto, Tomohiro Koseki, and Hirofumi Ueda) to gauge their opinions of my interpretations of the role and position of SMEs in the Japanese political economy. In addition, I obtained survey data on SMEs across industries that seem to indicate that the patterns and trends I identified in high tech manufacturing reflect the overall situation of small and medium-sized businesses in Japan.

It is generally agreed that small and medium-sized firms serve as the innovative and productive foundation of healthy economies. This is particularly true in periods of new business formation. Japan's economy, however, has been characterized by "keiretsification," in which small and medium-sized firms are routinely assimilated into vertically integrated production structures. Access to capital for new product development and to the market for goods is dominated by big business and bureaucrats beholden to big business. These big business-biased structures, found throughout most intermediating hierarchies in Japan, have been an overall bad thing for SMEs and have become a detriment to innovation and healthy new business formation in Japan.

A System in Crisis

Teikoku Databank, a private Japanese credit reporting agency, reported in July 1998 that corporate bankruptcies during the first half of 1998 rose more than 29 percent over the same period in 1997. Small and medium-sized firms comprise more that 99 percent of businesses and employ three-quarters of the workforce in Japan. This was the second highest level of bankruptcies since the end of World War II. Reasons cited for bankruptcies by firms going under have included: lack of capital, bankruptcies in related firms (such as a main client), credit crunch, and depressed sales. Start-up rates of new businesses remained lower than closure rates throughout the 1990s. Start-up rates for firms of less than 20 employees--a foundation for new industry creation in an economy--remain significantly below closure rates of firms of the same size. Producer confidence remained low into the first quarter of 1999.

Several interviewees commented on the mood among Japanese businesspeople as similar to that leading up to the Great Depression. The National Police Agency reported in July 1998 that there were 24,391 suicides in Japan during 1997, up 5.6 percent from 1996. Suicides attributed to work-related stress increased by 17.6 percent. Less than 60 percent of new college graduates found full-time employment in 1998.

A 1998 NHK survey of Osaka area firms reported that 90 percent of the firms described their situation as worse than a year earlier. Sixty percent of the respondents said that obtaining funds from banks and other financial institutions had become more difficult in the past year. A mere one percent reported that it had become easier to obtain funds. Thirty-four percent of the firms reported that in order to alleviate cash flow problems, they had resorted in the past year to selling assets. Sixteen percent said they were forced to contract business operations. A low two percent said that they had appealed to a parent firm for assistance.

In a television documentary on the Asian economic crisis (shown on NHK November 14, 1998), the experience of one high-tech (computer mother board) producer, was chronicled. At the beginning of 1998, this firm was growing, expanding operations and making a profit. In order to facilitate its expansion, the firm had recently obtained technology loans from a local bank. In the spring of 1998, however, the bank suddenly cut off funds and called in all previous loans. Bank representatives repeatedly berated the president of the firm, reportedly saying that "your firm is a 'B' or 'C' rank firm and doesn't deserve financing." The firm was forced to make large monthly payments on the loans. To exacerbate the situation, in May of 1998, one of its main buyers paid a large outstanding invoice with a bad draft (fuwatari-tegata). Despite these extenuating circumstances, the bank refused to let up on the demands for payment. By October of 1998, the firm was forced into bankruptcy.

A July 1998 article in The Economist concludes by citing the current high bankruptcy rate in Japan as evidence of "creative destruction." What is not mentioned in the article is that the firms going under are not necessarily "bad" or nonperforming firms. Moreover, the high levels of bankruptcies in Japan in the 1990s have so far not been followed by significant new business creation. A comparison of Higashi Osaka and Tokyo's Ota and Sumida wards shows a steady decline in both the number of plants and manufacturing output over a ten-year period after 1985. Between 1985 and 1995, the number of manufacturing plants in Higashi Osaka fell from 9,933 to 8,933, a decline of more than 10 percent. In Ota ward, the number dropped from 8,897 to 6,787 over the same period, a fall of nearly 24 percent. Sumida firms had a similar decline, from 7,133 to 5,514, almost a 23 percent drop. In terms of manufacturing output, all three areas experienced similar declines.

Even large firms are feeling the pressure. Large keiretsu groups have laid off thousands of full-time employees and scaled back operations, both domestically and abroad. Even industry leaders such as NEC and Nissan are experiencing difficulties. The unemployment rate remains at a postwar high. In the past, large firms have been able to offset inefficiencies with easy access to capital through their main keiretsu banks as well as having preferential access to government funding. As I will describe below, most SMEs have found it extremely difficult to access capital and government programs. Recent big bank failures have exposed the immense scope and scale of weakness in the financial system. The "bad loan problem" (furyousaiken mondai) and credit crunch (kinyuu hippaku, shinyou hippaku) show a system in desperate need of fundamental reform. Until the true inefficiencies are routed, Schumpeter's creative destruction remains unlikely.

The Japanese government has tried to implement a series of policies to deal with these problems. For example, in November 1998, The Ministry of International Trade and Industry (MITI) introduced legislation aimed at promoting new business formation. Loan guarantees would be offered to regional credit associations to assist new start-ups (although funds would be forthcoming only to those that already had part of the funds needed to start up their businesses). Another policy was announced by MITI in October 1998, ostensibly to support struggling SMEs. Forty trillion yen in government backed loans was earmarked for SMEs, supposedly for investment in new equipment and technology. Before the ink was dry on the legislative books, however, firms were approached by bank representatives, urging them to apply for the funds. Many 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 said to firms that if they applied for this new program (with the bank's assistance), they would be certain to get the funds. But in return, firms were expected to use these funds to pay back old loans.

Since the program was backed by the Ministry of Finance (run by MITI), there has been speculation that this so-called program for SMEs was actually a veiled bailout for the banks. The sheer speed with which banks at the local level all over Japan were able to obtain detailed information and applications for the program further supports such speculations. At the same time, there are reports of strong-armed practices by banks to collect called-in loans. One bank employee, interviewed anonymously, told of a "Collections Manual" that showed how medium to smaller-sized firms were specifically being targeted. Mistakes in reporting and the like were identified (ochido), and then used as excuses for a sudden call on 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.

The Production Pyramid in Japan

The structure of the Japanese economy--in terms of the relative position of the firms that comprise it--is often described as a "pyramid," a "convoy system" (gosou-sendan soshiki), or a "flying geese formation." Each of these euphemisms accepts that big firms take the top, or lead position. Alongside the biggest firms at the top are of course the bureaucracies, namely the Ministry of Finance (MOF) and the Ministry of International Trade and Industry (MITI). Smaller firms, and also firms operating outside of the Kanto Plain, assume a subordinate position. They supposedly accept this since their larger counterparts are assumed to play a leadership role and take on certain risks.

The term "small and medium-sized enterprise" is used to describe various kinds of businesses, from small mom-and-pop stores to manufacturers employing 300 people. The technological nature of enterprises also varies across industries. For example, a medium-sized construction firm employing 250 people may have a significantly lower technology level than a small electronics firm of 10 people. Most U.S. research on the Japanese political economy to date has concentrated on large firms and their relations with MITI and other government ministries. Yet these large firms consist of less than one percent of total firms, employ no more than 25 percent of the workforce and contribute less than 50 percent to output and value added (see Table 1). At the same time, there is no doubt as to the immense political influence big business has in Japan, as a growing number of political corruption scandals involving major manufacturers, big banks, and their clients demonstrate.

Contribution to Japanese Economy by Firm Size (Manufacturing)
1983   1993
Firm Size1  % of Total # of Firms2 Employees Output Value Added % of Total # of Firms2 Employees Output Value Added
4-9 58.7 14.6 5.5 7.9 55.4 12.6 4.8 6.8
10-19 19.5 11.2 6.2 7.5 19.8 10.4 5.6 6.8
20-99 18.6 29.9 22.4 23.2 20.9 30.8 22.3 23.9
100-299 2.4 16.5 17.8 17.4 2.9 18.0 19.0 19.3
300-999 0.7 13.7 22.7 19.7 0.8 14.6 22.1 20.9
1000+ 0.1 14.0 25.5 24.2 0.2 13.6 26.2 22.4

1Firm size = by # of employees
2Total # of firms in 1983: 446,942 and 1993: 413,563

One common misperception of how firms interact in Japan is that "technological leadership is rewarded." That is, firms developing creative products and having unique technical expertise (or "innovativeness") are expected to dominate. But in Japan, political access to various ministries, particularly MITI, rather than innovativeness, has proven critical to obtaining expansion and R&D capital, including that from private lending institutions. Political access includes access to government-sponsored funding programs, as well as government-related financial institutions. Compared to big business, Japanese small and medium-sized manufactures have little access to political institutions. This has led to a vicious circle in that R&D for promising innovations is not funded at the SME level. This in turn lowers the overall technological base of SMEs over time.

There is one important caveat about Japanese SMEs, and that is, the policy access of industries such as agriculture, construction, and retail stores, far outweighs their contributions to GDP. Looking at manufacturing, which consistently contributes more than 50 percent of GDP, one might expect to see a corresponding level of policy influence, but this is not the case. In comparison, SMEs in the U.S. employ 20 percent fewer people than in Japan, and contribute a bit less in terms of output, but they enjoy far greater access to policy-making institutions. SMEs in Italy, where they contribute 20 percent less in terms of employment compared to their counterparts in Japan, and 15 percent less to GNP, also enjoy higher levels of policy access. In short, it is important to get a firm grasp of the role of SMEs and their relation to the institutions around them, not because they are the "unsung heroes" of the Japanese economy. Rather, they should be examined because they serve as the basis for the economy and indeed society as a whole.

Intermediating Hierarchies

Situated between the state and private sector, an intermediating hierarchy is an institution providing a basis for political and economic interaction between the industrial base (represented by small and medium sized enterprises) and the state. Political and economic interaction in this case includes such things as lobbying for particular industrial policies and maintaining general awareness at the state level concerning the macro-economic policy needed by the industrial base. Intermediating hierarchies, taken together, comprise the institutional framework around which the productive functions of a given political economy operate.

In the empirical evidence I gathered in Japan, I found five sets of intermediating hierarchies that play key roles: 1) the convoy-led financial system; 2) vertical keiretsu group links; 3) the patent system; 4) local horizontal (informal and formal) business networks among small and medium-sized enterprises; and 5) key SME governmental and non governmental organizations (e.g., SME industrial promotion centers, and peak industry/trade associations). The type of intermediating hierarchy having the most negative impact on SMEs and overall innovation in Japan consists of vertical keiretsu group structures. SMEs that have avoided being assimilated into these vertical production structures, are the most likely to be innovative, relative to other SMEs.

One problem of the intermediating hierarchies in Japan's economy is that over time they have emphasized not creativity and ideas, but rather, procedures and connections. The vertical structure, and procedural nature of these intermediating hierarchies have also been combined, over time, with a bias toward big business. Furthermore, these intermediating hierarchies have become rigid, particularly where political and financial power is concentrated (ministries and big banks in Tokyo), allowing access to the most lucrative only by large corporations and their sponsored firms. New points of access are slow to form. This has been an important factor in the failure to foster a venture capital market in Japan and has contributed to the stifling of new business formation.

Firms have had few alternatives to operating within these biased structures. Some, however, have tried to deal with these intermediating hierarchies in two ways. First, they have tried to establish intermediary links with government and other institutions through independently formed organizations such as the Chuudoukyou (National Committee for Medium and Small Enterprises). Second, they have formed local business networks.

The Convoy-Led Financial System

Critics of the Japanese financial system describe it as too risk averse, too dependent on collateral-based evaluations, and lacking expertise on venture business. Interviewees cited five aspects of the Japanese financial system that have posed barriers to their business operations. In order of magnitude, these are: the lack of an "idea-based" venture capital system, lack of expertise on the part of bank employees, flawed lending practices, unfair lending practices, and difficult borrowing procedures. In addition, many interviewees cited the system of demand draft payment (tegata) as being unfair to small business. Like the vertically integrated production pyramid in Japanese manufacturing, the financial sector is structured as a "convoy system" (gosou-sendan soshiki), led (and arguably controlled at all levels) by the Ministry of Finance (MOF). Furthermore, in contrast to MITI's SME Agency, which has a limited budget, the funds brought to bear by MOF to keep banks in business is exponentially vast. Bureaucratic control of the day-to-day operations of banks at the micro level is one factor that has led to an overall aversion to risk in the financial system. Bank employees, particularly loan officers, are loath to do anything that goes against the standard procedures handed down from MOF. This aversion to any risk, is manifested most severely in the low levels of lending to start-ups and new businesses. Many interviewees reported similar experiences over time with MITI and private financial institutions. One of these is the experience of H firm:

When I first approached MITI's SME Agency and financial institutions for venture funds, I was turned down. I couldn't get any funding from elsewhere, so I sold my house. Now (20 years later) I have so many so-called "venture- capital" firms, government funds, and banks approaching me and wanting to give [lend] me money. I send them away, because I no longer need it. I made it based on my own strengths and risk taking from the beginning. Even when figures can be cited by banks on lending to start-ups, these are often spin-offs of larger firms, and cannot be considered to be independent start-ups. These "spin-offs" have also often served as receptacles for post retirement placement of management in the original firm (i.e., amakudari).

Finally, the system of demand draft payments has had a negative impact on SMEs. Large firms routinely pay their debts to subcontractors in tegata. In theory, the draft is paid at some specified later date at face value (30-90 days). In practice, large firms postpone payment for as long as six months, and even longer in some cases. Many small firms experience serious cash flow problems as a result. Having few avenues of recourse in the legal system, and weak market power vis-à-vis large assemblers, small firms must turn to local banks, who make a discounted payment to the firm in cash, in exchange for the draft. Small firms, on the other hand, are not allowed to make use of the tegata system of delayed payment for themselves. Instead, payments to large firms for capital equipment, and to suppliers for materials are expected in cash on the date of purchase. Several interviewees outside of SMEs have shrugged off this inequality as part of the traditional pyramid structure of doing business in Japan, implying that it is unlikely to change in the near term.

Vertical Keiretsu Group Links

Over the past five decades the privileged access afforded big business to the state has allowed big business and its keiretsu groups to grow and dominate the Japanese economy. This privileged access has been further reinforced and extended via amakudari practices (employing retired bureaucrats and using them as conduits to their former ministries). Peak industry associations such as Keidanren, and big business policy councils (shingikai) are the primary institutions of this access to the state. Consequently, the resources made available to firms via government programs targeting industry tend to be channeled through these privileged connections. The result over time has been an increasing bias toward firms having established connections to government via keiretsu and the peak industrial associations representing their interests. Independent small and medium-sized firms have had little access to these channels. Of the 43 firms I interviewed, more than 67 percent had experienced some kind of negative impact as a result of preferential policies toward big business. Nearly 35 percent had suffered a significant negative impact from biased government policies.

"Parent-child" subcontractor relations have been the basis of contractual relations between Japan's large assemblers and small and medium-sized firms. However, many of my interviewees took offense at the term "parent-child" subcontracting relations (oyako gaisha). Several said that this term has always been a misnomer, since large assemblers have rarely acted toward their subcontractors in a manner remotely resembling that of a benevolent parent figure. Ronald Dore has argued that Japan has been able to prosper and nurture competitive industries precisely because of such social institutions, particularly so-called "trust-based" relations in business. Dore refers specifically to the utility of long-term, exclusive subcontracting relations as one example of how the Japanese system works (see his Flexible Rigidities: Industrial Policy and Structural Adjustment in the Japanese Economy 1970-80, Stanford University Press, 1986). In Dore's and other works, this "relational contracting" is often placed in contrast to the "spot contracting" prevalent in the U.S.

By "keiretsification" I mean the assimilation of small and medium-sized firms into vertical production structures. If 70 percent or more of a firm's sales go to a single client, despite a lack of a formal link, the buyer is often referred to as the "parent." In some cases, firms having 50 percent or more of their sales going to single buyer refer to that client as their parent. Several interviewees referred to such integration into the vertical production structures of a main buyer as being "keiretsified" (keiretsu ni kuwaerareta, keiretsu ni irareta, keiretsu sareta tokoro). Traditional studies on Japan, done mostly from the perspective of large assemblers, have focused on the benefits to large business of encouraging exclusive subcontractor relations. In the immediately postwar period, when Japan was a net recipient of technology transfers (from the U.S.) and thus was a technology follower, this system may have worked. But as we approach the 21st century, Japan can no longer depend on unilateral technology transfers (indeed, Japan has become a technology leader in its own right), and the costs of stifling independent small businesses far outweigh past benefits.

Benefits to big business in the past have included: stable business relations, the ability of large firms to control supplier prices, reduced competition for suppliers among large keiretsu groups, and the ability to externalize costs by exporting them to subcontractors in the form of "cost-down" measures. Large firms are able to impose "cost down" measures on their suppliers, who are beholden to them for orders. In "cost down," the unit prices of supplier products are routinely (twice a year or more) reduced on demand from the parent. Through this practice, large firms can avoid dealing with internal inefficiencies by pushing the costs of them off onto suppliers. At the same time, large firms are able to exploit their power over suppliers, by routinely engaging in slow payment of debt.

Many interviewees recounted their personal experiences with large assemblers. A typical story goes as follows:

I have learned from the mistakes made by other firms around me. Large assemblers come in and make an order for a few thousand pieces. Several months later, they ask for more and more, paying on time at first. Then, before you know it, their orders take up most of your production time. That is when they stop paying (on time), when they know you have no choice. Then they start with "cost down," and again you have no choice, because they know that they have become your primary customer. [A large assembler] tried it with me, but I wouldn't let them put me in that position. It was difficult at first, but we have survived and done well.

As this and other examples show, these supplier relations are based much more on power differentials than on trust. Even in cases where the relationship could have been described as trusting in the past, this trust has long since been betrayed. Large firms have increasingly exploited their market power through imposing unrealistic JIT ('just-in-time') production deadlines, forcing 'cost down' prices on subcontractors, and engaging in the slow payment of debt to their captive subcontractors.

The Patent System

Patents are said to encourage innovation through the public introduction of new products and technologies. In theory, a patent guarantees the owner a limited exclusive right to potential economic rewards provided by the market. In terms of the law, any violation of a patent is considered an infringement. Patent protection is said to be most important in industries producing products that are easily duplicated, have long product cycles, and have high initial research and development costs.

Japan's patent system can be called a semi "first-to-file," system. Most industrialized nations have "first-to-file" systems, with the exception of the United States, which adheres to "first-to-invent." There is apparently a push in the United States to switch to a first-to-file system, but it is unclear who stands to benefit from such a change. It is interesting to note, however, that big firms in the U.S. appear to be behind the recent push for this change, while small firms in the U.S. have been opposed. Japan's patent system cannot be considered a full "first-to-file" system for four reasons: laying open with no protection, the time lag between application and review, the need to apply through benrishi, or patent agents, and the extremely high costs.

First, Japan allows "laying open," that is, the publishing of patent applications for "opposition" before patent review (kokai). Thus, product specifications and the layout of inventions are made publicly available. At this time, "oppositions" to the proposed patent are accepted, in which competitors can argue that they too have invented the same product. The Japanese system is unique in that applicants have no legal protection (in the form of a preliminary patent) at the time of publication, unlike all other industrialized systems. In comparison, European systems, while allowing the laying open for opposition, do so only after a preliminary patent has been granted (see Table 2).

Patent Systems in the U.S., Europe and Japan
U.S. Europe Japan
System Type "first to invent" "first to file" semi "first to file"
Time to Obtain Patent 18-24 months within 2 years 5-15 years
Public Disclosure Upon issue of patent 18 months after application (with preliminary patent being granted) 18 months after application (with no legal protection)
Validity of Patent 17 years from issue 20 years from filing 20 years from original application
Surrounding Patents (around core technology) "picket fence"   "great wall" and "patent flooding"
Languages Accepted any (followed by an English translation within 2 months) any European (EPO)* language Japanese only
Cost and Fees moderate moderate high (including necessity of hiring benrishi)
Scope of Claims Granted broad broad narrow
* European Patent Office member states are Austria, Belgium, Denmark, France, Germany, Greece, Ireland, Italy, Liechtenstein, Luxembourg, Monaco, the Netherlands, Portugal, Spain, Sweden, Switzerland, and the United Kingdom. Although patent applications may be filed in any of the member state languages, a translation in either English, French or German must subsequently be filed.

It will come as no surprise that in Japan patent applications are routinely opposed, especially by large firms opposing smaller firms' patent applications. Another common practice by large firms is to engage in "patent flooding." In this practice, large firms obtain information on the core technology in patent applications of other firms, and use their considerable resources to gain patents on all conceivable permutations of the core technology and its surrounding technologies. In the past, this tactic was often used to pressure foreign, particularly U.S., firms to license their technology. Such practices obviously have a significant impact on a small firm's long-term profit potential on a given patented technology.

Second, the Japanese patent system is unique in the unusual length of time between the initial application and patent review. While this takes about a year in the U.S., and a little longer in Europe, it takes around five years or longer in Japan. This practice worked nicely when Japanese industry was a recipient of U.S. technology, since it gave time for then infant domestic producers to "catch up" with Western technologies (while U.S. firms waited in the long queue for Japanese patent approval). As with most institutions, once established, this aspect of the patent system has become immune to change. Thanks to five decades of preferential government policies, the postwar "infants" have grown enormously. In their dominant position, large Japanese firms have been unable to resist taking advantage of the technology and skills of smaller domestic firms, as they once did in the past with foreign technology.

Third, Japan's patent system involves voluminous paperwork that places an emphasis on procedural skill in filling out forms, as opposed to the new product innovation itself. The need to retain benrishi, or patent agents, to act as intermediaries between a given firm and MITI's Patent Agency is unavoidable. While benrishi routinely refer to themselves in English translation as "patent attorneys," there is some debate as to whether the relative level of expertise and education measured by the benrishi exam qualifies them to be called attorneys. There is also concern among firms about the overall quality of benrishi--e.g., complaints that they do not act as a firm's advocates. Nonetheless, the employment of them is unavoidable, since they are the MITI-designated intermediaries in patent issues. In fact, over a quarter of the current benrishi are former Patent Agency employees. For example, of the 135 people passing the 1997 benrishi exam, 51, or 38 percent, were former Patent Agency employees.

The fourth problem in obtaining a Japanese patent is that the financial burden of obtaining even one is overwhelming. Application fees and retainer fees for benrishi are just the start. The standard fee (set by the national organization of patent agents) for a full patent (tokkyoken) is twice as much as for a "utility model" (jitsuyoushinnan) (¥300,000 versus ¥150,000). While the former protects the entire product and process by which it is produced, the latter only protect the physical appearance of the product. It is estimated that for one patent (not including the surrounding patents), the cash outlay is at least ¥2,000,000--and that is just in the first year. The capital outlay to protect surrounding technology is exponentially costly.

What is called a "picket fence" strategy is employed in the U.S., whereby firms patent surrounding technology in addition to their core innovation, in order to protect themselves from other firms making slight modifications in the product and obtaining separate patents. In Japan this strategy can be called a "great wall" put up by large firms. Large firms habitually file multiple versions of their own patents, while at the same time engaging in "patent flooding" of other firms' patents. Whereas small firms can rarely afford to patent more than the core technology of an innovation, large firms routinely engage in patent flooding. This is one reason why patent data are not a sufficient measure of innovation in Japan, since most of these "flood" patents never become products.

Of the 43 firms I interviewed, nearly half either had personally experienced or knew another firm that had its technology expropriated by a large firm. More than 90 percent of all firms studied were aware of innovations being stolen from small firms by large ones. Of the firms having personal experience, most mentioned how large firms are able to exploit their market position in the process of patenting a new product or technology. One small firm recounted how a large firm's representative went so far as to carry a hidden camera in his jacket when on a plant tour:

Parts makers have high skills, while their parents have skill in stealing. For example, we had a large firm representative come in here on the pretense of "kengaku" (making a plant visit) a while back, and found out later that the person had managed to conceal a small camera in his jacket. That firm then copied our production methods on a particular product. Another firm recalled that a person presenting himself as a university researcher came in and asked for advice on how to improve on the design of a product:

Since I thought he was a student, I gave him specifications for design improvements for free. Later on I learned that he was a paid consultant for a large assembler.

Even if a firm obtains patents on its technology, this offers only limited protection. If a large firm sees a growing market potential for an existing product of another firm, a common practice is to make minor surface modifications in the extant technology and obtain a separate patent. Although this can be interpreted as an infringement of the core technology of other firms, large firms routinely get away with it. One defensive strategy by SMEs is to develop only products for small (niche) markets, thus making their products less attractive to mass producers. Of course this niche-market specialization also tends to limit the scope and thus profits of the firm.

Another example of the unequal treatment of small businesses in Japan is the differential impact of government regulations. Almost all the firms I interviewed felt that government regulations pose a huge hindrance to doing business. The three most cited regulations were government licensing procedures, government fees, and onerous government reporting requirements. Asked separately about the degree of negative impact of government regulations, 32 percent reported some or a significant negative impact from reporting requirements, while more than 72 percent reported a negative impact from other time-consuming procedures. Specific cases cited involved the process of obtaining patents and government funding procedures.

In contrast, several sources, including a bank loan officer, said that large firms having established connections to MITI and other ministries often did not have to prepare much paperwork in order to obtain government funds. Instead, bank employees were expected to do the paperwork on behalf of the firms. At the same time, smaller firms were given voluminous paperwork (along with unrealistic deadlines for completing it) in the hopes of discouraging applications from independent firms in general.

Strategies for Independence

In all aspects of doing business, innovating firms have sought to avoid any significant links with keiretsu big business. In terms of how these firms went about avoiding keiretsification, there are four main strategies of independence. First and foremost, they sought to maintain a balance in sales ratios among their clients. Second, they engaged in joint R&D with other SMEs. Third, they made international connections via joint ventures and by employing foreign technologists in-house. Fourth, they engaged in internet-based networking. These strategies helped firms to subvert traditional mediating structures. Other strategies having mixed success included creating local business networks and participating in independent national level (SME) organizations.

In managing the sales ratios among their customers, few of the firms interviewed allowed the proportion of total sales to a single buyer to be more than 40 percent. Most stated that if possible, they tried to reduce ratios for large customers to around 15 percent. In some cases, firms have achieved these sales ratios by selling directly abroad to foreign, often U.S., firms, thereby circumventing traditional mediating agents, such as Japanese trading firms.

In 1998, about three quarters of the firms I interviewed were engaged in research and development of new products. To survive and prosper in high-technology industries, firms are under pressure to keep moving into new technologies. Although many were engaged in new product R&D, a much smaller proportion had significant success in terms of holding high-value-added patents and moving into new technologies. Of the firms engaged in new-product research and development, 60 percent maintained external links supporting in-house R&D with other SMEs, including foreign firms. The firms were generally satisfied with the results of these collaborations. Other relationships included government research institutes (43 percent), university faculty (41 percent), and private research institutes (25 percent). Few expressed satisfaction with these relationships, citing a lack of market expertise in government institutes and universities, and the high cost of private research institutes relative to output. Firms were least likely to forge a relationship with a parent firm (a main buyer) for R&D purposes. (Of the 43 firms, 8 considered themselves to have a parent, while a total of 16, or 37 percent, had a degree of keiretsu linkage either as a subcontractor or in terms of sales. The latter would often consider its main buyer a 'parent' of sorts.)

In addition to technological ties with foreign firms via joint R&D, innovating Japanese firms have forged joint business ties with SMEs in other countries. M firm, which manufactures precision fiber optic devices for computer applications, forged ties with SMEs in other countries early on. Like other independent, innovative SMEs, the president of M was unable to get assistance from any government agency or private financial institution in Japan. Using one million yen of his own money, M began producing lighting equipment in the 1970s. Establishing ties with a sensory equipment manufacturer in the U.S., and a producer of lighting equipment in Taiwan, M firm has been able to grow based on sales made through these firms.

T firm, a producer of semiconductor machinery, and among the most successful of the independent SMEs I interviewed, established international ties in two ways. First, T firm actively sought out foreign technologists to employ in-house. Currently, 10 percent of T firm's workforce is foreign, the bulk of them technologists. T firm also employs foreign legal and management staff from the U.S., South East Asia, and Eastern Europe. Second, T firm has forged strong sales ties with major U.S. firms. In the beginning Japanese assemblers rejected T firm's products because T firm refused to enter into exclusive sales relations with any keiretsu group. After three major U.S. high technology firms purchased T firm's machinery, however, a few Japanese firms came around and made purchases. Now serving the U.S., European, and Japanese high technology markets, T firm has been able to amass the funds to establish its own state-of-the-art R&D Center.

In recent years, firms have been able to circumvent traditional market structures and limited access to information through the use of internet-based networking. Three examples include the Kiseiren network in Kyoto, TOPS in Higashi Osaka, and OtaNet in Ota ward. These networks have allowed member firms to circumvent traditional information channels and avoid mediating agents by exchanging market and technical information directly with other SMEs.

Innovative Leaders and Failures

The 43 firms I studied can be divided into four groups: innovative leaders (9), moderately successful (22), mixed successes (8), and innovative failures (4). Firms considered innovative leaders were active in producing new products in growth sectors such as fiber optics and medical machinery. The value-added level of these products was high, and firms were currently engaged in R&D for additional products. A standard definition of "value-added" defines it as the value of a firm's output minus the value of inputs bought from other firms. These innovative leaders found it easier in 1998, despite the recession-plagued Japanese economy, to do one or more of the following: engage in new product R&D, obtain skilled workers, and seek new clients. The most innovative firms were the least likely to have any kind of keiretsu link.

Firms considered moderately successful at innovating were less likely than "lead" firms to be producing final products in growth sectors. The products of these moderately successful firms had an average value-added level. These firms may not have been currently conducting new product R&D, and they found it neither easier nor more difficult to seek new clients, engage in research and development, or obtain skilled workers.

Firms having a mixed experience with innovation, were unlikely to be currently engaged in the production of goods for growth industries (although these firms may have been active in earlier growth sectors, such as semiconductors). Products of "mixed" firms had average to lower value added levels. These firms generally were not engaged in current research and development or they were finding it increasingly difficult to do so. Mixed firms were likely to find no change or a decline in successfully seeking new clients or obtaining skilled workers.

Firms that had failed to innovate were not engaged in the production of growth sector products. Such firms held few patents, and those held tended to be of a low value added level. The firms were not engaged in R&D and had found it more difficult to seek new clients and obtain skilled workers in recent years. They were also most likely to have a strong keiretsu link across the board (i.e., have a parent, be a fully owned subsidiary of that parent, and have 70 percent or more of total sales going to one client).

As an example of an innovative leader, S firm was formed in 1975 by four people who at the time were working for a firm that was going bankrupt. The president of S had studied electronics and had the idea of starting a manufacturing firm. With a modest investment of less than ten thousand dollars, the four started operations. For many years after starting up, S firm found it difficult to obtain funds from financial institutions and government programs. Now S firm has become an established producer of factory automation systems, employs 80 people and produces high value-added computer peripheral products. In recent years, S firm has found it easier to attract skilled workers, and significantly easier to purchase new equipment. Twenty-five percent of S firm's sales goes to its top client. Like other technology leaders in this study, S firm has avoided becoming an exclusive subcontractor for any large firm.

The president of S firm noted that the financial system in Japan is weak in providing venture capital to new businesses because the system as a whole lacks the know-how needed to assess the viability of new ventures. At the same time, S finds that capital and other resources for new business remains concentrated in large firms as a result of biased government policies. Instead of getting better, as some MITI proponents have argued, S believes that "over time, the relationship between the large and the small has become increasingly unfair."

R firm was founded in the mid-1960s and produces connectors and cables for personal computers, mainframes, and aerospace equipment. It has 10 employees. The highest percentage of sales going to one client is 10 percent. The president of R firm said that although small firms can obtain patents, if they do not obtain patents on all of the surrounding (shuuhen) technology, a given patent is useless. "Small firms cannot afford to pay the fees, and a large firm inevitably comes in and gets the patents, because they have the assets to do so." R firm is a part of a local business network, but finds that compared to R firm's customers (and itself) other network members are not very high-tech.

D firm is an example of an innovative leader in Kyoto. It was formed in 1979 and produces semiconductors and thin film machinery. The founder of D worked for some years for a Silicon Valley research institute before returning to Kyoto to begin D firm in his garage with a few friends. D's founder and president says that from the beginning he forged strong relationships with several universities, and now commits 10 percent of sales revenue to joint R&D firm was able to utilize the machinery at these universities and obtain assistance from students, so that initially no significant capital was needed. D's president says that from the beginning, his firm sought to avoid an exclusive sales relationship with any keiretsu group. Currently less than 10 percent of sales goes to its top client. D firm plans to go public by around the year 2000. D firm is a member of several local networks, and finds them helpful in the exchange of information and studying new technologies.

E Firm is an example of a leader in Higashi Osaka. It was formed in 1945 and initially produced steel parts, screws, and so forth. E firm benefited from the jump in procurement of steel parts during the Korean War and was able to begin upgrading its machinery. E firm now produces original machinery, communications cable, and thin film technologies. E firm has a cooperative research relationship with the Osaka area government, and other SMEs. Forty percent of E firm's sales goes to one client, while 55 percent goes to its top three.

An example of a moderately successful firm, B firm was founded in the 1950s, specializes in computerized control panels, and has 40 employees. B firm recently merged with another firm making similar products. With the merger, 39 percent of B firm's product now goes to a single customer. B firm has made it a priority to increase the relative product share of other clients in the future. The current president of B firm has had personal experience with a large firm expropriating SME technology when he was the former president of the firm that merged with B firm. According to B's president:

Years ago, a large firm came in (on the pretense of being a potential client) and copied a cable product after seeing it. They made a few minor changes and produced it themselves. We did not have a patent yet for the product, so we had no recourse.

Moderately successful C firm was founded in the 1930s and produces mainly integrated circuits (ICs). C firm employs 174 people and its highest percentage of sales going to one customer is 10 percent. C firm has tried to work with universities and other research institutions in the past on joint R&D, but it did not find these relations terribly helpful, so it ended the involvement. The president of C firm has not found local business networks to be useful, but instead accesses information on competitors and potential customers via an independent SME organization.

J firm has been in business for more than 30 years, and has remained a small, family-owned electronics parts producer. The president of J is an exclusive subcontractor for a large keiretsu electronics firm. Although J firm has been active in a producer network of about 10 firms in Ota ward, it has been unable to avoid being keiretsified over the years, and has been subject to "cost down" measures by its main client (where J firm's sales prices have been unilaterally reduced). Although the president of J firm agrees that SME-based associations are important, he says that SMEs in Japan lack a strong national organization to represent the interests of members. J's president noted that this is very different from what organizations such as Keidanren can do on behalf of big business. J's president says that SMEs have tried to create national-level organizations in the past but have been prevented by keiretsu pressure (danatsu sareta).

K firm, a moderately successful firm in Kyoto, was established in 1962 and employs 25 people. K specializes in the production of high tech medical machinery and 30 percent of its sales goes to its top client. K's president finds the keiretsification of the Japanese economy unfortunate but no surprise. He believes the suppression of individuality, especially in large firms (kosei o osaeteshiyou) has exacerbated the tendency of Japanese people to "lack strength of expression" and to allow themselves to be "caged-in" by organizational structures (tojikomeru). K's president says that the tendency for firms, especially large ones, to erect walls around themselves, within which employees are expected to make the firm the number one priority in their lives--over family, friends, and other social networks--discourages the free flow of information and ideas among all sizes of firms. Although K firm is finding it easier to attract skilled workers, the continuing credit crunch has made purchasing new equipment more difficult. K firm often uses the internet to access new customers, and its own web page has helped as well. K's president notes that "until recently, it was very difficult to access information, as large assemblers have had privileged access to information from government programs and industry organizations."

F Firm is a moderately successful innovator founded in 1963 in Higashi Osaka. It employs 220 people and produces sputtering machines used in thin film application technologies (e.g., for liquid crystal displays) and etching technology. F's president noted that although there are many networks in Higashi Osaka, the quality of personal networks in Kyoto is higher. "This is because in Kyoto networks are generally established and run by the owners/managers of small firms. In Higashi Osaka, many are created by bureaucrats, who have the best intentions but do not understand market needs." Although F firm has had many interactions with MITI, F's president finds that bureaucrats in MITI also lack an understanding of the market and the needs of SMEs. He gets the impression that MITI bureaucrats come around primarily to look for post-retirement amakudari posts.

Based on my interviews with 43 firms of various sizes, types, capitalization levels and types of product, the following patterns emerge:

  • The less keiretsu link, the higher new product and process innovation.

  • The higher the number and value-added level of new product innovations (patents held): the larger the customer base (and balance in proportion of total orders among these customers); the more likely that those innovations were financed using "in-house" capital; the more likely the majority of stock is held by the proprietor.

  • The greater the number of customers, the more likely a firm is to have historically participated in either an SME organization and/or local network.

  • The less "exclusive" buyer-seller relations, the more likely a firm is to participate actively in local business networks (rather than merely being dues-paying members of the Chamber of Commerce, for example).

  • The less "exclusive" buyer-seller relations, the less susceptible a firm is to "cost-down" measures.
About 25 percent of the firms I studied found their situations somewhat improved in recent years in terms of seeking new clients, employing skilled workers, making investments for R&D and purchasing new equipment.

A good question to ask at this point is whether technological independence came as a result of eschewing exclusive keiretsu (vertically integrated) ties, or whether, due to a firm's original technological independence, it was able to avoid being assimilated. While this is to some degree a "chicken-or-egg" question, several interviewees stated that in many a case of bankruptcy, the firm that went under had in the past been a technological leader, but had eventually been put out of business by its parent's (primary buyer's) cost down measures, as well as falling prey to abuses in the patent system.

Firms that have both survived and continued to innovate and obtain their own patents have successfully maneuvered around traditional intermediating hierarchies and utilized local and foreign business networks for information and market access. This has possible implications for would-be emulators in other countries, as the successful firms in this research represent the exception to the rule of doing business in Japan. SMEs that have survived and prospered have generally done so despite the structures that dominate the Japanese political economy.

KATHRYN C. IBATA-ARENS is a Ph.D. candidate in the Department of Political Science at Northwestern University. From 1997 to 1998, she was a Fulbright Research Fellow at the Research Center for Advanced Science and Technology, University of Tokyo. She did field research among the medium and smaller enterprises of the Kanto and Kansai regions of Japan from 1996 to 1998. This paper is abstracted from her doctoral dissertation.

Downloaded from