Software rights and Japan’s shift to an information society

The 1993-1994 copyright revision process

Joel West

Japan’s abortive 1993-94 efforts to reduce intellectual property protection for computer software highlight a tension between two visions of a technologically based Japanese economy: its industrialized past versus a postulated future “information society.” The postwar era was characterized by largely successful Japanese efforts to catch up to Western levels of industrial technology in order to achieve national autonomy. These efforts created world-class competitors and major exporters in a number of key industrial sectors, such as automobiles and electronics, which generated employment and technology spillover benefits throughout the economy. But Japanese firms have been less successful in developing computer software, the essential building block for the “information society” long postulated for 21 st century Japan. Recent discussions of “information highways” and “multimedia revolutions” have given such information change (joho-ka) an added sense of urgency for Japan’s government and business leaders.

The revision in computer software copyrights proposed in 1993 - allowing competitors to legally glean the key technical secrets of successful products through a process known as “decompilation” - was widely understood to be designed to help Japanese firms “catch up” once again in a technologically important industry with high spillover potential. At the same time, however, the long-term growth and success of indigenous software producers depend on strong intellectual property rights; compared to smokestack or even traditional high-tech industries, the development of these new information industries is far more dependent on the property rights regime assigned by the government.

The copyright revision process highlights a number of conflicts between Japan’s technological past and future. One is the tension between hardware producers and users, who want software easily imitated, and software developers, who want to preserve their unique designs; another is between those in government and industry who are focusing on “catch up” with U.S. software companies, and a few successful Japanese firms who are fighting the efforts of others (sometimes U.S. firms) to catch up with them. These conflicting interests did not play an important part in Japanese public discussion of the 1993 revision effort, which was dwarfed by a more familiar controversy - the vociferous intervention by the U.S. government and computer industry that eventually stalled the proposal in 1994. That it was shelved without clear resolution suggests that key economic and political issues inherent in joho-ka are also far from being resolved.

Software and an Information Economy

The most visible reason for abandonment of Japan’s contemplated copyright revision was the heated opposition by key U.S. companies, soon joined by U.S. trade representatives. This is not surprising, since the development and sale of prepackaged computer software remain one of the few high-technology areas where American firms are dominant, with an estimated 74% of the $72 billion annually in worldwide sales.(1) But this success is being challenged by other countries who note the industry’s 12% annual growth rate, its strategic importance in other industries, and the high ratio of labor to capital investment in software R&D.(2) As elsewhere in the computer sector, the Japanese are among the largest competitors to U.S. firms. Of the world’s ten largest suppliers of software, six are based in the U.S., one in Germany, and three in Japan.(3) Sales by Japanese producers have been primarily limited to their domestic market, but that was also the case in the early stages of other industries where Japan now enjoys a trade surplus with the U.S. and Europe.

American software producers have mobilized political support as part of overall U.S. efforts to protect copyrights and patents. There are two reasons why intellectual property protection has become a major trade issue for the United States. First, the sale and licensing of intellectual property is one of the few areas in which the U.S. has an advantage in the balance of payments, particularly with East Asia. Such industries account for more than $200 billion in GDP, five million jobs, and $39 billion in total export sales.(4) Second, products such as software, books, and compact disks have become more vulnerable in the last decade through technological developments in the products and their means of reproduction.

Intellectual property creators have fought to insulate themselves from two types of competition: exact duplication and imitation. For software products, the unauthorized production of exact duplicates is extremely easy - unlike computer hardware, the manufacturing process provides no barrier to competition. Like audio and video tapes, software is unusually vulnerable because a single copy of the product contains all the information needed for a “pirate” to make unauthorized copies, and these can be made using standard consumer equipment. This capability has been identified as a major U.S. trade concern for the past 10 years, as a 1987 GAO report noted:

Intellectual property pirates are often in a better position than legitimate producers to satisfy demand in newly industrialized countries since they generally enjoy lower production costs. Because pirates merely copy products rather than developing their own, their design and/or research and development costs are often minimal. . . . Moreover, because they copy only products with proven market success, pirates escape the cost of developing products that turn out to be market failures.(5)

The same advantages also apply to imitative products, which are usually constructed using reverse engineering techniques such as decompilation, and legal prohibitions against such imitation are neither widely accepted, readily defined, nor easily enforced.

Software’s economic role is by no means limited to direct sales of software packages; it has also gained a pervasive enabling role in our high-tech society. In the computer, electronics, and communications sectors, software linkages are crucial to the success of not only products but entire industries. The importance of state-of-the-art software to such industries cannot be overestimated, and when unobtainable by legitimate means, it is often stolen. When that is sanctioned by foreign governments, either by explicit governmental policy or through lax enforcement of nominal protections, Washington has held such competition to be an unfair trade practice and has used threats of trade sanctions to enact, strengthen, and enforce laws protecting copyrights and patents. The major focus of such efforts in the past decade has been in East Asia, where, according to one American industry estimate, 1994 U.S. trade losses due to unauthorized duplication totaled $4.4 billion, with $2.2 billion of that in software.(6)

Most countries have had to amend existing laws (usually copyright) to cover software specifically, although software has been covered in a few countries by judicial interpretation of existing laws. Adjusting the treatment of software under intellectual property laws is an important part of a nation’s technology policy, because these laws determine the right to profit from investment in developing a product, which in turn determines what products a company can afford to develop. Such policies are crucially important in the future direction of industrialized economies, particularly for Japan; the inevitability of an information society revolution has been postulated in Japan for more than 25 years, and elsewhere for nearly as long. Futurists have claimed that the switch from industry to information is as unstoppable as the shift from farms to factories two centuries ago. But such a shift assumes a property-rights regime that assigns private returns for the nontangible inventions sufficient to reward those who develop them; that assumption - made by neo-classical economists - is hardly borne out by history. As Douglass North noted, “Throughout history there has almost always been an immense gap between the private and the social returns to invention and innovation. . . . It has been more difficult to devise property rights over technological development than over products or resource inputs. The difficulty of measuring the dimensions of intellectual property and innovations and of enforcing any such property rights has been a basic reason for the divergence between private and social benefits.”(7) While the state’s interest is in generating the benefit to overall society, failure to assign adequate private returns means there will be few innovations - and thus few returns to allocate between public and private interests.

Globalization of Intellectual Property Issues

Years ago, the state’s allocation of social and private returns could be made within a national economy whose boundaries coincided with the state. But in today’s global economy, such a simplification is no longer valid. The most vivid example is the recent U.S.-China conflict over the copying of American products. The Clinton administration charged that the Chinese government ignored 29 factories producing 75 million counterfeit copies of American music, movies, and software, and with 95% of the output exported to other countries; in response to the estimated $1 billion in losses to U.S. firms, the administration threatened a like amount in punitive tariffs on Chinese goods.

Among the counterfeit products are a CD-ROM containing Microsoft Windows and more than 60 of the top application software packages, all sold in Hong Kong for about $15.

But despite the U.S’s leading role in pushing intellectual property trade issues, it is far from the only nation hurt by the absence of enforcement. The Nikkei Weekly reported that “Japanese firms are faring no better than U.S. firms in battling the problem of piracy of intellectual-property rights. From CDs to superhero dolls, Japanese companies are watching copycat operations walk away with millions in stolen profits.”(8) Japanese policy, however, has been more equivocal. Like Sega with video game players, Nintendo faced a serious problem with Chinese counterfeit manufacture of its game machines, but the Japanese government was reluctant to confront Beijing for fear of jeopardizing trade. Instead, Nintendo asked for help from the U.S. trade representative via its American subsidiary. At the same time that the United States was threatening sanctions to protect new-style information industries (including Nintendo), the president of Toyota visited China in hopes that the trade conflict would give his firm an advantage over U.S. rivals. Conflict over counterfeit entertainment and software is only the forerunner of a new style of trade conflict that will be common in the information century. And it may not be limited to trade: if military conflict in the 20th century could start over the nationalization of manufacturing or mining interests, the 21 st century could see it start over the unilateral nationalization of intellectual property.

Forms of legal protection. To protect their interests, intellectual property producers lobby for both strong laws and enforcement of laws against all forms of imitative products. As the dominant producers worldwide, U.S. software companies generally favor laws to restrict the development of such products, and regard the absence of such laws virtually as a form of theft. Discussions of software intellectual property laws usually focus on the types of restriction, usually one of three: (1) trade secrets, typically enforced by employment agreements and litigation against direct competitors; (2) copyright restrictions on software as a creative work, with protection of both the actual work itself and specific elements of its expression; and (3) patents, the most controversial mechanism for software legal protection, since it forbids differing expressions of the same fundamental idea. But such mechanisms reflect the precedents of Western-based legal systems, not inherent requirements per se for the protection of computer software; also, focusing on mechanisms misses the consequences of the success or failure of such restrictions. More important is the consideration of restrictions in terms of the various forms of imitation:

  • unauthorized duplication (piracy) - making exact copies of the original product.(9) From a legal standpoint, it is relatively easy to establish piracy and set a value for damages; in the United States, it is usually covered by copyright law.
  • imitative implementation (cloning) - the construction of another (often competing) product by analyzing the characteristics of the original and then reimplementing them in the imitation. To defend against this, software source code is tightly held as a trade secret, while copyright restrictions are used to control the use of reverse engineering.
  • imitation of fundamental concepts - a producer restricting a competitor from making a similar product, although this remains controversial even in the U.S. Examples of such restrictions are software patents and interface copyrights.

Reverse Engineering in Software Development

While software piracy limits revenues to producers, it can create brand loyalty so that the market share among pirated software sometimes translates into additional sales of legal copies. However, cloning also produces products that compete with an existing product. Successful cloning by rival developers reduces the market share of an existing developer, often reducing prices and increasing sales to the benefit of consumers. For software, such cloning is normally achieved by “reverse engineering” an existing product.

Because of the role of formal and de facto standards, the incentive for cloning software is stronger than for most other types of intellectual property. Although many products are distinguished by their newness, closely imitative computer products have an advantage in terms of interoperability with existing products, ease of user training, and compatibility with existing data. This interoperability provides strong incentives for imitation, often resulting in near-duplicates (clones) of popular products, particularly in national markets where domestic innovation is weak and there is a well-established market leader. Such cloning can be encouraged by companies seeking to establish a computer architecture as a widely accepted standard. For example, IBM established its PC as an open standard, allowing the creation of companies such as Compaq and Dell to manufacture rival PC clones; in 1994 the worldwide sales of one such clone (Compaq) even surpassed those of the original standard-setter. Where cloning is not sanctioned and encouraged, competitors may instead design their products by reverse engineering the existing product. This is common practice in software, although U.S. statutes and court rulings have held such practices to be illegal in some cases.

Reverse Engineering and Decompilation

If engineering is the process of turning a product design into an actual product, then reverse engineering means taking the actual product and deducing from it as much as possible about the original design. In software development, the most significant (and controversial) form of reverse engineering is decompilation or disassembly.(10) A commercially distributed version of the product is taken to recreate an approximation of the original source code. The role of source code and decompilation in software engineering can be better understood by way of analogy. Imagine a software product as a non-tangible machine, deriving value from taking inputs and producing outputs; the relationship between those inputs and outputs corresponds to a utilitarian pattern of value to a consumer - e.g., aggregating individual sales transactions to produce total sales and cost of goods. This machine is referred to in software law as “binary software,” i.e., the pattern of binary 1s and 0s interpreted by a computer as instructions that determine the operation of the program. Such binary software is not produced directly by programmers but instead is written in algorithmic pseudo-languages that use typewriter symbols to represent program rules in a quasi-human readable form. This textual representation is known as “source code,” akin to the plans used to build a machine. The job of a good programmer, therefore, is to craft a comprehensive and accurate plan for the machine (software product) within the limitations provided.

An unusual quality of software is that its plan is endogenous and complete. The translation from source code to binary software (compilation) is entirely automatic, by means of the computer. Hand any good programmer the source code to Microsoft’s best-selling product, and by the end of the week he or she will be producing machines (software products) that are just like Microsoft’s. A second software characteristic is that plans can be deduced from the original machine - i.e., a source code can be decompiled from the binary software. This translation is incomplete and lacks much of the context that makes the original source code so valuable, but the tools for preparing such decompilation have gotten progressively better.(11) The practical effect is that a skilled microcomputer programmer can go into a computer store, pay $400 for an original copy of a software product, and usually within a day or two describe some of the innermost secrets that the program’s authors took years to develop.(12) The process is essentially identical for all software-based products, even when the software is embedded as part of an electronics product, whether it be a mainframe computer, video game cartridge, or communications device.

Competing and Compatible Products

For a popular product, imitative software is quite common, and absent some protection of fundamental concepts, legal. Even if not legal, decompilation is often used to reduce the time needed to bring a product to market. It is usually possible to build competing products without decompilation, but the rate of change in the industry is so rapid that such products will fail if they take too long to develop. This distinction - whether it is technically (or practically) possible to make a product without decompilation - has been held by courts to be important in determining whether decompilation should be allowed. A distinction is usually drawn between two types of products developed using decompilation: (1) competing, or clone products, which are intended to supplant the existing product, and (2) compatible products, which supplement the existing products.

Although the decompilation process and usage are the same in both cases, the economic impact on the copyright owner is different. The clone will take sales away from the copyright owner, while the compatible product will not, and will sometimes help sales. This distinction has been held to be legally important, particularly where decompilation is justified as a fair use under U.S. copyright law. The practical differences are also significant. Decompilation is quite common in the U.S. software industry, where poorly documented operating system interfaces are understood by decompiling their implementation. When such decompilation is used to produce complementary products, it is usually in the business interests of the copyright owner to look the other way because such products benefit the copyright holder. However, where the decompilation produces competing products, such decompilation is often the basis for litigation against the competitor.

Existing Precedents

In the 1993-1994 Japanese copyright revision, three major legal precedents were cited by both sides of the issue:

  • In 1991, the European Community (EC) issued its “Directive for the Legal Protection for Computer Programs,” unifying software law throughout the EC. This directive explicitly permits decompilation to achieve interoperability.
  • In 1992, the 9th Circuit Court of Appeals in the U.S. ruled in Sega v. Accolade that despite a presumption that for-profit copying is an unfair use, Accolade could incorporate 25 bytes of decompiled Sega code in its games to make them compatible with Sega’s player.
  • Also in 1992, the U.S. Federal Circuit Court of Appeals argued in Atari v. Nintendo that some forms of decompilation constitute fair use; however, this precedent is generally considered less binding because it was decided against the company doing the reverse engineering (Atari) on other grounds.(13)

The status of these precedents and the lack of a “bright line” test for infringement suggest that this is an evolving area of law worldwide subject to future legislation and judicial interpretation.

Japan: Catching Up in Software

Despite early identification of information industries as an essential part of the country’s future, and numerous government efforts to promote such a shift, Japan has failed to establish a position in the world software industry commensurate with its leading role in other high-tech fields. The phrase joho-ka - usually translated by the quasi-English word “informatization” and denoting change to an information-oriented society (joho shakai) - has been a slogan of Japanese government policy for more than two decades even though its actual effect has been minimal. In the early 1960s, the phrase “information industry” (joho sangyo) was popularized by Tadao Umesao, while the later joho-ka is credited to Yujiro Hayashi of the Economic Planning Agency. The terms corresponded to the vision of a “post-industrial society” promoted in the United States by Daniel Bell, and resulted in an early information society boom in the late 1960s and early 1970s.

In his history of the Ministry of International Trade and Industry (MITI) Chalmers Johnson attributes a shift in the ministry’s policies away from heavy industries to its realization after the 1960s pollution crises that Japan could not continue its resource-intensive, high-speed growth indefinitely; software neither requires precious metals nor produces the toxic byproducts associated with electronics components and subassemblies. In May 1971, MITI formally amended its criteria for targeting industries to include an “overcrowding and environmental standard” and a “labor content standard.” The oil shock of 1973-74 brought home the country’s vulnerability as a resource-poor industrial nation, and Johnson places MITI’s first detailed vision of a “knowledge-intensive industrial structure” in November 1974. Certainly from the 1980s onward, Japan’s shift to an information society was repeatedly cited as a national goal, evident in Prime Minister Nakasone’s speech opening the Diet in February 1984, and it became the subject of various books, articles, and television programs.(14)

Lags in Packaged Software Development

Despite the desirability to shift from producing tangible (hard) to intangible (soft) goods, Japan’s role in the global information technology market has remained primarily in electronics components and peripherals, with a limited role in complete computer systems and a negligible role in software; by one calculation, the size of the information industry increased only from 3.1% to 4.0% of GDP in the period 1975-85.(15) Public policy debates on information technology are still dominated by considerations of manufacturing and selling hardware; this disproportionate political influence of the largest electronics hardware companies over smaller software-only firms would figure prominently in the 1993-94 copyright debate. In fact, Baba et al. note that most of the top software firms in Japan are spin-offs of computer makers or larger user companies (e.g., steelmakers, banks) and are still controlled by their parent companies, with independent firms holding less than 25% market share of the top 100 firms.(16)

Many of these captive firms focus on custom software development, targeted at their parent company’s specific needs, rather than general-purpose software products that can be sold to many firms. Nakahara attributes this to large variations in corporate business procedures, the scarcity of software engineers skilled enough to design general-purpose software, and the demand for custom solutions from Japanese mainframe computer makers.(17) Such reliance on custom software has hindered the widespread adoption of the software products - so called “packaged” or “shrink wrap” software - that have fueled the explosive worldwide growth of computer usage, particularly of personal computers. Meanwhile, personal computer acceptance in Japan still remains low (in 1993 per capita usage in Japan was 97 computers per 1,000 vs. 287 in the U.S., and even 125 in Singapore),(18) further limiting the largest potential market for packaged software. Only with the 1995 collapse of the dollar would Japan’s largest packaged software firm, JUST Systems, make it on a list of the world’s top 10 PC software companies as measured by sales volume.

Cottrell argues that fragmented PC standards made it less profitable for Japanese firms to develop packaged software. He attributes such fragmentation largely to Japanese language processing difficulties, slow sales of PCs that delayed pressures toward a single standard, and the dominance of mainframe computer companies that prevented the entry of entrepreneurial PC computer makers.(19) Various barriers to entrepreneurial firms in Japan must be considered in any evaluation of the Japanese software industry, as the largest PC and software firms in the U.S. began as entrepreneurial start-ups in the 1970s and 1980s. Cultural barriers to entrepreneurship include risk aversion. Hofstede lists Japan as the most risk-averse country among 39 studied, while the United States was among the least, ranking 35th out of 39.(20) This, in turn, is linked to the institution of lifetime employment for selected employees of major Japanese corporations, thus diminishing the talent pool of would-be entrepreneurs and their employees. Finally, the various obstacles to financing a company through bank loans, venture capital, or stock offerings make it more difficult for a new company to compete with established Japanese firms or start-ups in other countries.

Efforts to catch up with U.S. software technology. Japanese software industry development has thus been a major priority for the past two decades, stemming from two basic MITI goals: the shift in the early 1970s to “information” industries and MITI’s interest in enabling technologies. MITI-sponsored software R&D projects have included Sigma, TRON, and the Fifth Generation project, with total government funding during the 1980s of ¥84 billion. Software is also a key enabling technology for the computer and other industries. Software design skills are essential components in designing successful computer architecture (such as Microsoft Windows) that can accrue billions of dollars of profits to its owner over the course of a decade. Okimoto argues that MITI will be unwilling to concede U.S. dominance in computers because the industry holds the key to competitiveness in a wide range of industrial sectors, including automobiles, telecommunications, banking, and chemicals. Software has even been a factor in Japan’s security policy, delaying completion of its FS-X fighter. In the past two years, efforts to develop indigenous software technology have become more urgent as part of the drive to build Japan’s information highway.(21)

But MITI’s attempts to incubate the Japanese software industry as it did with consumer electronics and mainframe computers have been far less successful. The inability of the Japanese software industry to develop innovative new products has left its primary focus on understanding and imitating rival American software products. American companies alleged that such imitation sometimes involved illegal activities and, when such activities were discovered, efforts were made to change Japanese law to make them legal.

U.S.-Japan Conflict Over Software Intellectual Property Rights

The 1993-94 software copyright revision process, its potential impact on U.S. suppliers, and the intervention of American industry and government interests must be considered in the overall context of U.S.-Japan trade relations in regard to computer software. Computer software has been a trade issue only for the past two decades, and a major issue for half that. Significant sources of friction have included end-user software piracy, the means used by Japanese companies to design IBM-compatible mainframes, and the 1983-85 revision of copyright law to include computer software.

Japan’s Intellectual Property Philosophy

The U.S.-Japan conflict over software rights must be examined in the context of differences between the U.S. and the Japanese philosophy on intellectual property rights, often cited by Japanese government and industry representatives. For example, Swinyard et al. argue that Asian culture, as evidenced by the imitation inherent in Japanese calligraphy, is less supportive of proprietary copyright. In Japan, Womack notes that intellectual property law provides a comparatively narrow scope of protection, i.e., an imitative product can be more similar than under the U.S. system:

As a result, marketplace competition among Japanese firms is based on speed-to-market, finesse, and scale in production, and strong distribution systems, rather than knowledge creation.

This system meshes nicely with the Japanese philosophic premise that knowledge should be shared widely across industry to maximize social utility. It also fits the structure of the Japanese enterprise system, consisting of a few giant players in every sector supported by elaborate production and distribution networks.(22)

Since actual manufacturing (disk duplication and printing) is rarely a source of competitive advantage in software, such advantage comes down to a system of R&D vs. distribution strengths. In addition to the marketing investment common to all products, computer products usually require a high-cost training and support network. Recently, the trend has been to high-volume, low-cost distribution with even wider reach. For foreign firms selling to the Japanese computer market, this requires partnering with a Japanese firm - often one that manufactures competing products - for some or all of the distribution infrastructure. As Teece notes, firms that contract out for complementary assets run the risk that the “partner may imitate the innovator’s technology and attempt to compete with the innovator.(23) Facing such barriers to market entry, a foreign software developer usually has only the advantage of technological innovation; if that innovation is easily imitated, then the other disadvantages will assure failure. The only way that a foreign software product can successfully participate in a market is through intellectual property restrictions against imitative products, and with such restrictions determined by government policy, not the market. If government policy makers favor the interests of domestic producers over consumers - as has often been the case in Japan(24) - then foreign software producers are doubly disadvantaged.

Software Piracy in Japan

One source of ongoing U.S.-Japan trade friction has been the high rate of software piracy in Japan, primarily the unauthorized duplication by end-users. By one industry estimate, the Japanese piracy rate has been dropping, but at 67% was the highest of the G-7 nations in 1994. The industry estimate of $2 billion annually in lost software sales in Japan was second only to the $2.9 billion losses in the much larger U.S. market, where the piracy rate is only 35%.(25) Of such unauthorized copies, $1.1 billion in Japan represents one year’s losses by American software producers,(26) large enough to attract the attention of U.S. trade representatives. This piracy also diminishes the economic incentive for Japanese software producers to develop new products, particularly for the vast majority of firms developing only for the domestic market. Because of such copying and the proposed decompilation copyright revision, a coalition of U.S. copyright interests in early 1994 nominated Japan for “Priority Watch List” status under Special 301 legislation. No sanctions were imposed, but the Japanese agency responsible for administering copyright law responded with a public education plan “to try to shake Japan’s growing reputation as a software pirates’ paradise.”(27)

IBM’s Software and Industrial Espionage

The most notorious form of software imitation figuring in U.S.-Japan trade relations centers around the technology of IBM’s long-dominant mainframe computer product line. For years, Japanese makers of IBM-compatible mainframe computers purchased new IBM computers and reverse engineered the hardware and software in order to make better clones. Apparently, such reverse engineering was not enough because in 1982 two firms - Hitachi, Ltd., and Mitsubishi Electric Corporation - were caught in a joint IBM-FBI sting operation seeking to buy the secrets of IBM’s mainframe computer architecture. After admitting that it had agreed to pay more than $500,000 for confidential IBM information, Hitachi pied guilty in a U.S. district court to a charge of conspiring to transport stolen IBM computer secrets, and both companies settled with IBM. The largest IBM-compatible vendor, Fujitsu Ltd., was not a party to the industrial espionage suit because it obtained information through its long-standing investment in the largest American IBM-compatible maker, Amdahl Corp. But after settling with the other Japanese firms, IBM threatened Fujitsu with a copyright violation lawsuit over use of similar technology, so Fujitsu entered into a licensing agreement with IBM for the same technology.(28)

The IBM supai jiken (spy incident) was extremely controversial. American observers viewed it as industrial espionage; the Japanese reaction was initially that Hitachi and Mitsubishi had violated no laws, and once a guilty plea was stipulated, that such violations were due in part to ignorance of U.S. laws. But a contemporaneous Japanese report said the three companies “cannot ignore the possibility that such software may be in violation of relevant American copyright law and therefore, till now, IBM-compatible computer makers have refrained from marketing their operating systems in the United States.”(29)

In 1985 IBM’s settlement with Fujitsu fell apart, and by mutual agreement, resolution of their differences over IBM’s intellectual property rights was assigned to the American Arbitration Association. In 1988 a two-arbitrator panel required Fujitsu to pay IBM a lump-sum payment of $833 million. Fujitsu secured the right to obtain future IBM operating system software secrets (under tightly controlled conditions) in exchange for an additional annual license fee, initially set at $25-$51 million, depending on the information desired.(30) Although it obtained access to a wide range of compatibility information, Fujitsu nonetheless began a long-term campaign to reduce copyright restrictions on reverse engineering. This campaign included 1989 legal briefs in the European Community’s aforementioned efforts to unify its software laws. According to the account of an opposing IBM attorney, “Fujitsu had advised the EC officials that in addition to strengthening the interface exceptions that already appeared in the directive, a new exception should be added: the directive should explicitly permit ’reverse engineering’ of software. Here was a notion worthy of the preeminent software clone.”(31) Fujitsu won allies among a number of European and American companies, including Bull, Olivetti, NCR, and Unisys, forming a trade association called the European Committee for Interoperable Systems (ECIS) to push for a reverse engineering exception in the new copyright law.

The 1983-1985 Revision

Two American consultants who studied the IBM charge that the supai jiken was not an aberration but part of a pattern of stealing American software concluded:

[The Japanese] subjected all the larger American firms to massive intellectual property theft. The sordid dark-of-night stealing from IBM by Hitachi and Mitsubishi is only the most egregious example .... In an extraordinary exercise in cynicism, in 1984, after the exposure of widespread Japanese intellectual property theft, MITI introduced legislation that effectively legalized stealing from foreigners.(32)

This perceived tie-in with Japanese reverse engineering of IBM’s software is but one of the similarities between the 1983-85 and 1993-94 revisions of software copyright law. Other similarities include conflict between MITI and Ministry of Education objectives, complaints by U.S. trade officials over lack of openness in the policy-making process, and a stated goal of reducing the development costs for software imitators.

In the early 1980s, Japan’s copyright laws did not specifically mention computer programs. However, an advisory committee to the Agency for Cultural Affairs (ACA) had concluded in 1973 that computer programs were works of authorship entitled to copyright protection, and several Japanese video game manufacturers brought successful copyright infringement lawsuits against counterfeiters from 1982 to 1984. In 1983 MITI introduced plans for a unique “Program Rights Law” that would control legal protection of software without using copyright law. MITI began its push to revise software laws by consulting with its Industrial Structure Council advisory board - a month after the arrests in the Hitachi/Mitsubishi supai jiken. The link between the two events was made in the Japanese press:

A trial in the United States involving Japan’s Hitachi Ltd. accused of purchasing stolen International Business Machines (IBM) documents has produced an acute need for some sort of measures to legally protect software, although the ministry publicly disavows any relationship between the new law and the IBM espionage case.(33)

MITI’s proposal provided 15 years of protection, as compared to the 50 years in effect under the Berne Copyright Convention (which Japan adheres to) and 75 years for software protected under U.S. copyright law. MITI’s 1984 software plans were in line with those of the large, export-oriented companies, which include major customers for computer software systems as well as manufacturers of electronics equipment who depend on the availability of software. Its most controversial provision provided for compulsory licensing of software when “it is necessary for public interest or when it is not properly worked”(34) - referring to software that is written outside Japan but not freely licensed within the country. MITI’s plans were protested by U.S. officials, and the ACA within Japan’s Ministry of Education proposed a plan more favorable to the interests of Japanese and foreign software authors by amending copyright law to protect software for 50 years and without compulsory licensing. In the face of international pressure and threatened U.S. sanctions, MITI dropped its plan in March 1985, and the ACA plan prevailed.

The 1993-1994 Copyright Revision

Eight years later, many of the same issues reappeared when the Japanese government reexamined software copyright law. One key issue was the annual royalties paid by Fujitsu to IBM. In July 1993 Fujitsu announced it would not renew its annual license for compatibility data from IBM’s mainframe operating system, even though Fujitsu had felt that such information was worth millions of dollars in the past; industry analysts predicted this would hurt Fujitsu’s sales. The announcement came four weeks after published reports said the ACA planned a major revision of software copyright law, including studies of EC and U.S. laws. Fujitsu’s chairman later stated, “We want the government to allow us to pick up ideas from the software so we can develop products compatible with the other firm’s standards.”(35) Fujitsu was represented in both the ACA’s 16-member shingikai (collaborators’ council) and in an influential subcommittee of Keidanren (Federation of economic organizations).

Keidanren’s Proposed Changes and Shingikai Deliberations

Of concern to American companies was a report submitted to the ACA shingikai by the “Working Group on Legal Protection of Computer Programs,” a Keidanren standing subcommittee. Its 11-members included officials from the four major Japanese computer companies - Fujitsu, NEC, Hitachi, and Toshiba; IBM Japan; and representatives of large Japanese companies that are major computer users. In September the subcommittee considered software copyright and drafted a confidential proposal to the ACA, which was subsequently leaked to the public and widely distributed in both Japanese and an unofficial English translation. The proposal listed a series of majority and minority opinions, the latter assumed to be that of IBM Japan, and included three split recommendations regarding software decompilation: (1) copying and adaptation of computer programs for decompilation should be permitted; (2) decompilation should be made compulsory to prevent it from being restricted by software license terms; and (3) unlike the EC directive, copying and adapting programs for decompilation should not be limited to achieving interoperability, but should be allowed regardless of purpose to prevent the author from obstructing reverse engineering. Among the reasons cited for the provisions was the societal benefit of “preventing redundant investments in similar technologies.” This is similar to the 1983-85 revision in which the Information Industry Committee of the Industrial Structure Council “called on the government to promote software development and expand utilization of software by avoiding overlapping investment and promoting software distribution.”(36)

Throughout the 1993-94 revision process, the Keidanren proposal was used as the starting point for discussion by proponents and opponents of changing Japanese software law. An ACA official denied that the report was ever the proposal of his agency or its shingikai: instead, the ACA had initiated the revision to address legally the issue of “reverse engineering or decompilation,” citing the European and U.S. cases mentioned earlier.(37) As with the 1983-85 revision, rival factions clashed over conflicting goals. In the earlier revision, MITI pushed proposals aimed at promoting its constituents (Japanese high-tech companies) over foreign rivals, while the Ministry of Education (ACA) was more closely allied with the interests of its constituency (software authors); these factions were again represented in the 1993-94 process. The shingikai’s stated goal was to develop a final public report for the ACA; after receiving comments from the various ministries the agency would propose legislation for approval by the Diet. The shingikai was characterized by two of the traits considered typical for such deliberative councils: the use of a public-private discussion group to build consensus for the ministry and informalism (what trade rivals call nontransparency) to arrive at consensus. As with other shingikai, the informalism was manifested by nonpublic deliberations that allow clearance of controversial positions behind the scenes prior to any public commitment to a particular position.(38)

But such secret deliberations were a major source of friction with the U.S., and Commerce Secretary Ron Brown and Trade Representative Mickey Kantor drafted a November 2 letter to the ministers of education and foreign affairs expressing concern over the lack of U.S. input in the revision process. In response to U.S. protests, the shingikai held an unusual heating December 13 to take testimony from a series of speakers from American firms opposed to the proposed change, later supported by an attorney from the U.S. Patent and Trademark Office authorized to speak on behalf of the Clinton administration. Also testifying were two American industry representatives who favored the Keidanren proposal and sought to discredit the views of the other Americans.

American Opposition and Support

IBM led the opposition to the Keidanren proposals along with an industry group called the “Alliance to Promote Software Innovation” (APSI), which had been formed in 1991 to fight against decompilation in the European Commission proposal and is closely aligned with another industry association, the Business Software Alliance. For the APSI allies, innovation is the rallying cry to protect the lead held by a number of dominant suppliers of computer software and hardware. On IBM’s side were five of the six largest U.S. microcomputer software companies, as well as the leading developers of proprietary operating systems (IBM, Apple, Microsoft, DEC) and two major vendors of proprietary semiconductors (TI, Motorola). The APSI members said that allowing decompilation would eventually result in reduced financial rewards for innovation. They also claimed that reverse engineering could be used to construct less detectable copies of popular programs so that decompiled, modified, and recompiled code would become a common form of disguised plagiarism by competitors.

In addition to the letter by Brown and Kantor, the American opponents of decompilation were supported by visits from Ambassador Walter Mondale and patent and trademark officials. Testifying in support of the APSI position at the December heating, Christopher A. Meyer, a senior U.S. Patent and Trademark Office copyright attorney, argued that under U.S. law, the fair use copyright exception for decompilation would not be applicable to for-profit development of a directly competing product.

The Keidanren proposal, however, was supported by American interests organized as the American Committee for Interoperable Systems (ACIS) led by Sun Microsystems.(39) On Sun’s side were IBM’s mainframe competitors (NCR, Bull, Amdahl), peripherals companies (Emulex, Storage Technology, Seagate, Western Digital), semiconductor companies (AMD, Chips & Technologies, Phoenix), and three software companies (Accolade, Broderbund, and Informix). Sun’s leadership (and hence its role in AcIs) is driven by the philosophical zeal of its chair and cofounder, Scott McNealy, for so-called “open systems” - the idea that computer architectures should have open specifications allowing access by all vendors rather than proprietary specifications, which provide an advantage to the developer of that architecture. But just as many of the APSI members have a business interest in preventing imitative products, most of the ACIS members have a business interest in using reverse engineering to build such products. For example, Amdahl is one of the two leading U.S. providers of IBM clone mainframes and is part-owned by Fujitsu. Meanwhile, AMD is the most aggressive developer of substitute chips for Intel microprocessors, and Chips & Technologies and Phoenix sell chips and firmware for IBM PC clones. Accolade was the defendant in the case involving Sega-compatible video games.

ACIS argued that first-comers will lack incentive to innovate if they are protected from competition. Instead, consumer choice is threatened by “lock-in” to a product that wins market dominance because alternatives (clones) cannot be legally developed. If innovation is the mantra of APSI, interoperability is what drives ACIS. Such interoperability is defined both in its customary sense - complementary products such as application and peripherals - but also to include competing products that replace the original, the same position adopted by the Keidanren panel. Because the U.S. government intervened against the ACIS position, ACIS in turn sought to undercut whatever influence Washington might have with Tokyo. Thus, in testimony before the shingikai, the ACIS representative said that “decompilation is not properly a U.S.-Japan trade issue” and also argued that Meyer was not authorized to represent the U.S. government.

The presumption has been that, because U.S. firms generally lead their Japanese competitors in software development, all Japanese firms would favor loosening copyright restrictions. But this does not allow for those cases where a Japanese company with a strong product in its home market has more to lose at the hands of its domestic competitors than it has to gain against American ones. It also does not allow for those Japanese firms that dominate markets, such as the video game companies Nintendo and Sega. The two firms sued to prevent American firms from using decompilation to make unlicensed video games that compete with licensed games and those developed by Nintendo and Sega. Despite vigorous arguments in the United States against decompilation, both firms were conspicuously absent from the 1993-94 revision process, unwilling to publicly cross the position of MITI and the large electronics firms. Also absent from the discussion were representatives of the European Union, whose difficulties in dealing with the earlier APSI vs. ECIS duel over its own software directive made it unlikely that they would revisit the issue within Europe, let alone seek a common position in the Japanese debate.

Is Japan Ready for the Information Century?

Unable to agree on how computer software copyright law should be changed, the shingikai concluded its work in May 1994 with a series of options rather than specific recommendations. This led Fujitsu’s chairman to charge that “the Cultural Affairs Agency surrendered to pressure from the U.S.”(40) However, given Japan’s recent ability to say “no” to U.S. trade demands, one must also conclude that pro-copyright sentiments among Japanese panelists allied with the ACA played a significant role. The inconclusive ending to this revision effort suggests that Japan is far from resolving both the narrow issue - reverse engineering of computer software - and the broader political and economic changes inherent in its long-contemplated shift to an information economy. Such a shift is crucial to Japan’s technological competitiveness in the 21st century in which information technologies will play a role comparable to that of steam engines and railroads in the 19th century.

But despite the rhetoric of joho-ka, information technology policy continues to be dominated by the major Keidanren firms that want cheap software - that is, big electronics firms that use it to promote computer sales and large manufacturers who are the major users. The smaller, innovative software-only firms in Japan have yet to reach the influence of their U.S. cousins. Thus, in the near term Japan’s information technology policies will reflect a continuing “catch up” mentality, as evidenced by MITI’s abortive efforts in 1995 to create ISO-9000 quality certification procedures for packaged software. If such policies reduce intellectual property rights for software - more characteristic of a developing nation than the world’s second-largest economy - they will spark continuing trade conflict with the United States and with other nations largely sitting on the sidelines. Such an approach might be justified if it successfully inculcates a domestic information industry, but it also inherently carries the seeds of destruction of that industry. Societal attitudes toward intellectual property are less easily changed than a mere regulation, and enforcement of intellectual property rights depends as much on moral legitimacy as the enforceability of legal sanctions. An attitude that dismisses intellectual property rights within Japan would make it difficult for Japanese films to appropriate adequate domestic returns to support the technological innovation necessary for global competitiveness. At the same time, the experience of Japanese firms in China demonstrates that the export of Japan’s intellectual property depends on the global acceptance of such property rights, rights that the Japanese will be unable to promote until they are unequivocally established in their own country.


1. U.S. Department of Commerce, U.S. Industrial Outlook 1994, pp. 27-5.

2. The software industry has the highest rate of R&D spending (compared to sales) of all U.S. industries, according to an annual survey. See “What’s the Word in the Lab? Collaborate,” Business Week, June 27, 1994, pp. 78-103.

3. U.S. Industrial Outlook 1994. Note that the non-U.S. firms are all computer makers rather than software-only companies.

4. “Special 301 Report to U.S. Trade Representative,” International Intellectual Property Alliance (IIPA), February 13, 1995.

5. International Trade: Strengthening Worldwide Protection of Intellectual Property Rights, GAO/NSIAD-87-65 (Washington, D.C.: U.S. General Accounting Office, April 1987), p. 9.

6. “Special 301 Report,” February 13, 1995.

7. Douglass C. North, Structure and Change in Economic History (New York: Norton, 1981), p. 16.

8. “At the Mercy of China’s Pirates,” Nikkei Weekly, February 6, 1995.

9. The word “piracy” has judgmental connotations not present with “unauthorized duplication” but it is the most common term, even used by governments of countries where such duplication is common.

10. An “assembler” is a program that translates primitive, human-readable assembly-language computer programs into binary form, while a “compiler” translates more sophisticated higher-level languages (e.g., c++) into the same binary form. Thus a “dis-assembler” and a “de-compiler” are functionally equivalent, differing only in degree.

11. This refers to commercially sold tools, internal tools have existed a lot longer, and because of the obvious intellectual property issues, are unlikely to be publicly acknowledged.

12. This highly simplified discussion does not consider steps that a program author can consciously take to disguise program structure, not the practical problems of dealing with imperfect decompilations of 100,000 lines of code. The process should be compared to cryptography - as the technology gets better for creating ciphers, so, too does the technology for decoding them. The difference is that a primary purpose of a cipher is to subvert analysis, while for most software products, this is a secondary consideration.

13. See Charles McManis, “Intellectual Property Protection and Reverse Engineering of Computer Programs in the United States and the European Community,” High Technology Law Journal, 8:1; Sega Enterprises Ltd. v. Accolade, Inc., 977 F.2d 1520 (9th Cir., 1992); Atari Games Corp. v. Nintendo of America, Inc., 975 F.2d 832 (Fed. Cir., 1992).

14. Chalmers Johnson, MITI and the Japanese Miracle (Stanford, Calif.: Stanford University Press, 1982), pp. 289-301; Tessa Morris-Suzuki, Beyond Computopia: Information, Automation, and Democracy in Japan (London: Kegan Paul, 1988), p. 28.

15. Takeshi Hiromatsu and Gosei Ohira, “Information Technology and Japanese Economy: An Empirical Analysis on the Size of the Information Economy,” Working Paper no. 19 (Tokyo: University of Tokyo, Komaba Department of Social and International Relations, 1991), p. 2.

16. Yasunori Baba, Shinji Takai, and Yuji Mizuta, “The User-Driven Evolution of Japanese Software Industry: The Case of Customized Software for Mainframes,” in The International Computer Software Industry: Structure and Evolution, David Mowery, ed. (London: Oxford University Press, forthcoming).

17. Tetsushi Nakahara, The Industrial Organization and Information Structure of the Software Industry: A U.S.-Japan Comparison (Stanford, Calif.: Center for Economic Policy Research, 1993.)

18. Egil Juliussen and Karen Petska-Juliussen, Computer Industry 1994-95 Almanac (Austin Tex.: Reference Press, 1994), pp. 393-95.

19. Tom Cottrell, “Fragmented Standards and the Development of Japan’s Microcomputer Software Industry,” Research Policy, 23 (March 1994), pp. 143-74.

20. Geert Hofstede, Culture’s Consequences, abridged ed. (Newbury Park, Calif.: Sage, 1984), p. 122.

21. See Charles R. Morris and Charles H. Ferguson, “How Architecture Wins Technology Wars,” Harvard Business Review, March-April 1993, pp. 86-96; D.I. Okimoto, Between MITI and the Market (Stanford: Stanford University Press, 1989), pp. 28-29; Richard Samuels, Rich Nation, Strong Army: National Security and the Technological Transformation of Japan (Ithaca, N.Y.: Cornell University Press, 1994), p. 244; Joel West, “Building Japan’s Information Superhighway,” JPRI working paper no. 7 (Cardiff, Calif.: Japan Policy Research Institute, February 1995.)

22. James P. Womack, “Competing with Japan in Intellectual Property: A Strategic Overview” (Cambridge, Mass.: MIT, Japan Program, MITJP 92-02), p. 7.

23. David J. Teece, “Profiting from Technological Innovation: Implications for Integration, Collaboration, Licensing, and Public Policy,” in The Competitive Challenge: Strategies for Industrial Innovation and Renewal (Cambridge, Mass.: Ballinger, 1987), p. 201.

24. E.g., Japanese companies forced by MITI to buy inferior Japanese mainframe computers in the 1960s, as noted in Made Anchordoguy, Computers Inc.: Japan’s Challenge to IBM (Cambridge, Mass.: Harvard University Press, 1989), p. 33.

25. Piracy rates and losses as reported by the Business Software Alliance, April 1995. Another industry group, the Software Publishers Association, estimated 1994 piracy losses of business software at $1.3 billion for Japan but only $1 billion for the U.S., and piracy rates of 56% and 25%, respectively.

26. See IIPA, 1995.

27. “Culture Agency Issues School Guideline in Hope of Slowing Software Piracy,” Japan Digest, April 13, 1994.

28. Anchordoguy, p. 113; “Fujitsu Chief Is Confident of IBM-Compatible Computers,” Nihon Keizai Shimbun, December 6, 1993, p. 20.

29. Nihon Keizai Shimbun, editorial: “Trade Espionage,” November 1, 1993, p. 6; “Computer Spy Cases Ended but Problems Remain on Software,” JIJI Press Ticker Service, October 27, 1993.

30. Ilyse Barkan and Joseph L. Badaracco, Jr., “The IBM-Fujitsu Settlement,” Case 9-391-149 (Boston: Harvard Business School, 1991, revised 1/15/92); Anthony Lawrence Clapes, Software, Copyright, and Competition (New York: Quorum Books, 1989), p. 172.

31. Clapes, Softwars: The Legal Battles for Control of the Global Software Industry (Westport, Conn.: Quorum Books, 1993), pp. 127-28.

32. Charles H. Ferguson and Charles R. Morris, Computer Wars: The Fall of IBM and the Future of Global Technology (New York: Random House, 1994), paperback ed., p. 155.

33. “Japan to Enact Program Law to Protect Computer Software,” Kyodo News Service, December 9, 1983.

34. Teruo Doi, “The Role of Intellectual Property Law in Bilateral Licensing Transactions Between Japan and the United States,” in Law and Trade Issues of the Japanese Economy, Gary R. Saxonhouse and Kozo Yamamura, eds. (Seattle: University of Washington Press, 1986), p. 177.

35. “Agency to Overhaul Law Covering Software Copying,” Kyodo News Service, June 22, 1993; “Tokyo Hit for Copyright Law ‘Surrender’,” Nikkei Weekly, June 6, 1994.

36. “Japan to Enact Program Law to Protect Computer Software,” Kyodo News Service, December 9, 1983.

37. Interview with Misato Kitani, director of the International Copyright Office of the Agency for Cultural Affairs, January 12, 1994.

38. Frank Upham, Law and Social Change in Postwar Japan (Cambridge, Mass.: Harvard University Press, 1987), p. 168; David Litt et al., “Politics, Bureaucracies, and Financial Markets,” University of Pennsylvania Law Review, 139 (1990), pp. 370-71.

39. The hardware companies included Bull, NCR, and Unisys - all members of the aforementioned ECIS - as well as Amdahl.

40. “Tokyo Hit for Copyright Law ‘Surrender’,” Nikkei Weekly, June 6, 1994.

Joel West is a doctoral student at the Graduate School of Management, University of California at Irvine. The author wishes to thank Jason Dedrick and Steven Vogel for helpful comments on an earlier version of this article.

© 1995 by The Regents of the University of California

Citation: West, Joel, “Software rights and Japan’s shift to an information society: the 1993-1994 copyright revision process,” Asian Survey 35, 12 (Dec. 1995): 1118-1139