GOVERNMENT-SPONSORED COLLABORATIVE RESEARCH

TO PROMOTE INFORMATION TECHNOLOGY:

JAPAN'S CHALLENGE TO THE WEST?









Tim Ray

PREST,University of Manchester









1993”N3ŒŽ

March 1993















First Theory-Oriented Research Group

National Institute of Science and Technology Policy

Science and Technology Agency



Contents



	Page



Section 1	Introduction 	3



Section 2	An Overview of the Issues 	9



Section 3		Government Policies to Support 

Information Technology	20



	3.2 	Towards a Classification of IT Policies 	24



Section 4	Japanese IT Policy and the Use of 

Engineering Research Associations 	29



	4.1 	A Review of MITI's computer-related 

Engineering Research Associations 	35



	4.2	A note on Collaborative Software 

Projects 	46



Section 5	The UK Alvey Programme: A Case-study 

of a European Response to Japan's 

Fifth Generation Computer Project 	48



Section 6	The Rise of Collaboration in the 

United States 	54



Section 7	Conclusion 	59



		References 	63





Abstract 



To many Western observers, it has seemed that collaborative 

research schemes organised by Japan's Ministry of International 

Trade and Industry (MITI) have played an important role in the 

country's rapid development of its computer and semiconductor 

component industries. Japan's 1981 announcement of its Fifth 

Generation Computer Systems initiative prompted a number of 

Western attempts to match Japan's competitive performance by 

using "Japanese-style" collaboration to support research in the 

various constituent areas of information technology (IT). 

However, there is evidence to suggest that many of these Western 

schemes failed to take full account of a number of special 

features associated with the Japanese environment.



The present paper considers why governments intervene to support 

IT and presents a classification of different types of support 

policies. This model is used to assess the changing role that 

collaboration has played within the development of Japan's 

computer and related industries. Comparisons with Western 

versions of collaboration are then used to help identify factors 

that affect the relative success of collaborative research 

projects as mechanisms for promoting increased competitiveness. 



A central message that follows from these comparisons is that the 

use of collaboration as a "market modifying mechanism", should 

take appropriate account of the complex nature of market 

structures and the ways in which these structures vary between 

different national environments. Japan's collaborative schemes 

evolved gradually and their development has been shaped by many 

factors that are specific to Japan. If these factors are 

ignored, there is a strong possibility that attempts to 

transplant the collaboration model to different national 

environments will suffer from adaption problems and a failure to 

function in the way that was intended. 



Section 1: Introduction

 

The motivation behind the present paper is derived from the 

author's experience with the evaluation of a major government-

sponsored initiative which was designed to promote Britain's 

competitiveness in the information technology (IT) sector by 

supporting "Japanese-style", " pre-competitive" collaborative 

research. This five-year British initiative, known as the "Alvey 

Programme", was a direct response to Japan's 1981 announcement 

of its government-sponsored collaborative project to develop 

"fifth generation" computer systems. At that time, the apparent 

success of previous initiatives sponsored by Japan's Ministry of 

International Trade and Industry (MITI) had given the concept of 

"collaborative research" a high degree of credibility in the 

West.



During the 1980s, collaborative research experienced a rapid 

growth in popularity among Western countries struggling to 

maintain a competitive position in evolving IT markets. Yet, while 

many of these schemes have made important contributions to 

the "national innovation systems" of the countries concerned, 

improvements in competitive advantage have often failed to match 

Western perceptions of the achievements made by Japanese 

initiatives. With this problem in mind, the present discussion 

considers the nature of MITI's collaborative IT initiatives and 

the way in which they functioned in the Japanese environment. 

The argument is based on interviews with members of Japanese 

firms, government officials, researchers that have been involved 

with MITI's collaborative research schemes and published sources. 

On the basis of this information, it is proposed that a deeper 

understanding of Japan's research initiatives provides a useful 

basis for re-appraising attempts to imitate these policies in the 

West. The West's use of Japanese-style collaboration to compete 

with Japan should take careful account of the context in which 

Japanese schemes were developed and use this as the basis for 

considering how a similar initiative might interact with the 

national innovation system in question. 



With regard to assessing the nature of Japanese-style 

collaboration as a "policy challenge" to the West, Alvey is of 

special interest in that it embodied key features that were 

subsequently adopted by a number of other attempts, in both 

Europe and the United States, to use collaborative research to 

pursue enhanced competitiveness. It is used as here as a case-

study of a government-sponsored European response to Japan's 

Fifth Generation Computer Programme. Mention is also made of 

the rise of collaborative research in the US since the early-1980s. 

The development of these US initiatives may be traced from 

private sector collaborative schemes which were instrumental in 

securing a relaxation of the Anti-Trust Laws: thereby paving the 

way for an expansion of private sector ventures and the launch 

of federally-funded collaborative ventures.



One message that appears to follow from attempts to imitate 

Japan's collaborative schemes is that collaboration, as a "market 

modifying mechanism", should take account of the complex nature 

of market structures and, also, the fact that there are 

significant variations in the way that these structures operate 

in different national environments. If the West fails to 

recognise the significance of contextual factors that have shaped 

the evolutionary development of Japan' s collaborative schemes, 

there is a strong possibility that attempts to transplant this 

collaboration model to different national environments will 

suffer from adaption problems and a failure to function in the 

way that was intended. The following quotation relating to a 

failure by the West to understand the Japanese corporation, or " 

kaisha" provides a convenient analogy for summarising the 

shortcomings that can be associated with many Western 

perceptions of Japanese collaborative research schemes. 



But now there is a danger of the strengths of the kaisha 

being overestimated just as these companies were for so 

long underestimated. Some Japanese firms do well; others 

do badly. The strengths and tactics of the kaisha need not 

be surprising to those who analyze their behaviour and who



develop plans for effective competitive response. 



(Abegglen and Stork: 1985 p17) 





A central theme of the present paper is that, in the early-1980s, 

the strengths of Japanese collaborative research schemes seem to 

have been overestimated in the same way that they were previously 

underestimated. In reality, Japanese schemes have included 

failures as well as successes and their strengths and tactics are 

best understood by looking closely at how they have functioned 

within the context of Japanese industrial development in the IT 

sector. 



The case of Alvey can be used to illustrate a scheme which was 

largely justified in terms of enhancing UK competitiveness, but 

produced achievements were seen in mainly in terms meeting 

technological research targets and a "structural" broadening of 

the UK research base, which was supported by the formation of 

extensive "networks" within the IT community (Guy, Georghiou, Ray 

et al: 1991 pii-iii ). This gave the UK IT research base 

increased coherence and created new channels of communication 

between industry and academia. Nevertheless, the evaluators 

final report on the initiative noted that the intended boost to 

national competitiveness did not occur on the scale that was 

hoped for: 



The major disappointment (of the programme) concerned the 

goal of enhancing competitiveness. All the major 

indicators of the programme show declining competitiveness 

in the UK IT sector during the lifetime of the programme; 

market shares declined and major firms passed wholly or 

partially into foreign ownership. 



(Guy, Georghiou, Ray et al: 1991 p xviii)



The apparent failure of British industry to capitalise on the 

research outputs produced under the Alvey Programme raises



questions about the extent to which it is realistic to expect 

that pre-competitive collaborative research will promote national 

competitiveness. Prior to Japan's Fifth Generation announcement, 

the competitive instincts of Western companies tended to restrict 

collaborative research to a very limited range of circumstances 

where the advantage of sharing the costs and uncertainties of 

research outweighed the disadvantage of having to share that 

knowledge with other parties. The European system of Industrial 

Research Associations provides an example of research funded 

jointly, by firms and the governments, which has been described 

as "worthy" but "hardly more exciting than the work of the local 

Post Office" (Woodward: 1965 p38). As the Director of a well 

known British Research Association remarked: 



Sir, running a Research Association is like a dog's walking 

on its hinder legs. It is not done well but you are 

surprised to find it done at all. 

(Woodward: 1965 p39) 



This image of collaborative research as a rather peripheral 

activity makes the West's apparent conversion to the paradoxical 

principle of "collaborating to compete" all the more remarkable. 

Moreover, even though there is evidence to suggest that the 

extent of increased competitiveness brought about by these 

schemes is questionable, pre-competitive collaborative research 

has become more prominent in Europe and the United States. 





Structure of the Paper 



Section 2 considers the significance of Japan's announcement of 

its Fifth Generation System Computer Systems Project and presents 

an overview of the principal issues which are addressed in 

subsequent sections. These cover three main areas: general 

perspectives on IT support policies; a case-history of 

government-orchestrated collaboration in the Japanese IT 

industry, and; Western responses to the "Japanese Model". 



The perspectives on government policies to support IT presented 

in Section 3 serve to locate collaborative research within a 

broader range of policy options. This subject is approached by 

considering why governments should support IT and the nature of 

different types of IT initiatives. The section closes with a 

classification of schemes. 



Section 4 focuses on Japan's development of policy initiatives 

to support IT. It shows how collaboration emerged as only one 

of four basic types of IT support policies. MITI's schemes in 

the computer hardware and semiconductor components sectors are 

reviewed to identify various steps in the process by which 

Japanese firms gradually closed the gap with IBM and other 

leading Western producers. It is argued, that while government-

sponsored collaboration has provided a mechanism for accelerating 

the pace at which Japanese firms could "capture" and exploit 

best-practice technology, this was done in a distinctly 

competitive (as opposed to cooperative) environment. Japanese 

firms preserved their competitive identities and generally 

managed to avoid sharing knowledge associated with near-market 

technology. In addition to competition between firms, it is 

noted that there were conflicts between the industrial ambitions 

held by companies and MITI's view of its policy objectives; not 

to mention rivalry between MITI and other ministries. MITI's 

Fifth Generation project is then contrasted with previous schemes 

as a "new direction" to the established trajectory of 

collaborative research ventures. Finally, Section 4.2 contrasts 

the success of hardware support schemes to the much more 

problematic history that has been associated with collaborative 

research in software. 



In section 5, the Alvey Programme is used as an example of a 

leading European response to the Fifth Generation "challenge". 

By contrast with the preceding Japanese schemes, this was mainly 

focused on research that was some distance from the market and 

was designed on the assumption that firms would engage in the 

sharing and joint production of knowledge. It featured a major



input from the academic sector and served to consolidate and 

restructure the national IT knowledge base through the creation of 

new communication networks and an expansion in the number of 

researchers. 



A perspective on the rise of collaboration in the US is provided 

in Section 6. This provides a further dimension to the picture 

in that pressure for increased collaboration may be traced from 

the private sector. A relaxation of the country's Anti-Trust 

Legislation, which had been a major distinguishing feature of the 

US "national innovation system" followed and there have since 

been federally-funded collaborative schemes to promote 

competitiveness. 



Finally, Section 7 presents conclusions that follow from the 

study. These serve to highlight the complex nature of market 

structures. While national innovation systems are generally 

concerned with translating knowledge into commercially 

exploitable technology, there are significant variations in the 

way in which they do this. The use of collaboration as a "market 

modifying mechanism" should take careful account of these 

variations. Japanese-style collaboration emerged in the Japanese 

environment and was shaped by a number of factors that are 

specific to that environment. 





2 An overview of the Issues 



During the 1970s, developments in computer and communication 

technologies converged to create a new arena of competition: 

information technology (IT). The combination of computing power 

with the message-sending capability of telecommunications 

redefined traditional boundaries between the gathering, 

processing and distribution of information. At the same time, 

advances in microelectronic technology paved the way for 

spectacular improvements in the performance-to-cost ratio of 

computing, thereby fuelling the rapid expansion of IT markets and 

the diffusion of the technology into new areas of economic 

activity. The pervasive nature of IT helped to ensure that it was 

quick to emerge as an essential ingredient in the economic 

development of industrial economies. 



Japan's government policies to support its computer and 

semiconductor industries during the 1960s and 1970s have often 

been cited as an important element in its ability to exploit 

opportunities created by the " IT revolution". As the 1980s 

unfolded, many countries looked to Japan as a prime example of 

the apparent benefits that could be derived from pursuing 

appropriate policies to support the constituent elements of IT. 

There was a growing realisation that a vibrant IT sector embodies 

enormous potential for stimulating significant growth in 

employment, together with enhanced overall competitiveness 

arising from performance improvements in a whole host of 

downstream user industries. A number of governments also view 

access to state-of-art IT as being of crucial importance to the 

maintenance of an effective defence capability. To a greater or 

lesser extent, governments have been forced to take note of IT and 

a proliferation of policies to promote national capabilities in this 

sector has followed. 



Western awareness of Japan's role as a leading exponent of 

effective IT policies increased dramatically during the late-

1970s. At that time, there were signs that Japan had started to



draw level with the United States in certain areas of computer 

hardware and semiconductor component technologies. To many 

Western observers, it seemed that close cooperation between the 

state and private industry had fostered the development of a 

highly-effective "national innovation system". Respect for the 

apparent achievements of this system was a major factor in 

causing Japan's 1981 announcement of its Fifth Generation 

Computer Project to send shock waves of concern reverberating 

around competitor IT industries. It seemed that Japan's national 

innovation system was about to be trained on a new target: 

advanced research.



Whereas previous policies had been concerned mainly with 

transferring existing technology to Japanese firms, progress 

towards "fifth generation" computing was to be research-driven. 

The project was directed towards uncharted territory, beyond the 

prevailing forefront of international best practice research. 

Fifth generation computers were intended to understand spoken 

instructions, emulate human reasoning and explain how conclusions 

are reached: thereby providing a practical realisation of 

sophisticated artificial intelligence (AI). The programme was 

regarded widely as a bid by Japan to lead international best-

practice technology in this sector. Although estimates of the 

ultimate tractability of this ground involved considerable 

uncertainty, the West's fear was Japan might be able to use its 

"national innovation system" to create an ability in an area of 

AI which had suffered a credibility crisis in the West caused, 

in part, by the over optimistic claims of enthusiasts for the 

discipline. Irrespective of the its eventual outcome, the fact 

that Japan was prepared to launch such an initiative was 

acknowledged widely as being a clear signal that the country had 

reached a position where it could mount a serious offensive 

on the very frontier of international best-practice, advanced IT 

research. The Fifth Generation programme was used extensively 

by Japan's competitors as a justification for adopting a more 

interventionist approach to IT policy. Ferné has noted that some 

20 other national programmes followed in the wake of the Fifth



Generation announcement (Ferné: 1989 p10).



Although Japan had implemented a range of different policies to 

support different aspects of IT, the model that captured the 

West's attention was the system of Engineering Research 

Associations (ERAs), organised by the Ministry of International 

Trade and Industry (MITI). These ERAs, which were initiated in 

the early-1960s, involve government support for "horizontal" 

collaborative groupings that allow competing firms to pursue a 

specified agenda of research on a temporary basis. During the 

late-1970s, the widely-acclaimed success of collaborative 

initiatives, such as the Very Large Scale Integration (VLSI) 

project, helped to reinforce the idea that MITI's ERAs had a 

major role to play in helping Japanese firms to become more 

competitive. It seemed as if the government had conspired with 

industry to exploit the paradoxical notion of "collaborating to 

compete". Japanese firms appeared to have become competitive by 

colluding in government-sponsored initiatives to suspend the 

concept of competition during the development phases of 

technology life-cycles. Arguments to the effect that this sort 

of policy would only bring short-term gains became hard to 

sustain in the face of Japan's record of continued increases in 

its share of world markets for semiconductor components and 

computer hardware. There was growing support in the West for the 

diagnosis that Japanese collaborative research "worked", although 

details of the design and implementation of these schemes 

received rather less attention. While the exact extent of MITI's 

contribution to increased industrial competitiveness is a matter 

for debate, it did appear that MITI's "vision" for the computer 

and semiconductor components sectors had coincided with dramatic 

improvements in "revealed industrial performance". MITI's 

computer-related ERAs had become synonymous with success. The 

fact that the Fifth Generation project was also going to be 

collaborative raised the spectre of extending this momentum of 

sustained progress into a new area of basic research.





Suddenly, it seemed that the West's view of collaborative 

research had changed. Instead of being seen as a rather 

peripheral activity, collaborative research became "fashionable" 

and emerged as a key ingredient in a number of European and US 

ventures to promote competitiveness. However, many Western-style 

attempts to imitate the perceived success of Japan's cooperative 

schemes ("fighting fire with fire") do not appear to have taken 

full account of several key points relating to the practical 

operation of "Japanese-style" collaboration. Little importance 

seemed to be attached to the fact that the structure and 

objectives of MITI's ERAs evolved gradually and were continually 

redefined as a consequence of a "push-me-pull-you" dialogue 

between the government and participating firms. There was no 

standard "blueprint" for designing ERAs. Over the years, 

modifications were introduced as part of an evolving process to 

correct past problems and react to changes in technological and 

market environments. Structural variations have included shifts 

in the respective level of government and industry funding; 

differences in the duration of projects and shifts in the 

mechanisms for knowledge sharing. At different times, the 

Japanese models of collaboration have varied between a clear 

division of labour, based on self-contained "modular" work 

programmes, and the joint creation of knowledge in a central 

research laboratory staffed by researchers seconded from 

collaborating firms. This has reflected a shift away from "near 

market", application-oriented schemes ( where commercial 

sensitivities necessitated a degree of modular organisation) and 

towards the more basic research of the type conducted under the 

Fifth Generation Computer Systems project.



Japan's collaborative schemes have evolved against the background 

of simultaneous shifts in the relative ability of Japanese firms 

to operate effectively in the global arena of competition. This, 

in turn, has had implications for the extent to which MITI can 

act as an "honest broker" and "legitimate" agent for mediating 

between rival firms. In this respect, it could be argued that 

improvements in the international competitiveness of Japanese



firms undermined the significance of MITI's role of providing 

support at the national level. At the same time, there does not 

seem to be any industrial consensus about the shape and direction 

of a "new role" for MITI. Catching-up with the West removed the 

"technological signposts" that had assisted in the formation of 

a government-industry consensus for the design of past, 

"application-oriented" schemes. Basic research is typically 

concerned with fostering long-term "creativity" and is often 

prone to drift away from the more rigorous disciplines that are 

imposed by firms' "current commercial commitments". Companies 

that are actively involved in the conduct and exploitation of 

basic research find it hard to generate accurate "visions" for 

the future. The need to cover the "technological waterfront" 

means that much of the research work will eventually lead to 

benefits that are only of limited or indirect value. It is 

difficult to produce sharply-focused research agendas that are 

capable of close correlation with the future evolution of "user 

requirements". Such a task is even more difficult for government 

agencies.



The question of the justification for governmental intervention 

to support IT is assessed in Section 3. Attention is drawn to 

the nature of government views on the "ideological acceptability" 

of support for collaborative research as part of an attempt to 

answer the question "why should governments be involved in IT?". 

It is argued that, while government support for basic research 

has long been justified on the grounds of "market failure", a 

case for supporting research as a means to enhance industrial 

competitiveness in IT started to become increasingly more 

prominent during the 1980s. Attempts to promote competitiveness 

through IT research lead on to the question of "what constitutes 

IT policy?". As IT has become more closely connected with 

economic growth, policies that affect IT have become more closely 

inter-twined with broader measures, designed to regulate trade 

and enhance competitiveness. The theoretical model proposed 

relates government-sponsored research and development to the 

overall range of policy options. Japan's development of



collaborative research and Western responses are considered 

against the background of this model in Sections 4, 5 and 6.



Japanese policies for the development of "targeted" industries 

have been mainly shaped by a brand of pragmatism which has a long 

history. The state has played a significant role in structuring 

economic development in Japan since the country started to 

industrialise in the late-1800s. At that time, the fear of 

colonisation by the United States or a European power helped to 

establish the principle of the state acting to manage the market 

economy. Intervention to manipulate the "invisible hand" of 

capitalism subsequently became accepted as a legitimate role for 

the state and has been an important feature in Japan's post 

Second World War development in a range of new industries. 

Little attention was given to ideologically-based concerns with 

limiting intervention to correcting deficiencies caused by market 

failure. Government support of "near market" applied development 

has been considered to be perfectly acceptable as long as it does 

not simply take the form of an unalloyed subsidy to a particular 

firm of industrial grouping. Rather, the philosophy has tended 

to reflect the idea that there should be sufficient competitive 

activity in the system to allow the recipient of the subsidy to 

digest its benefits without acquiring the economic equivalent of 

obesity. Competition should be constructive in stimulating rival 

firms to make progress without eliminating players from the game. 

In some instances, creating a healthy environment for innovation 

has represented the limit of policy intervention. For example, 

Japan's highly successful consumer electronics industry was 

established with support to create a favourable business 

environment using macroeconomic policies, generous tax provisions 

and compensation for deficiencies in the market mechanism 

(Okimoto : 1986 p549). The sector than prospered through Japanese 

firms' rapid assimilation and improvement of US technological 

advances, coupled with sustained capital investment, effective 

process engineering and aggressive marketing.



By contrast with consumer electronics, the computer and 

semiconductor industries were targeted for special attention. 

MITI's image of orchestrating collaborative projects which 

successfully enhanced the competitiveness of participating firms, 

was based on initiatives in applied rather than basic research. 

In this respect, pragmatism centred on a clear target: catching 

up with the West. Moreover, pragmatic catch-up projects tended 

to be near market ventures and were not really all that 

"collaborative" in terms of the joint production and sharing of 

knowledge. This modularisation enabled the benefits of so-called 

collaborative research projects to be "internalised" by 

participating firms without compromising the distinctive aspects 

of their technological knowledge bases. Firms did not have to 

share their proprietary knowledge and so retained elements of 

diversity in their approaches to innovation. This meant that 

rival firms were pursuing moving frontiers of best-practice 

technology from a range of different angles: thereby promoting 

the elements of technological diversity that drive the very 

process of competitive innovation. Nonaka and Yoneyama have 

argued that an imprecise understanding of the exact capabilities 

of rival firms operating in semiconductor component markets can 

lead to a redoubling of competitive efforts. A process which 

they refer to as "overshooting" leads to firms innovating to a 

greater extent than is necessary to capture market share at 

successive levels of technological development: thereby "bidding-

up" the currency of best practice technology in a virtuous cycle 

of innovation (Nonaka and Yoneyama: 1992).



Genuine knowledge sharing has tended to be confined to Japan's 

more basic collaborative research projects which do not have 

immediate implications for competitive performance. Moreover, 

the direct importance of these collaborative projects to the 

competitive performance of participating firms declined sharply 

once Japan started to catch up with best-practice IT technology. 

This is somewhat paradoxical in that the period of declining 

relevance has coincided with a dramatic expansion in the West's 

use of collaboration as a policy support mechanism to promote



competitiveness in the IT sector using "pre-competitive" (ie "far 

from market") knowledge-sharing projects. It was Japan's success 

in applied collaborative research that gave credibility to its 

Fifth Generation Computer Programme, even though the country had 

no track record in this type of basic-research-driven venture. 

Thus, it could be argued that many Western initiatives reacted 

in the wrong way to the wrong signal.



The subject of "lessons for the West" arising from Japan's 

experience with collaborative research draws heavily on the 

example of the UK government's £350 million "Alvey Programme for 

Advanced Information Technology" (1983-1988). Many features in 

the design of this programme can be traced to the recommendations 

of a committee established in 1982 under the chairmanship of John 

Alvey (who was then the Senior Director of Technology at British 

Telecom) to improve the competitiveness of the UK IT sector by 

supporting "pre-competitive" collaborative research. Much of the 

justification for this programme was based on arguments to 

mobilise national resources in the face of the challenge posed 

by Japan's Fifth Generation initiative. Alvey stands out as a 

notable example of intervention on the part of a Conservative 

Government, led by Margaret Thatcher, which was committed to 

minimising government intervention in industry. The sheer scale 

of the Alvey initiative was without precedent and the programme 

has been described as the largest single venture of its type ever 

attempted by the UK during peacetime. The case-study of the 

Alvey Programme is of particular interest because many of its 

features are reflected in approaches to collaboration have since 

become well established both in the UK and across Europe with 

programmes such as ESPRIT, Eureka and a host of similar 

initiatives.



A further important manifestation of collaborative research which 

can be related to the period following the launch of the Fifth 

Generation Programme can be seen in the case of the United 

States. This shift towards collaboration was all the more 

remarkable because any collusion of this type had previously been



prohibited by the antitrust laws which have been an important 

distinguishing feature of the US "innovation system". By 

contrast with Japan and Europe, the commercialisation of 

semiconductors and computers has also featured significant 

efforts on the past of new firms (Mowery: 1992). In this 

respect, the entrepreneurial flexibility of new firms can also 

mean that they lack the "broad gauge" central core of research 

that enables them to restructure their technological knowledge 

bases to meet the challenges of dynamic shifts in the prevailing 

technological and commercial environments.



A common feature of many Western models of collaboration is the 

use "pre-competitive" research projects as policy support 

mechanisms to underpin the competitiveness of domestic IT 

industries. At some point, the research should become 

"competitive" and erstwhile collaborators are expected to stop 

collaborating and start competing. One justification advanced 

for this model of collaboration is that it creates a form of 

"temporary monopoly" whereby the differences between rival firms 

that normally drive the engine of competitiveness are suspended. 

The aim of the policy is to combine the ability of a large 

organisation to devote resources to research, with the active 

rivalry of independent firms seeking to use exploitation to gain 

a competitive advantage.



Part of the case for arguing in favour of horizontal 

collaborative groupings hinges on the idea that "the whole is 

greater than the sum of the parts". The hope is that there will 

be a multiplier effect on government funding as project 

participants spur each other on to achieve greater technological 

progress than would have been possible if they had acted 

individually. Nevertheless, while collaboration suddenly became 

"fashionable" in the 1980s, this form of joint research is by no 

means new (the case of Industrial Research Associations was noted 

in Section 1). Collaborating to share the costs and risks of 

research has a long history although, in the past, the value of 

such ventures was generally thought to be limited to very



particular sets of circumstances, where the benefits of "low-cost 

access" to research results outweighed any potential commercial 

disadvantages associated with sharing that knowledge with other 

participating organisations. For this reason, firms tend to be 

most comfortable with projects that are at the basic end of the 

research spectrum. If horizontal collaborations do involve "near 

market" technologies, it is generally necessary to ensure that 

there good grounds for expecting that there will be an equitable 

distribution of benefits, as for example might occur in the 

development of an industry standard. Otherwise, so-called 

"natural" collaborations that involve "near market" technologies 

frequently centre on the type of non-horizontal links that exist 

between suppliers and customers or firms dealing with 

complementary technologies.



While the concept of "market failure" embraces a large number of 

potential deficiencies in the ability of competition to allocate 

resources in a desirable manner, the focus of the West's use of 

the term for justifying government-sponsored research lies with 

the unwillingness of firms to commit resources to longer-term 

projects where estimates of likely commercial returns are subject 

to high degree of uncertainty. This form of justification for 

intervention implies a primary concern with basic research areas 

which are outside the sphere of "natural" collaborations formed 

by firms for commercial reasons. Yet, support of far-from-

market-collaborations as a means to pursue competitiveness relies 

on a high degree of optimism about matching research agendas to 

evolving market needs. Unless sufficient attention is given to 

potential exploitation routes at the outset, there is a very real 

risk that outputs will fail to feed into emerging trajectories 

of commercial development (Ray: 1992). The fact that research 

is collaborative does not insulate it from any of the traditional 

arguments about the uncertainties that are linked to "technology 

push" research projects. Markets, competing technologies, 

changes in standards, legislation and exogenous shifts in 

relative prices can all change dramatically in the time that it 

takes to bring a technology-push project to fruition: thereby



reshaping the window of commercial opportunity that is available 

for exploiting the technology (Ray et al: 1989). At the same 

time, pursuing a research project collaboratively also introduces 

its own problems associated with costs of coordinating the work 

undertaken by different partners, fears related to the ownership 

of intellectual property and non-performance or withdrawal by 

partners during the life of the project.



Section 3: Government Policies to Support Information Technology



Reference has already been made to the "traditional" differences 

between Japanese and Western approaches to intervention. This 

section further explores more general aspects of the nature of 

intervention in the West and Japan. It concludes by proposing 

a "triangular representation" of different policy options. The 

triangle is defined by three basic axis of variation in the 

dimensions of government policies: (1) the extent to which the 

policies are either sector-specific or general (ie are they 

focused on a single industrial sector or do they affect a range 

of industries?); (2) the degree to which governments shape the 

direction and outcome of activities supported under 

interventionist initiatives, and; (3) the position of the 

policies on the spectrum between "supply side" and "demand side". 

In some respects, this could be thought of as a "four sided 

triangle", with time forming a fourth dimension to the image. 

The triangular model is introduced as a point of reference for 

subsequent sections relating to Japanese policy and Western 

responses. The thrust of the argument is that, overtime, Japan 

has pursued a relatively balanced mix of policy options within 

the "space" defined by the triangle of possibilities. While this 

mix includes collaboration, for various reasons the degree of 

commercial relevance associated with that option has declined 

since the mid-1980s. By contrast, the West has given a greater 

priority to collaboration during that period. As section 5 will 

argue, one consequence of this move has been to draw attention 

to the disappointing level of exploitation that can be associated 

with public support for pre-competitive, collaborative research. 

One response to the problem of improving exploitation has been 

to consider a further area of the policy triangle concerned with 

assisting the diffusion of IT products.



Western economic policies are usually coloured by the relevant 

government's approach to "free markets" and "market failure" to 

induce industrial investment in longer term research projects. 

It has long been recognised that a competitive market system is



likely to under-invest in technologies that are based on the 

production of new knowledge. Technological uncertainty means 

that firms operating in a competitive market might not be either 

willing or able to devote a level of resources to long-term 

research that is commensurate with ensuring a long-term ability 

to maintain a share of evolving markets. This sub-optimal 

allocation of resources tends to be most pronounced in 

technologies at the basic end of the R&D spectrum. A prevailing 

view in free market economies is that public support for basic 

research tends to be less controversial than "near market" 

research because it is distanced from the commercial world. In 

the early-1960s, Arrow (Arrow: 1962) and others argued that 

government funding of this type can offset this type of market 

failure without compromising competition. Basic research, by 

definition, is not directed towards commercial objectives and 

represents what could be described as "soft intervention". It 

increases technological variety without exerting any direct 

effect on the economic selection environment.



By contrast, what might be called "hard intervention", started 

to achieve prominence in the West during the 1980s. This 

emerging area of interventionist activity is concerned with 

policies that aim to create new technological infrastructures, 

directed towards specific economic growth paths which go beyond 

the goals of individual firms (Justman and Teubal: 1991). If 

these policies are to be effective, they require a deep 

understanding of the processes which shape trajectories of 

innovation and the parallel evolution of the needs of users at 

different point in the "supply chain" between basic components 

and finished products (PREST: 1992). The identification of 

strategic technologies in advance of market indicators almost 

inevitably reduces technological variety, since there are strong 

pressures to exclude options that appear to be less favourable. 

This reduction in variety impinges on the breadth of 

technological competition, although the speed with which firms 

pursue targeted trajectories of innovation is likely to be 

enhanced. In this respect, support for specific technologies



focuses the competitive process on particular fields of enquiry 

at the opportunity cost of restricting the attention paid to 

technological options that embody a greater potential if they 

were explored further.



Japan's interventions to target its infant computer and 

semiconductor component sectors for support during the 1960s and 

1970s avoided the difficulties that follow from a reduction in 

variety because they were mainly oriented towards transferring 

technology developed in the US, to the Japanese environment. 

The direction of technological development had already been 

defined by US firms. As the burgeoning literature on evolutionary 

economics has made clear, established technological trajectories 

that are supported by a similarly established base of users 

generally have a powerful "exclusion effect" on alternative forms 

of technology -- even if those alternatives embody a greater 

theoretical potential for exploitation (Georghiou, Ray et al: 

1986). Japan's strategy of pursuing technology options that 

could be exploited in the economic niche excavated by IBM and 

other leading US firms, involved little commercial uncertainty. 

The challenge played to one of Japan's acknowledged strengths. 

It involved the pursuit of relatively predictable trajectories 

of technological development which centred on known technological 

parameters. This is seen particularly clearly in the case of 

semiconductor components. Once successive generations of 

technology had been "captured", attention could be given to 

process engineering. As the level of technology progressed, 

Japanese firms demonstrated a increasing ability to develop 

production process technologies which allowed them to meet price 

and quality targets. The realisation of economies of scale at 

each successive level of technology was assisted by a large 

domestic market and an orientation towards volume production 

which was unaffected by the "quality pull" effects of defence 

procurement policies of the type that had been a factor in 

shaping semiconductor certain aspects of developments in the US 

and other Western countries. From Japan's point of view, the 

problem has come in cases where trajectories of development are



less clear as is frequently the case in the software engineering 

and AI (this is discussed further in Section 4.2).



By contrast to the history of government intervention in Japan 

prior to the late 1970s, the challenges to Western governments, 

that started to address the problem of developing appropriate IT 

support policies during the 1980s, was generally focused on the 

problems of assisting established industries that were being 

exposed to increasingly stronger competitive challenges. The 

justification for such initiatives was typically expressed in 

terms of the critical role that can be played by a strong 

indigenous capability in IT. Such an argument is frequently 

associated with bids to secure access to state-of-the-art IT for 

use in defence applications, and to justify support to civilian 

IT users that would by disadvantaged by a dependence on buying 

in technology from foreign suppliers. It is based on the premise 

that the ability to use IT effectively is linked to a national 

capability to supply IT.



Promoting a technology in the absence of clear market signals can 

by analogous to "pushing a piece of string" in the sense that 

research inputs might not be translated into commercial outputs: 

mechanisms are required to secure investment and give 

entrepreneurial direction. Clearly, advanced IT research will 

only acquire economic significance if it is exploited 

commercially and successful exploitation typically hinges on 

effective diffusion. This issue of diffusion raises two further 

categories of "market failure" that are of critical importance, 

namely: "capability failure" and "information failure". 

Capability failure encompasses such things as :deficiencies in 

the technical skills required to assimilate new technology that 

embodies a potential advantage over existing ways of doing 

things; inflexible organisational structures and an inability to 

form strategies which are appropriate to capitalising on new 

technological opportunities. Information failures relate to 

either a shortfall in the level of knowledge to make a "rational" 

decision or ,alternatively, difficulties in processing available



information effectively. While capability and information 

failures affect the IT industry in general, small firms are 

especially vulnerable to these problems. Overcoming these 

barriers to diffusion represents a formidable challenge to IT 

policy formation.



In comparison to the West, the structure and organisation of the 

Japanese IT industry tends to mean that the problems of 

capability and information failures are less significant. The 

Japanese industry centres on large firms with vertically 

integrated and diversified business activities. As Nonaka has 

noted, different phases in product development processes in 

Japanese firms are overlapping and when necessary staff can "move 

along with the project "as research is translated into 

competitive advantage. Knowledge is passed between team members 

"rugby style" as projects progress towards commercial fruition. 

The "Integrated" organisational structures frequently found in 

Japanese firms appear to offer greater "capability" than Western 

models which feature greater separation between research, 

production and marketing functions (Nonaka: 1991).





3.2 Towards a classification of IT policies



Once the case for supporting IT has been accepted, there is a 

wide range and diversity of policy measures that can be adopted 

as mechanisms for assisting IT. These might focus directly on 

IT or, alternatively, include IT as a component part of more 

broadly-based policy packages, designed to promote economic 

growth and national competitiveness. In both cases, there is also 

scope for implementing policies at different points on the 

spectrum between hard and soft intervention. One further source 

of variation, that occurs in the case of policies directed 

specifically at the IT sector, is associated with the extent to 

which the policies address demand or supply-side factors. A 

convenient way of classifying different types of support 

mechanisms has been proposed by Guy and Arnold (Guy: 1991).



Policies that affect the production and use of IT can be 

characterised, in very general terms, as being either "sector-

specific" or "environmental" in nature. The former category 

comprises policies that are targeted specifically on the IT 

sector and might include such things as government-sponsored R&D 

programmes, procurement contracts and schemes to promote 

awareness. At a more general level, environmental support 

schemes represent broad-brush measures that can be exploited by 

the IT sector as well as by other sectors of the economy (for 

example, tax incentives and regional development aid).





Both types of policy provide scope for emphasis which is either 

"developmental" or "regulatory", with the difference being 

determined, to a large extent, by the degree of government 

direction implicit in the action. Developmental policies have 

a directed, "hands-on" character. Firms are encouraged, via 

policies such as procurement, to follow paths which are intended 

to fulfil priorities identified by the state, with the result 

that technological variety is reduced. By contrast, regulatory 

policies such as the liberalization of telecommunications, are 

more "hands off" and do not involve specifying how economic or 

technical development is to take place. R&D tax incentives 

constitute a "hands-off" policy-measure because government is not 

involved in decisions concerning the type of R&D undertaken by 

firms, whereas government is involved in such decisions in the 

allocation of loans or subsidies for specific R&D projects. 

These latter examples illustrate "hands-on" or developmental 

policy measures.



Within the spectrum of sector-specific policies, it is possible 

to distinguish three broad policy types: "demand-side" actions, 

"supply-side" actions and "bridging" actions, each of which can 

be further sub-divided into a range of policy options. Supply-

side initiatives are oriented primarily towards bolstering 

indigenous IT supply capability, whereas demand-side policies are 

intended both to stimulate demand from indigenous producers and 

to improve the efficiency of user sectors via the diffusion of



IT products and services, be they supplied from indigenous 

sources or from elsewhere. Bridging actions aim to link supply-

side and demand-side actions rather than relying on independently 

conceived "technology-push" and "demand-pull" measures.



In the broader sphere of "environmental" measures, it is more 

difficult to describe actions simply as supply-side or bridging 

policies. This is partly a consequence of one person's supply 

being another person's demand. For example, non-sector-specific 

capital equipment investments can be considered as supply-side 

measures by IT producers when used to purchase new production 

equipment for their own use, but these incentives can be regarded 

equally as demand-side measures when users take advantage of them 

to purchase IT products. Similarly, control of interest rates 

can be used simultaneously to regulate investment in new plant 

and to influence the overall level of demand in the economy.



The IT Policy Triangle defines three main policy clusters.





Cluster A Hands off, Non-sector specific, Supply-side policies



General R&D tax incentives. Available to all firms from a 

variety of sectors, with government having little or no say 

in the type of R&D conducted.





Cluster B Supply side, Hands-on, Sector Specific Policies



These policies range from the direct subsidy of national 

champions to IT R&D programmes in areas that are deemed to 

be strategically important. Government usually has a 

strong say in the choice of priority areas and in 

sanctioning participation. 





Cluster C Sector-specific, Demand-side, Hands-off policies



A good example of this is an IT equipment diffusion 

incentive. By stimulating general demand for IT goods it 

is possible to benefit the IT sector.



Although there are a wide diversity of policy instruments which 

can be used to stimulate IT R&D, the main instruments in use tend 

to fall into one or more of the three main clusters. Cluster B 

policies (supply-side, hands-on and sector specific) are almost 

universally applied and constitute the main policy instruments 

in many countries. Government-sponsored collaborative research 

fits into this category, although its practical implementation 

varies significantly between countries. As the next section will 

make clear, a main feature of the Japanese case is not so much 

its use of collaboration but its deployment of a mix of policies, 

which collectively cover more than one policy cluster. It could 

be argued that problems with the exploitation of technology push 

collaborative projects has led to a growing realisation in the 

West that Cluster C policies (sector-specific, demand-side and 

hands-on) can be important instruments for stimulating the 

diffusion of IT. By contrast, it could be argued that close 

linkages between technological development and the evolution of 

different points on the "user chain" have been an enduring 

feature in the development of Japan's IT industry.





4 Japanese IT Policy and the use of Engineering Research 

Associations



The pragmatic dimension to Japanese government policy for 

industrial development has sometimes prompted the diagnosis that 

there is more concern with "winning the game" than creating a 

"level playing field". Supporters of the "Japan Incorporated" 

view often promote an image of MITI as the central headquarters 

of an industrial policy formation mechanism that is distrustful 

of government laissez-faire capitalism and a keen advocate of 

"hands-on" intervention. However, the extent to which 

subscribers to the metaphor actually consider Japan to be 

"incorporated" varies considerably.



Some of the more extreme interpretations of the "incorporated" 

view can give the misleading impression that Japan is an 

integrated, monolithic entity when, it reality, there is 

generally strong competition between firms and frequent 

disparities in the relative positions adopted by industry and 

different government ministries. Indeed, many of the supporters 

of the "Japan Incorporated" view would probably agree that 

competition has played an important role in Japanese industrial 

policy and acknowledge that in certain circumstances, 

"Schumpeterian" competitive rivalry, involving the use of 

technological innovation to capture an increased share of the 

market, is at the very heart of Japan's economic development. 

Their point is that Japanese government policy has operated in 

terms of a "desirable" level of competition, which is related to 

the relative ability of Japanese firms to compete in a national 

or global arena. From an external viewpoint, this can give an 

impression of an "incorporated" system. New industries might, 

for example, be afforded protection and subsidies until they are 

sufficiently well established to stand on their own feet. In 

practice, this involves the use of intervention to over-ride free 

market mechanisms and thereby "short-circuit" the market 

selection process.



There is, of course, no guarantee that this short-circuiting 

procedure will be effective. Competitive rivalry frequently 

leads to the production of alternative technical solutions to a 

particular problem. A reduction in rivalry can mean that 

technological options which do not appear to embody much 

potential for commercial exploitation are discounted more quickly 

than would otherwise have been the case. This can cause major 

problems if the market selection mechanism fails to evolve in the 

way that was predicted and, as Okimoto has pointed out, MITI has 

made some costly errors (Okimoto: 1989 p4-7) . However, the case 

of the computer industry is acclaimed widely as an example of 

MITI "getting it right". 



With regard to assessing the extent of the government's role in 

Japan's rapid industrial development, Hart has noted that the 

small group of Western scholars who write seriously about Japan 

is divided between those who see the role of the state as being 

central to the country's economic development and those who 

believe it to be more peripheral (Hart 1992: p37). While the 

importance of the state-industry partnership is generally 

acknowledged, there is often considerable debate about who 

actually has the upper hand. In some respects, the debate is 

redolent of features associated with past concerns about whether 

innovation was a product of "technology push" or "need pull". 

A widely-accepted way of dealing with this difficulty occurred 

with growing interest in a "combination view" whereby it was 

possible to subscribe to both models simultaneously, but with a 

degree of emphasis that varied according to contingent factors 

such as the type of technology in question, its relative maturity 

and so on. By the same token, it could be argued that there is 

a "middle ground" which can be used for interpreting Japanese 

support initiatives. The nature of this middle ground varies 

over time and according to the technology in question: it is 

contingent upon prevailing circumstances.



Japan pioneered the interventionist policies to support the 

computer industry in the late-1950s. At that time, the country



recognised the importance of building a computer industry and was 

quick to see the strategic advantage that would arise from 

technological spillover into related areas such as 

telecommunications. Although the technology was new and in many 

ways of unproven value, companies already engaged in 

manufacturing telecommunications equipment and electronic devices 

were keen to take advantage of government support to try and 

exploit its potential as a high-value-added extension to their 

business activities. Legislation was passed in 1957 to exempt 

computer technology from anti-monopoly law put in place under the 

US Occupation. MITI helped to prioritise the acquisition of 

relevant technology by Japanese firms and thereby established a 

pattern for government support to promote the development of the 

computer and related industries. A critical ingredient in this 

pattern of support has been its consistency. For most of the 

post-war period, Japan has been ruled by the Liberal Democratic 

Party creating a record of political stability that is 

unprecedented amongst the world's large democracies. Against 

this background of stable government, the Ministry of 

International Trade and Industry (MITI) has achieved 

extraordinary scope for autonomous action (Okimoto: 1989 p181).



Japan's political stability has been matched by stable population 

of leading IT firms. By contrast with the IT industries in many 

Western countries, the Japanese IT industry has not experienced 

the effects of significant restructuring caused by mergers and 

takeovers. Such activities are comparatively rare in Japan. 

Interlocking patterns of company ownership associated with 

Japan's keiretsu -- business groupings centred on either a bank 

or a supplier chain dominated by a leading manufacturer --

provide a solid financial base for supporting longer term 

investments. On occasions, keiretsu are able to create barriers 

to foreign participation in Japanese business, for example, by 

preventing the acquisition of a Japanese company that could 

provide a gateway to the domestic market. At the same time, 

internal competition between companies in a particular sector 

belonging to different keiretsu tends to intensified.



Japan's tradition of lifetime employment mitigates against 

technology transfer being brought about by the movement of 

individuals between organisations and the formation of "spin-off" 

companies to develop specific technologies. Japan's leading IT 

companies have "broad gauge" technological knowledge bases which 

tend to include more across-the-board coverage of relevant 

technologies than is the case in many of their more specialised 

Western counterparts. Boundaries between companies tend to be 

much stronger than in the West and, while there is close 

interaction at the interface with suppliers and customers, the 

"true core" is well guarded from competitors (thereby providing 

scope for the "overshooting" phenomenon mentioned earlier).



The development of Japanese IT policy has followed from a 

long-term dialogue between MITI and the major firms operating in the 

IT sector. In the course of this dialogue, MITI and the firms 

have reached a high level of "understanding" about the basis on 

which policy support measures are given. Failure of firms to 

abide by the accepted "rules" could result in them being 

disadvantaged in future MITI initiatives. While the relative 

importance of MITI and the private firms is matter for debate, 

the growth of Japan's IT industry has been impressive. 

Anchordoguy's scholarly analysis of how Japan built a computer 

industry has identified four dimensions to Japanese government 

policies: protectionist regulation; heavy subsidies; the 

establishment a national company to rent domestically-produced 

computers to Japanese users at very favourable rates and; 

cooperative R&D projects (Anchordoguy: 1989 p15). 



MITI's first step towards building a computer industry was to 

impose controls on computer-related foreign investment in Japan 

and restrict imports. IBM Japan, which was established as a 100 

per cent subsidiary of its American parent in 1950, was put under 

pressure by MITI to share technology with the Japanese industry. 

In 1960, IBM was persuaded to give Japanese computer firms access 

to its patents, although the strategy was not entirely successful 

as no arrangement was made for the practical transfer of IBM



technology to the Japanese industry (Fransman, 1990: p27). 



An effective "bridging policy" to support Japanese computer 

manufacturers and users was established by the formation of the 

Japan Electronic Computer Company (JECC) in 1961 . At a time when 

users tended to rent rather than buy computers. JECC bought 

systems from Japanese suppliers and rented them to domestic 

customers at subsidised rates. Vendors received a prompt return 

on their investment, while interest-free loans from the JECC 

provided capital to support improvement innovations. Over a 20 

year period, $2 billion in low interest government loans was 

channelled into JECC (Anchordoguy: 1990). Although this sum 

might appear relatively small by today's standards the key point 

to note is that, at the time, it represented a significant amount 

of money in relation to what the firms themselves were able to 

spend . Moreover, the scheme preserved some aspects of market 

selection mechanisms in that it helped firms who were prepared 

to help themselves. Although computers produced by IBM during 

this period were frequently cheaper, more reliable and more 

powerful, Japanese public policy was oriented towards a long-

term view of efficiency based on the ultimate acquisition of an 

indigenous state-of-the-art capability. 



The fourth dimension to Japan's policy for IT has centred on 

government-sponsored collaborative R&D projects. During the 

period since the 1960s, collaborative research schemes 

orchestrated by government agencies have become firmly 

established as a major vehicle for public policy in the 

development of the Japanese industry. These schemes typically 

involve horizontal groupings of rival firms, together with the 

participation of government agencies or research laboratories. 

The aim is to use collaboration as a vehicle for sharing the 

costs and uncertainties of technological development in order to 

enhance long-term competitiveness.



Although Western commentators often speak as if collaborative 

research was intrinsically Japanese, MITI's system of research



associations actually has foreign origins. In the late-1950s, 

the Director of MITI's Mechanical Engineering Research 

Laboratory, Dr Masao Sugimoto, was impressed by the British 

system of Research Associations (Levy and Samuels: 1989 p31). 

These associations had been introduced in 1921 to counter concern 

that Britain was loosing its technological leadership. Similar 

associations was subsequently created in a number of European 

countries and, although there was considerable diversity in the 

ways in which the British prototype was implemented in different 

national environments, a major survey published in 1965 reflected 

a generally low level of respect for the achievements of these 

bodies (Woodward: 1965). Nevertheless, Sugimoto was particularly 

interested in the assistance given by the British government to 

small and medium-sized firms and sought to transfer this feature 

of support to Japan. The first research associations in Japan 

were ad hoc ventures to support the manufacture of parts for 

motor vehicles. In 1961, the ERA system was put on a formal basis 

by the Research Association for the Promotion of Mining and 

Industrial Technology Act. This law gave research associations 

the status of a legal corporation which had the effect of making 

it possible for them to receive significant tax benefits. In the 

period up until 1965, 12 ERAs were established in a range of 

sectors.



Between 1965 and 1970, there was a moratorium on ERAs. This 

period coincided with the launch of MITI's National R&D Programme 

(often referred to as the large-scale project) which represented 

the Japanese government's first attempt to finance 100 per cent 

of the costs of certain R&D projects carried out by private 

firms. At the time it was believed that this form of support 

precluded the possibility of organising cooperative research 

schemes. However, it was subsequently decided that the two forms 

of support could complement each other and it became standard 

practice for MITI to implement national projects through the use 

of cooperative research projects (Kodama 1991 p86). A rapid 

expansion in the number of ERAs followed and an increased number 

of larger firms became involved as MITI started to use ERAs for



broadly-based national projects in areas such as 

microelectronics, materials science and biotechnology.



Prior to 1980, computer-related ERAs had been concerned with 

transferring existing technology to Japanese firms. Japanese 

hardware and component manufacturers caught up with US best-

practice technology, which have been seen as a sign that ERAs had 

served their purpose. In the event, a new type of ERA was 

pioneered which was designed to push back the frontiers of basic 

research. While the Fifth Generation Computer Project perhaps 

provides the most publicised example of this phenomenon, similar 

initiatives were launched in optoelectronics and component 

technologies for a supercomputer. Unlike previous IT ERAs, which 

were "distributed" in that participants worked from their own 

premises, these three new-style projects have been run from 

central research institutes. With less proprietary knowledge at 

stake, collaborators evidently feel more happy with closer 

cooperation (Levy and Samuels: 1991 p140).





4.1 A Review of MITI's Computer-related ERAs



The computer industry has seen more Engineering Research 

Associations (ERAs) than any other industrial sector, reflecting 

the potentially high returns from increased market shares that 

can be gained as a consequence of coordinated technological 

innovation. These ERAs may be traced from the Computer Basic 

Technology Research Association which was launched in 1962 to 

help Japanese manufacturers build a machine that would be able 

to compete with IBM's second generation 1401 series. It lasted 

until 1966 and funded on 50:50 basis by MITI and the 

participating firms. The association was also known as the 

FONTAC project, which drew the first three letters of its name 

from the initials of the participating companies: Fujitsu. Oki 

Electric and NEC. Fujitsu worked on the main processor and punch 

card equipment, while Oki and NEC addressed issues related to 

sub-processors and input/output equipment. Research was done in



house and there was minimal communication between the 

participating companies (Fransman: 1990 p29). In many respects, 

the FONTAC project could be regarded as being unsuccessful. 

Integration of the modules produced by the participating 

companies was problematic. Moreover, in the meantime, IBM had 

introduced its 360 series machines which effectively rendered 

the project's outputs obsolete.



MITI's response to IBM's 360 series machines took the form of the 

Very High Speed Computer System (VHSCS) which was launched in 

1966. It lasted for six years under the overall technical 

leadership of MITI's Electro-Technical Laboratory (ETL). By 

contrast with the FONTAC project, the VHSCS involved all six 

major computer manufacturers: thereby defining the basic format 

for participation that existed through most of the subsequent 

ERAs. There was a clear division of labour in the project with 

Hitachi, Fujitsu and NEC working on mainframes and integrated 

circuits, while Oki Electric, Mitsubishi Electric and Toshiba 

worked on peripheral equipment . Hitachi took the lead and ETL 

provided some of the more basic research input. Although the 

project achieved most of its objectives there was still a wide 

gap with IBM. Many of the more significant benefits were 

indirect . According to Fransman (Fransman: 1990 p32), much of 

the background to NEC's current domination of the world market 

for memory devices can be traced to its specialisation in this 

technology during the VHSCS project. There was very little 

sharing of knowledge created in the project and none of the 39 

patents produced involved two or more companies (Fransman: 1990 

p34).



The gap with the US was made all the more apparent in 1971 when 

the launch of IBM's 370 series once again outflanked Japan's 

efforts (and also forced RCA and GE out of the US industry). 

MITI felt that there were too many computer manufactures to 

compete effectively and sought to strengthen the Japan's position 

by promoting the formation of national champions. Substantial 

financial assistance was promised if the six major firms agreed



to some for of rationalisation into two or three groups. 

However, a combination of strong resistance by the firms and a 

significant cut in MITI's budget by the Ministry of Finance meant 

that the proposed mergers never took place. MITI' s policy 

instead centred on the formation of a further ERA.



The New Series Project ran from 1972 to 1976 and was something 

of a turning point in that it allowed Japanese manufacturers to 

divide up the market so that they could collectively provide a 

full range of products that were capable of presenting an 

"across-the-board" challenge to IBM. It was the first Japanese 

project to aim at IBM compatibility. Government subsidies of 

70.3 billion yen were made available to participating firms who 

were then required to contribute a similar level of funding from 

their own resources.



MITI forced the six firms in the project to form three teams 

although, in practice, the firms remained independent and there 

was conflict both between and within the partnerships. Fujitsu 

and Hitachi worked on the largest machines which represented a 

direct challenge to IBM's domination of the world market for 

mainframe computers. NEC and Toshiba worked on middle-sized 

machines, while Mitsubishi and Oki focused on smaller computers. 

Extensive use was made of private links that the Japanese 

companies had been able to establish build with US firms. By 

exploiting these links, the Japanese companies were more easily 

able to assimilate the technologies that were necessary to ensure 

compatibility with IBM machines. The project's strategy appeared 

to be highly effective and was instrumental in enabling Japanese 

hardware producers to match the performance of IBM 370 Series 

machines at prices which, by 1970, were between 15 and 20 per 

cent lower. For the first time, Japanese and US producers 

competed head-on.



Japan's progress in hardware during the late-1970s, focused 

attention on the US lead in very large scale integrated circuits 

(VLSI) and gave rise to the famous VLSI project which ran from



1976 to 1980. It cost 72 billion yen, 30 billion of which was 

provided by the government. Much of the justification for the 

project had its origins in the realisation that the US had a 

head-start in the development of VLSI circuits associated with 

"fourth generation" computing. In 1975, IBM was rumoured to be 

contemplating a "future system" line of computers using VLSI and 

MITI's view appeared to reflect the idea that their were too many 

computer makers in Japan to cope with a giant like IBM. The 

project might also be seen as a competitive response to a similar 

VLSI project launched in 1975 by NTT which gave substantial 

financial support to NEC, Hitachi and Fujitsu. MITI evidently 

felt that NTT had intruded into its territory (Anchordoguy: 1989 

p140) and sought to establish a larger project which included 

five of the six domestic manufacturers: NEC, Toshiba, Hitachi, 

Fujitsu and Mitsubishi. Oki was excluded as a consequence of its 

failure to exploit the results of the New Series Computer Project 

and also because the company was experiencing severe financial 

problems which cast doubt over the probability that it would be 

able to make a useful contribution to the venture.



Research was carried out at three sites : a cooperative 

laboratory; the Computer Development Laboratories, (CDL) 

established jointly by Hitachi, Fujitsu and Mitsubishi; and NEC 

Toshiba Information Systems (NTIS) Laboratories owned jointly by 

NEC and Toshiba. The cooperative laboratory was distinguished 

by the fact that it existed on a single site and was staffed by 

research engineers from all five participating companies, as well 

as members of the Electrotechnical Laboratory which is part of 

MITI's Agency for Industrial Science and Technology. According 

to Okimoto et al, the work at this laboratory concentrated on 

common basic technologies which only accounted for a minor 

proportion of the project's overall research activities. The 

main part of the project was concerned with applied development 

leading to more immediate commercial exploitation and was 

undertaken by the companies (Okimoto et al: 1984 p19).



The VLSI project is generally acclaimed to have been an 

outstanding success. Considerable advances were made in process 

technology and by 1980, Japanese firms became world leaders in 

VLSI by producing products that exhibited 1.5 micron feature 

sizes. These achievements helped Fujitsu, NEC to introduce 

highly competitive fourth generation computers . The project also 

produced over 1,000 patents, although less than 20 per cent of 

these were jointly held by more than one company.



In commenting on the role of MITI in the success of the project, 

a Director of one of the joint research laboratories has been 

quoted as saying: ". . . The role of MITI was important. Also, 

much more money went to NTIS than to the joint research 

laboratories. That money was helpful for the individual five 

companies, I think, to develop practical technologies." 

(Fransman: 1990 p97). The implication appears to be that the 

project emphasised the practical development of technology rather 

than joint research. In this respect, the project falls into the 

"catch-up category" as Okimoto has observed:



Even the heralded VLSI project (1976-1980), hailed as an 

unprecedented model of collaborative research failed to 

push semiconductor technology beyond the frontiers of 

knowledge (except perhaps in liquid crystal displays). 

While the VLSI project did advance the state of Japanese 

semiconductor knowledge, especially in the area of 

production technology (eg silicon crystal growth and 

processing). Japanese companies probably would have made 

such advances anyway. If so, the project's main 

accomplishment may have been to hasten the timetable of 

development, a nontrivial but hardly revolutionary 

accomplishment.

(Okimoto: 1986 p541)



With hindsight, the VLSI project may be seen as representing the 

conclusion of the catch-up era. The "additionality" of the 

project achieved by government funding (ie outputs that would not



have been produced in the absence of government funding) was 

evident in an increase in speed with which firms were able to 

pursue a trajectory that had been more-or-less defined by the US 

industry leaders. Given the problems that existed with several 

of the earlier computer ERAs and the initial reluctance of firms 

to participate in the VLSI venture, it is instructive to consider 

why the eventual outcome was widely acclaimed to be a success. 

One of the key factors in this success seems to have been clear 

technological objectives that were directly relevant to the 

business strategies of all the participating firms. Exploitation 

routes were apparent at the during the life of the project and 

gave direction to the research. While the firms' competitive 

instincts were to avoid collaboration, the level of government 

funding was sufficient to persuade them that failure to take part 

would place them at a serious disadvantage relative to 

participating firms. (The project enabled participating firms 

to double or even treble their potential research expenditures 

on relevant aspects of semiconductor technology.) It was a 

market-oriented venture which introduced the possibility of 

delivering tangible commercial benefits within a relatively short 

time.



Participation in the VLSI project gave Japanese firms what the US 

Semiconductor Industrial Association regarded as government 

subsidies which would be in breach of the American Anti-Trust 

Laws. It was a source of trade friction; along with unfair 

tariffs (Japan 12% verses USA 6%), preferential treatment of 

Japanese suppliers by NTT, allegations of "dumping" and a high 

trade surplus (Imai: 1983 p3). By contrast with this "in the 

market" orientation of the VLSI project, the next chapter in 

MITI's history of computer related ERAs took a very different 

turn toward basic research. 



MITI's high-profile Fifth Generation Project involved ambitious 

plans to develop a new dimension of computing. While earlier 

generations of computers were associated with component 

technologies in the form of thermionic valves, discrete



transistors and VLSI, the fifth generation centres on computer 

performance characteristics. Fifth generation computers were 

intended to understand spoken instructions, emulate human 

reasoning and explain how conclusions are reached --in short, 

a machine that embodied artificial intelligence (AI) . Advances 

in VLSI had introduced the possibility that traditional Von 

Neumann computer architectures, in which processing functions 

were performed sequentially, could be replaced with parallel 

processing. The practical realisation of parallel processing 

could create scope for the enormous increases in computing power 

that would be required for AI. Planners of Japan's fifth 

generation project envisaged that "thinking" computers would 

provide solutions to long-standing problems such as poor 

performance in software development and low white collar 

productivity (Unger: 1987 p9). They would be "machines for the 

1990s ". From the West's point of view, this was a formidable 

challenge from a country that had hitherto been an imitator. 

Moreover, Japan was threatening to take up the baton of 

leadership in a technological area that had been much discredited 

by extravagant claims on the part of AI enthusiasts. 



Although foreign organisations were invited to participate in the 

FGCS project, the eventual outcome was an all-Japanese 

initiative. Work began in April 1982 at ICOT, which is a 

specially created central research facility, located in Tokyo. 

This use of a central research facility was a notable change in 

direction from the policy adopted by MITI's preceding closer-to-

market cooperative projects where, with the exception of the VLSI 

project, firms conducted their work on their own premises (the 

so-called "distributed" model of collaboration). The basic 

nature of ICOT's research was probably a major factor in 

convincing firms that commercial confidentiality would not be a 

problem with this more open form of collaboration. However, 

Cusumano has noted that, of the participating firms, only the 

company that agreed to produce the hardware showed any enthusiasm 

for the project. He suggests that part of the reason for this 

might have been apprehension about the role of the central



research facility, although a more important factor was perhaps 

the risky and difficult nature of a project that seemed to have 

no immediate commercial applications (Cusumano: 1991 p411).



ICOT was staffed by researchers seconded from the eight 

industrial participants in the project: Fujitsu, Hitachi, NEC, 

Toshiba, Mitsubishi, Oki, Matsushita and Sharp. In addition to 

supporting the central research facility, the FGCS project also 

spent a substantial amount of its budget on research contracts 

placed with the participating firms. Funding for the project 

came entirely from the government and, at the end of the 

project's original 10-year life-span, some 54 billion yen had 

been spent. Although MITI had planned that the participating 

firms would donate matching funding to the overall 

project budget, in the event, the industrial contribution was 

limited to sending researchers to the central research facility 

at the government's expense. In 1985, there were 50 

researchers and, by the end of 1990, there were about 100 

researchers.



Phase one, from 1982 to 1984, aimed to develop basic 

technologies. The next phase was to generate the building blocks 

of computers capable of "reasoning" and the final phase between 

1989 and 1992 was designated at the period within which the 

prototype fifth generation machine would be constructed. This 

was envisaged as "user friendly" machine with 1000 parallel 

processors. While the project was able to meet its target for 

phase one, subsequent goals proved to be more difficult and there 

was also a growing divergence between ICOT research and the more 

immediate commercial interests of participating firms. At the 

project approached its conclusion, a critical article in Nature 

(26th March 1992) proposed that the Fifth Generation project 

illustrated the problems of Japan's rigid bureaucracy, noting 

that:



By the mid-1980s, it was clear that other approaches to 

parallel computing not based on traditional artificial 

intelligence techniques, such as neural networks or the



massively parallel machines created by Thinking Machines 

Inc. of the United States looked more promising. But 

having told the Ministry of Finance that it would build a 

1090-processor machine, MITI had no choice but to continue 

towards that goal.



When ICOT held an international conference to report on its 10 

year work programme, the three largest "parallel inference 

machines" exhibited only had 256 parallel processors each (New 

Scientist: 13th June 1992). The project was extended for a year 

to pursue further the goal of the 1000 processor machine. It was 

subsequently announced that there would also be an additional one 

year extension, during which time researchers would rewrite 

programmes to run on conventional machines. 



The Director of ICOT, Kazuhiro Fuchi, defended the project in his 

keynote conference speech claiming that parallel will be a core 

for future technologies that will be able to beyond the framework 

of what is possible in conventional computing. Critics argued 

that ICOT's venture into basic (as opposed to applied) 

collaborative research suffered because it was difficult both to 

set goals and adjust objectives in the light of changed 

circumstances. Other technologies subsequently proved to be of 

much greater-than-expected importance but they were not in the 

original ICOT plan and ICOT could not adapt. (For example, the 

entrepreneurial flexibility that had for example enabled Sun 

Microsystems in the US to exploit Reduced Instruction Set Chips 

(RISC) was not a feature of the ICOT model.) Similarly, 

promising approaches to AI such as neural networks. which attempt 

to mimic the functioning of the human brain, fell outside ICOT's 

research agenda. Nikkei Business reported on tensions arising 

between the government and the firms, although the firms did not 

criticise the government openly, the article claimed that there 

was mounting dissatisfaction the nature of government 

initiatives. The case of NEC was cited a firm that was formally 

a keen member of ICOT but subsequently tried to distance itself 

from the work. NEC equipment on show at the 1992 exhibition did



not bear the company logo to avoid public association with the 

project. While one earlier benefit to NEC from participation in 

ICOT was an air crew management system which has been sold 

commercially, this was based on "traditional technology" 

developed in the first phase of the project (Nikkei Business: 

June 29. 1992).



An objective assessment of ICOT's achievements is difficult 

because the publicity at the beginning of the project led to 

unrealistic expectations. At a technical level, the project was 

acknowledged to have produced some notable achievements. 

However, there was some divergence between the project's research 

agenda and the parallel evolution of participating companies 

technological priorities (Washington Post: June 2, 1992). During 

the project's first 10 years, a total of 184 researchers, all 

under the age of 35 worked at ICOT's central research facility 

(New Scientist June 13, 1992). Yet, while the experience that 

these individuals gained might be one of the more significant 

benefits of ICOT, it is difficult to be clear about the time-

scale within which these benefits will be realised.



By contrast with the Fifth Generation project MITI's follow on 

10 year "Real World Computing Programme" appears to represent an 

attempt to circumvent this problem by using a diversity of 

research strategies prior to mid-term appraisal of which 

technologies to select for mainstream development (Science, 

October 23 1992). The Real World Computer will be oriented 

towards a flexible information system with an intuitive 

information processing capacity similar to that of human beings. 

Hajime lrisawa, who is a former MITI official and Executive 

Director of the Real World Computing Programme, has described the 

project as "very basic" : there is no intention of even building 

a prototype computer (Tokyo Business: March 1993 p3) . MITI is 

expected to provide something in the order of 90 per cent of the 

funding which is estimated to be in the order of 60 million yen 

over the 10 years. The project aims to create a new 

infrastructure for basic research which is expected to include



a number of non-Japanese organisations. 



In summary, while the achievements of IT ERAs are cited as 

examples of effective cooperation they have also embraced a 

potent mixture of competitive rivalry between participating 

firms. Anchordoguy points out that:



"Cooperative" R&D conjures up images of members of 

different firms working together on the same problem. 

While this did occur, it was rare. For the most part, 

tasks were assigned to different companies. In some cases, 

the firms divided up the work and gave one another access 

to the resulting patents; in other cases, the firms split 

into groups to take different approaches to the same 

problem while agreeing to share results.



(Anchordoguy: 1989 p43) 



Further support for the idea that inter-firm knowledge sharing 

is of comparatively minor importance to firms participating in 

Japan's national cooperative projects is provided by Fransman's 

detailed analysis of cooperation and competition in information 

technology in the Japanese system. Quantitative survey data 

obtained from companies participating in five major national 

programmes led him to conclude that: "Access to knowledge 

contributed by other participating organisations was not 

generally felt to be a very important benefit. . . " (Fransman: 1990 

p252) . While greater knowledge sharing has been a feature of the 

more basic research consortia launched during the 1980s, this 

type of sharing does not involve significant amounts of 

proprietary knowledge. 



MITI's collaborative ERAs show that if certain circumstances 

collaboration between competitors can be of considerable 

importance to the development of Japan's IT industry: for 

example, the New Series and VLSI projects stand out as successful 

projects. Yet, it should be stressed that "collaborating to



compete" is rarely easy. Effective collaboration requires a 

sufficient inducement to collaborate. Typically this involves a 

sufficient incentive to participate (usually if the form of 

government funding), a belief that the potential outcomes that 

will be commercially relevant and an expectation that it would 

be possible to "internalise" these outputs without leaking 

proprietary knowledge to competitors in the grouping. The 

commitment of industrial funding to the project can also be a 

major factor in encouraging firms to link the research agenda to 

commercial requirements.





4. 2 A Note on Collaborative Software Projects



In comparison to its promotion of computer hardware and 

components, the history of MITI's ventures into collaborative 

research into software have been of limited success. Prior to 

the 1980s, MITI did not channel significant funds into 

collaborative software R&D projects. While specific projects 

have experienced different problems, Cusumano notes that there 

have been some common themes: "poor planning disagreements on 

objectives and poor results" (Cusumano: 1991 p389).



Of the more recent software projects, SIGMA (Software Industrial 

Generation Maintenance Aids) is perhaps worthy of special mention in 

that it represented a high profile attempt to address Japan's 

emerging "software crisis" by automating the production of 

software and facilitating its re-use by improving the quality of 

software components. SIGMA was organised by MITI and run through 

its Information Processing Agency (IPA). Work began in 1985 and 

lasted 4.5 years. The overall budget was 25 billion yen, which 

came from government and industry sources (with the size of 

industrial contributions being determined in accordance with the 

company's turnover). A total of 194 industrial organisations 

participated including 15 hardware manufacturers, 109 software 

companies and 11 foreign firms with Japan-based operations. 

Sigma aimed to produce a UNIX based workstation which could be



used for developing applications software more efficiently. 

Following the project's completion in 1990, the SIGMA Systems 

Company was formed to assist in the commercialisation of the 

projects outputs. However, the company was not able to cover its 

costs and sales from SIGMA tools were negligible.



A major problem with the SIGMA project was that the tools 

developed did not have any significant advantages over the 

products of major software houses, while the range of tools was 

limited. Switching to SIGMA tools would involve companies in a 

change of software and hardware to adopt a system that lacked 

some of the facilities provided by current technology. Another 

difficulty was that the workstation and software makers who 

supported other operating systems were also expected to act a 

SIGMA sales agents and were thereby confronted with a direct 

conflict of interests. Some critics of SIGMA have suggested that 

problems arose because MITI was trying to push firms towards 

technologies that lagged behind best practice alternatives.





5 The UK Alvey Programme: A Case-study of a European response 

to Japan's Fifth Generation Programme



One of the first initiatives that can be directly related to the 

Fifth Generation announcement was the UK government's Alvey 

Programme to promote pie-competitive research in advanced IT. 

It is used here to provide a case-study of "European-style" 

collaboration. In Europe, government policies to support 

electronics related industries began to emerge as a subject for 

debate in the late 1970s. A traditional predilection for "soft" 

interventionist policies to support a more healthy investment 

environment began to give way to more blatantly interventionist 

sector-specific "strategic" policies. 



Prior to the fifth generation announcement, the UK had been 

actively seeking industrial links with Japan in order to offset 

declining competitiveness in the domestic IT sector. ICL, which 

was Britain's only significant indigenous computer manufacturer, 

had already forged links with Fujitsu and the official view was 

that Britain might be able to benefit from further links with the 

Japanese IT industry. However, approaches made to Japan during 

the period leading up to the October 1981 "Fifth Generation" 

conference revealed that the country considered itself to be 

self-sufficient in hardware terms and was interested in gaining 

access to British academic expertise, particularly in artificial 

intelligence. One academic, Professor Donald Michie, is Reported 

to have commented that collaborating with Japan in this area 

would be rather like: "cooperating with a vacuum cleaner" (Oakley 

and Owen: 1989 p17). This concern about a one-way flow of 

information was reflected widely in academic, government and 

industrial circles. One consequence was to focus attention on 

how best to emulate the Japanese research effort.



Alvey was set-up as the first stage in a 10 year plan to improve 

the competitiveness of the UK IT sector although, in the event,



it was only funded for five years. The programme began in 1983 

and broke new ground in that it was sponsored by three different 

sections of government: the Department of Trade and Industry 

(DTI), the Ministry of Defence (MoD) and the Science and 

Engineering Research Council (SERC). Its main thrust was to 

support pre-competitive, collaborative research projects in the 

enabling areas of IT. Some 200 collaborative projects involving 

partners from industry and academia were undertaken and the 

overall budget was 350 million pounds sterling, 200 million 

pounds of which came from government sources with industrial 

participants providing the balance. For the most part, a 

"distributed" model of collaboration was used with projects 

typically lasting in the order of three years. By contrast with 

the Japanese model of collaboration, there was a very real 

expectation that there would be substantial knowledge transfer 

between partners. Government funding was not provided until all 

the partners had signed a collaboration agreement which covered 

the ownership of intellectual property produced. Firms tended 

to be cautious in their approaches to these agreements and many 

projects suffered long delays as a consequence of the time that 

it took to draft an agreement that was acceptable to all the 

partners in the project.



Alvey was successful in promoting collaboration between UK 

industry and the national science base . It also provided 

valuable learning experience for the majority of participants, 

paved the way for further collaboration at the UK and European 

levels and offered lessons for the running of a rather more 

modest follow on programme to support IT through the use of 

collaborative R&D. Revitalising the UK IT sector was rather more 

problematic and, in many respects, the level of exploitation of 

Alvey outputs has been disappointing. In some cases, barriers 

to exploitation arose because of poor linkages between R&D and 

production facilities, while in other instances technical 

difficulties proved to be insurmountable. However, there were 

also difficulties which arose from the fact the research had been 

undertaken as part of a collaborative project. The involvement



of other partners mean that projects are relatively inflexible 

in the face of changing commercial circumstances. Major changes 

in a project's direction involved gaining the agreement of other 

partners and the government . Project progress might also be 

affected by a divergence between the interests of the academic 

participants and the commercially-related concerns of industrial 

collaborators. Withdrawal by partners, following a reappraisal 

of business and technical priorities, tended to create problems 

for other collaborators (Guy, Georghiou, Ray et al: 1991 p V).



An independent evaluation of the Alvey programme concluded that 

pre-competitive research and development programmes are well 

suited to a range of tasks but are not in themselves sufficient 

mechanisms to bolster the competitive performance of the IT 

sector. Complementary private sector and government initiatives 

are needed to: relate IT development to users; promote effort 

within firms to formulate technological strategies to facilitate 

the exploitation of research, and; a serious re-evaluation of the 

need for patient capital (Guy, Georghiou, Ray et al 1991). To 

varying extends, each of these three "deficiencies" are less of 

a problem in Japan.



While interventionist policies enabled Japan to build an IT 

industry in the shadow of IBM, they have not solved the 

shortcomings associated with the exploitation of outputs arising 

from programmes of basic collaborative research. Rather support 

policies in Japan interacted with the Japanese environment to 

create a "climate for innovation" through tax incentives for 

research and infant industry protection. This has been helped by 

structural features in the Japanese economy. Japan's success in 

consumer electronics was an example of close matching of 

innovation trajectories to prevailing patterns of user 

requirements, while commercial returns on this effective coupling 

process was helped by a domestic market which is more than twice 

the size of that in the UK. The absence of high-specification 

military procurement policies, combined with competition between 

a number of domestic producers is also instrumental in



stimulating both competitive innovation and aggressive pricing 

policies. Japanese industrial structure and, in particular, the 

broad technological and foundations of the leading players in the 

IT sector has been a source of relatively patient capital.



Even though collaborative research does not in itself appear to 

be an entirely satisfactory solution to the problem of 

revitalising the UK IT industry, it has been retained a leading 

position in the portfolio of British support mechanisms. A 

scaled-down national programme of collaborative research has 

followed from Alvey (The Information Engineering Advanced 

Technology Programme), while an expended commitment has been made 

to pan-European collaborative research.



The launch of Alvey was followed by a dramatic expansion in pan-

European collaborative R&D under the European Community's first 

Framework Programme (1984-87), which covered a broad range of 

research areas and has been followed by subsequent programmes. 

These programmes are intended to support pre-competitive research 

projects. A fourth Framework programme was agreed in 1992. The 

most prominent IT initiative in the framework scheme is probably 

the European Programme for Research in Information Technology 

(ESPRIT). This began in 1984 and covers microelectronics, 

software engineering, computer integrated manufacturing and 

advanced IT systems for business and the home. ESPRIT is a 

"technology push" initiative without any formal mechanism for 

linking projects to the evolving needs of IT users. The most 

important benefits derived by industrial participants in ESPRIT 

have been mainly associated with improvements in basic know-how 

the adoption of more ambitious research objectives (OECD: 1992 

p76). ESPRIT is complemented by the Eureka programme, which aims 

to be market-driven. It was launched in 1985 by President 

Mitterand of France as a response to President Reagan's Star Wars 

initiative. Eureka provides funding for a variety of 

technologies and involves all 12 member states of the European 

Community, as well as seven other nations in Europe and the 

European Commission. One particularly significant piece of work



funded under Eureka in the IT sphere is the Joint European 

Submicron Silicon Investigation (JESSI) project which could be 

seen as a European response to the USA's Sematech project, which 

is introduced in the next section.



At the pan-European level, the "overheads " associated with 

collaborative research tend to be amplified by the geographical 

separation between partners and language barriers. Concern has 

been expressed that the "pre-competitive" emphasis of ESPRIT is 

producing research that is a still some considerable distance 

from the market and vulnerable to the type of barriers to 

exploitation that beset the Alvey project. Successful 

collaboration between industrial firms is generally associated 

with clear research targets which embody a clear potential for 

exploitation and are shared by all the participants. In this 

respect it is instructive to consider the case of a "spontaneous" 

European collaborative response to the Fifth Generation programme 

which was initiated without government funding. The (ECRC) 

European Industry Research Centre located in Munich was the 

established in 1984 as an Anglo-French-German collaboration by 

ICL, Bull and Siemens. By contrast with ICOT, the ECRC has been 

able to shift its emphasis in the light of changing 

circumstances. It is only about half the size of ICOT, with some 

50 researchers, and has concentrated on more modest projects 

which are coupled to the commercial requirements of partners 

(Guardian: Aug 13 1992).



Collaboration has become well established in Europe as a way for 

spreading the costs and uncertainties of research. Yet these 

collaborative ventures frequently centre on research that is some 

considerable distance from the market and involve academic as 

well as industrial collaborators. Routes to exploitation are 

less clear than was the case of the more successful commercially-

oriented Japanese schemes. The European IT industry also differs 

from Japan in that it is fragmented and subject to constraints 

imposed by a large number of relatively small national markets. 

Differing national priorities, coupled with cultural and language



barriers, add to the difficulties of implementing a concerted 

programme of action. At the same time there are restructuring 

and adjustment issues associated with successive waves of mergers 

and takeovers. This means that the planned "temporary monopoly" 

that would be associated with a collaborative project is 

sometimes overtaken by an actual monopoly that occurs in the wake 

of moves toward increased industrial concentration.





6 The Rise of Collaboration In the United States



In the USA there are a few overtly "hands-on" actions in the 

civilian sphere and a plethora of "hands-off" actions, both non-

sector-specific and sector specific. The other major 

characteristic of the US support is a range of sector-specific 

actions determined by the Department of Defense; many of which 

have a determinedly hands-on nature.



During the 1970s, the United States led the world in component 

technology, computer manufacture and many aspects of what has 

since become known as "software engineering". At the federal 

level, the official US industrial policy has been avowedly non-

interventionist: the policy was not to have a policy. By 

contrast, many individual state governments have a long tradition 

of interventionist policies. In practice, federal defence 

spending has performed a key role in establishing America's post 

Second World War lead in electronics and computer-related 

technologies. National security was used as a justification for 

"hands on", sector-specific interventionist industrial policies.



Although a relative decline in US's industrial competitiveness 

over the 1980s led to calls for a more coordinated approach to 

industry, these were met with fierce opposition in official 

circles. In 1981, the Reagan Administration entered the White 

House with a strong commitment to minimal intervention in the 

commercialisation of technology. It was deemed that the 

appropriate role for federally-funded research in the civilian 

sector should be restricted to supporting basic research. At 

that time, the exploitation of research was a matter for market 

forces. Nevertheless, during the course of the 1980s, 

collaboration in support of the pursuit of competitive advantage 

gained ground and established a trend which looks set to continue 

into the 1990s. This is worthy of special note because it runs 

contrary to the tradition of Anti-Trust which has been a very 

deeply engrained feature of American economic policy. Moreover, 

in sharp contrast to MITI's collaborative ERAs and government



sponsored European responses to the Fifth Generation Programme, 

collaboration in the US had private sector origins. 



Japan's announcement of its Fifth Generation Computer Programme 

caused concern amongst members of the US IT industry and policy-

makers. While the perceived threat was not sufficient to break 

the "anti-centralist" tradition of US politics, it did provide 

an opportunity to mobilise Congressional support for an increase 

in funding to established agencies. For example, the Department 

of Defense's Defence Advanced Research Project Agency (DARPA) 

launched a ten year Strategic Computing programme in 1983 which 

set ambitious targets for the practical demonstration of advanced 

AI and parallel computing. The National Science Foundation (NSF) 

also expanded the scale of funding for computing and AI. In 

addition, NSF played a pivotal role in consolidating the US's 

leading position in packet-switched networking and computing. 

A further policy action with implications for the IT industry was 

the Strategic Defence Initiative (more popularly known as Star 

Wars). It was first announced in 1983, as a space-based system 

to defend the US from inter-continental ballistic missiles. By 

1993, the project had consumed some $32 billion -- mainly on 

ambitious military-oriented projects which are distanced from 

commercial markets.



Outside of government, one of the early responses to the Fifth 

Generation announcement was the Semiconductor Research 

Corporation (SRC) . This was created in 1982 as a permanent non-

profit-making institution linked to the highly influential 

Semiconductor Industry Association. SRC is supported by 

industrial corporations who pay a subscription in proportion to 

their turnover and, in return, gain access to a broad spectrum 

of semiconductor research projects, conducted at US universities 

under SRC sponsorship. (By contrast with Japan which is 

frequently characterised as having a relatively weak university 

sector, US industry has a solid history of using of universities 

to support research activities.)



The major non-governmental collaborative response to Japan's 

plans for fifth generation computing was the setting-up of 

Microelectronic and Computer Technology Corporation (MCC). This 

is an independent research institute funded by a consortium of 

US companies. It was set-up in 1982 to undertake pre-competitive 

research in semiconductor and computer technology on a permanent 

basis. For the US, MCC was nothing short of revolutionary. It 

involved unprecedented cooperation in a fiercely competitive 

industry and was only made possible by changes in the anti-trust 

legislation. A central research facility was constructed at 

Austin in Texas and work began in 1984. From the outset, MCC was 

seen as a bold initiative that would only produce results in the 

long-term. During the late-1980s, MCC embarked on a era of 

restructuring which led to an increased emphasis on interim 

deliverables while, at the same time, pursuing longer-term goals. 

In this respect, MCC's continued survival reflects a degree of 

flexibility which is way beyond what proved to be possible at 

ICOT.



MCC was at the forefront of a new approach to collaborative 

research in the US. During the 1980s, several Bills were passed 

by Congress which were designed to aid certain collaborative R&D 

activities that could be regarded as crucial to the national 

interest. In particular, these Bills served to clarify the 

position of collaborative R&D with regard to existing anti-trust 

legislation. This paved the way for a wide variety of research 

consortia and, by the late-1980s, collaboration had been 

established as a legitimate vehicle for sharing the costs and 

uncertainties of pre-competitive R&D (see: Evan and Olk, 1990).



While MCC paved the way for collaboration in the US, the launch 

of the Sematech Consortium in 1987 represented a sea change in 

the US administration's prevailing attitude to collaborative R&D. 

Sematech was founded with the largely commercial objective of 

enabling the US to compete more successfully with Japan in the 

manufacture of semiconductor devices. It is a non-profit making 

organisation funded by a total of $500 million over 5 years from



the US Department of Defence. via DARPA, and a similar sum made 

up from subscriptions paid by 14 leading semiconductor 

manufacturers.



Competitiveness in chip manufacturing processes had been driven 

by the quest for increased density in the production of memory 

chips. The problem for the US was that, by the mid-1980s, 

leadership in this technology was tending to shift to Japanese 

firms. Moreover, US firms were also being displaced from the key 

component technology of dynamic random access memory (D-RAM) 

chips by Japanese competition. Fears that US firms would fall 

behind in semiconductor production technology without an 

indigenous D-RAM capability were countered by the idea that 

SEMATECH would act as a "technology driver" or "forcing ground" 

for promoting state-of-the-art processes, although the extent to 

(LSI Logic, which was founder member of Sematech, withdrew in 

January 1992 arguing that the subscription could be used better 

internally.)



A further innovation in federal support for R&D has followed from 

the 1988 Omnibus Trade and Competitiveness Act. One of the 

consequences of this act is the creation of what some 

commentators have referred to as a 'Civilian DARPA' in the form 

of the National Institute of Standards and Technology's Advanced 

Technology programme (ATP). This small but innovative scheme was 

launched in 1990 as a mechanism for providing federal support to 

US business carrying out pre-competitive R&D in generic 

technologies (defined as concepts, components, processes or 

scientific knowledge that could be applied to a broad range of 

products and processes). The emphasis is on supporting 

technologies such as IT that will play a significant role in 

enhancing US competitiveness.



During the period since Japan's Fifth Generation announcement, 

US policies that impinge on IT have been changing. Collaboration 

has become established as a legitimate mechanism for conducting 

pre-competitive R&D both for private industry and, more recently,



for federally-funded research. An industrial policy is slowly 

beginning to emerge as part and parcel of the evolution of 

policies for national security, on the one hand, and a 

competitive trading position, on the other. Defence research 

funding is now being used to support advances in civilian 

technologies in the expectation that it would preserve a national 

capability that could be exploited in future military programmes 

which, as Mowery and Rosenberg have argued, is a reversal of 

earlier patterns of funding and technological spillover (Mowery 

and Rosenberg: 1989). However, there is a sense in which the 

national focus of these policies has been overtaken by the 

globalisation of the computer industry. Following the signing 

of the Japan-US Semiconductor Trade Agreement in September 1991, 

there has been a steady increase in the "natural alliances" 

between Japanese and American firms and many would argue that 

international rather than national ventures will become more 

relevant as the 1990s unfold.



While it is difficult to equate initial post-fifth-generation 

developments in the US to the UK's Alvey Programme and similar 

European government-sponsored responses to the "Japanese 

challenge", the rise of private-sector-initiated collaboration 

during the 1980s reflected a growing belief that "new tactics" 

were required to compete with Japan. Moreover, the principle of 

sharing research costs and uncertainties gradually permeated 

official policy making, leading to the rise of ventures such as 

Sematech which are US becoming established as legitimate industry 

support mechanisms for "keeping up" with leading Japanese firms. 

The principle of federal support for collaborative research in 

pursuit of competitiveness is also illustrated by the National 

Bureau of Standard's advanced Technology Programme. Although 

this only exists on a modest scale, it is nevertheless indicative 

of a new approach to policy formation.





6 Concluding Comments



Markets are very complex structures. Moreover, the nature and 

dimensions of complexity vary considerably between different 

national environments. While it could be argued that "national 

innovation systems" perform broadly similar basic functions in 

the sense that they translate technological knowledge into 

commercial products and production processes, there are 

considerable differences in the way in which different innovation 

systems perform these functions. In particular, substantial 

variations exist in the relationship between private firms, 

governments and bodies contributing to aspects of "public domain 

knowledge" that are relevant to technological development (eg 

universities and similar institutions).



The development of an effective IT capability by Japan's 

"national innovation system" was conditioned by a number of 

circumstances that were particular to Japan. Policies were 

initially directed towards creating economic "space" for the 

basis of a computer industry to be established in the face of an 

overbearing competitive treat posed by IBM. The use of US-style 

military procurement to assist in this objective was precluded 

by a post Second World War ban on defence-related exports. Since 

direct foreign investment ran against Japan's tradition of self-

sufficiency, the relative status of collaboration was higher on 

the agenda of policy options than might otherwise have been the 

case. Even then, collaboration was only one of four basic 

policies. Moreover, the building of a "collaborative culture" 

did not occur in an instant but rather as the consequence of a 

gradual learning process, which took place over a series of 

projects. Strong competition between the firms was accompanied 

by a sometimes less than harmonious relationship between the 

industrial sector as a whole and the government. Attempts by 

MITI in the early-1970s to restructure the industry into a few 

"national champions" were strongly resisted by the firms and in 

the end did not take place. Thus, a drift towards monopoly which 

is gennerally restrained by legislation in the West (eg through



US Anti-Trust Laws etc) was constrained by the "natural" 

competitive instincts of Japanese firms.



Given the extent of competition between Japan's IT firms, it is 

perhaps appropriate to ask why initiatives such as the VLSI 

project were apparently so successful. In prospect, the firms 

were reluctant to participate in a venture that could compromise 

their independence. Yet the degree of government funding was 

high enough to mean that non-participation would place firms at 

a commercial disadvantage relative to participating firms. This 

commercial incentive was accentuated by the project's "near 

market" orientation. Although there was a central facility which 

provided for an environment for shared knowledge creation, this 

was mainly concerned with more basic research. The bulk of the 

project's commercially-oriented development work was organised 

on a more "modular" basis enabling firms to internalise the 

benefits of government-sponsored research without compromising 

their competitive positions. Commercialisation of project 

outputs was helped by effective internal communication channels 

that exist between different sections within Japanese firms. 

Against the background of Japan's stable industrial structure, 

the project could be thought of as an example of the government 

providing a " balanced subsidy" to rival firms, rather than 

creating a mechanism for sharing the risks and uncertainties of 

research. These uncertainties were in any case already limited 

by the fact that the "window of commercial opportunity" for the 

technology had already been opened by leading US firms and the 

direction of technological development was well signposted.



Once Japan started to catch-up with the West, the position began 

to change. The clear target disappeared and it became more 

difficult to see how MITI could impose strategic direction on an 

industry that was subject to such a rapid pace of technological 

development, combined with turbulent changes in the structure of 

its associated markets. One dimension of MITI's response was a 

move towards collaborative basic research projects. The Fifth 

Generation Computer Systems Project was a major departure from



the preceding applications-oriented ERAs. In some respects, it 

could be seen as an attempt to build a basic research component 

into an innovation system which did not have a strong tradition 

of links with university research. Its achievements have not 

been closely correlated to the evolving agenda of participating 

firms' commercial requirements and the translation of outputs 

into competitive advantage is likely to be an indirect process. 

Many aspects of the project's work have entered the public domain 

and have a status which is not altogether unlike that of academic 

work conducted in Western universities. The Real World Computing 

Programme is a further extension of the use of collaboration to 

promote inter-firm, "close to the public domain" basic research. 

In this respect, collaboration is being used to foster the 

development of a new research network.



Given the very particular nature of the circumstances under which 

collaborative research promoted the competitiveness of Japanese 

IT firms, it is perhaps hardly surprising that the West's 

attempts to use the concept as a means for matching the dynamism 

of the Japanese economy have not been without their problems. 

The expectation that the Alvey programme could use pre-

competitive collaborative research to match Japanese-style 

competitiveness asked more from collaborative research than had 

been achieved in Japan. It was the application-oriented ERAs 

organised on modular basis that were most closely associated with 

promoting industrial competitiveness. The Fifth Generation 

Project's use of shared knowledge creation was a radical 

departure from previous practice. 



While problems with exploitation are likely to remain a problem 

for pre-competitive collaborative research programmes in any 

national environment, it does not necessarily follow that such 

initiatives are without value. New networks of the sort being 

pioneered in Japan's Real World Computing programme were 

established with some considerable success under Alvey. These 

networking benefits are also a feature of pan-European 

initiatives and have also been exhibited by collaborative schemes



in the US. Such networks are of special importance in national 

environments where there is scope to extend the role of public 

sector research in the national innovation system. Countries, 

like Britain and the US, which have a strong academic research 

in IT-related sectors have used cooperative research to 

consolidate their national knowledge bases. During the 1980s, 

Anti-Trust restrictions (which were eliminated from the Japanese 

environment in 1957) were relaxed in the US, thereby articulating 

a fundamentally new attitude to competition. Federally supported 

initiatives such as Sematech are also being commended to the 

Clinton Administration as policy as models for improving the 

national innovation systems. Trans-national links across Europe 

have also helped to unite disparate resources. However, the 

construction of these new communication infrastructures does not 

solve the problem of how best to proceed into the unknown 

territory of research-driven innovation. While they can provide 

an effective tool for supporting national and international 

knowledge creation, claims that they will automatically 

strengthen competitive performance should be treated with 

caution.



During the period when Japan's computer industry was catching-up 

with best practice Western technologies, the government's 

policies were often effective in minimising the effects of 

competitive disadvantages which confronted Japanese firms. Since 

the early-1980s, the Japanese government has faced the problem 

of producing policies that foster national competitive 

advantages. Governments in such a position are not usually well 

placed to "pick winners" and Japan's approach to supporting the 

IT industry now features a substantial element of basic research: 

thereby reflecting a long-established theme of policies practised 

in the West. Understanding future patterns of innovation will 

require a deep appreciation of how IT firms' organisational 

knowledge creation processes interact with evolving technological 

and market systems at national and international levels.



References



Abegglen, J. and Stalk, G., (1985), Kaisha, The Japanese 

Corporation, Basic Books Inc., New York.



Anchordoguy, M., (1989), Computers Inc.: Japan's Challenge to 

IBM, Harvard University Press.



Anchordoguy, M. in: Ferguson, C.H., (1990), Computers Keiretsu, 

Harvard Business Review, July-August, pp. 55-70.



Arnold, E and Guy, K., (1986), Parallel Convergence : National 

Strategies in Information Technology, Francis Pinter, London.



Arrow, K., (1962), 'Economic Welfare and the Allocation of 

Resources for Invention', in Rosenberg, N. (ed), (1971), The 

Economics of Technological Change, Harmondsworth, Penguin.



Buisseret, T., (1992), The Role of Users in the Information 

Engineering Advanced technology Programme (IEATP), PREST, 

University of Manchester, UK; Report to the UK Department of 

Trade and Industry. 



Commission of the European Communities, DG XIII 

Telecommunications, Information Industries and Innovation, 

(1991), The European Electronics and Information Industry: State 

of Play, Issues at Stake and Proposals for Action.



Cusumano, M., (1991), Japan's Software Factories: A Challenge to 

US Management, Oxford University Press. 



Even, M. and Olk, P.,(1990), R&D Consortia: A New US 

Organisational Form, Sloan Management Review, Volume 31, No. 3, 

pp. 37-46.



Ferné, G., (1989), R&D Programmes for Information Technology, The 

OECD Observer, August/September, pp.10-13.



Feigenbaum, E. and McCorduck, P., (1985). The Fifth Generation: 

Artificial Intelligence and Japan's Computer Challenge to the 

World, Michael Joseph, London.



Georghiou. L., Metcalfe J., Gibbons M., Ray T., Evans J., (1986), 

Post Innovation Performance: Technological Development and 

Competition, Macmillan.



Guy, K., (1991), Policy Instruments for Stimulating IT&T R&D in 

European Nation States, paper prepared for: DG XIII's Public 

Strategy Observatory, Brighten, Science Policy Research Unit, 

university of Sussex, UK



Guy, K. and Arnold, E., (1988), Performance and Competitive 

Success : Government IT Policies in Competing Countries, London, 

NEDO Books. 



Guy, K., Georghiou, L., Ray et al, (1991). Evaluation of the 

Alvey Programme for Advanced Information Technology, HMSO.



Gregory, G., (1986), Japanese Electronics Technology: Enterprise 

and Innovation ( 2nd Edition), Japan Times, Tokyo.

Justman, M. and Teubal, M., (1991), Strategic Industrial Policy: 

Infrastructure Development and Market Building, Jerusalem: 

Jerusalem Institute for Israel Studies. 



Hart, J. A. (1992), Rival Capitalists: International 

Competitiveness in the United States, Japan and Western Europe, 

Cornell University Press.



Imai. K., (1983), An Industrial Organisation Analysis of the 

Semiconductor Industry: A US Japan Comparison, Institute of 

Business Research, Hitotsubashi University, Kunitachi, Tokyo, 

Discussion Paper No 113.



Kodama, F. (1991), Analyzing Japanese High Technologies: The 

Techno-Economic Paradigm Shift, Pinter Publishers.



Mowery, D., (1992), The US National Innovation System: Origins 

and Prospects for Change, Research Policy, vol 21, no2, ppl25-

144.



Mowery, D. and Rosenberg, N., (1989), New Developments in US 

Technology Policy: Implications for Competitiveness and 

International Trade Policy, California Management Review, vol 32, 

no 1, pp. 107-124.



Nonaka I., (1991), Managing the Firm as an Information Creation 

Process, Advances in Information Processing in Organisations, Vol 

4, pp239-275.



Nonaka I., (1992), Inter-organisational Knowledge Creation: 

Collective Innovation in the Japanese Semiconductor Industry, 

paper presented at the International Conference on "New 

Imperatives for Managing Revolutionary Change" Shiuoka, Japan, 

September 3-6, 1992.



Oakley, B. and Owen, K.,(1989) Alvey: Britain's Strategic 

Computing Initiative, MIT Press.



Organisation for Economic Cooperation and Development,(1992), 

Science, Technology Industry Review, no 10.



Okimoto D., Sugano T., Weinstein F., (1984), Competitive Edge: 

the Semiconductor Industry in the USA and Japan, Stanford 

University Press, Stanford, California.



Okimoto D., (1989), Between MITI and the Market: Japanese 

Industrial Policy for High Technology, Stanford University Press.



Ray T, (1988), Shooting at a Moving Target: How to Time 

Innovations to Meet Changing Customer Expectations, Strategic 

Direction, No 38.



Ray T. et al., (1989), Competition and the Momentum of Technical



Change, in; The Competitiveness of European Industry, edited by 

Francis, A. and Tharakan P.K.M., Routeledge, pp165-199.



Ray T., (1992), Alvey and the Programmes Beyond, in: Gabriel J 

and Gessner W., (Eds), Government Promotion of Microsystem 

Technology in Europe, VDI/VDE Technologiezentrum 

Informationstechnik Gmbh, No 1, October 1992.



Ray T. and Guy K., (1993), Government Policies for Information 

Technology, chapter in: Longman Concise Guide to World Technology 

Policies.



Tokyo Business, Real World Computing: A Brave New World, March 

1993.



Unger, J.M., (1987), The Fifth Generation Fallacy: Why Japan is 

Betting its Future on Artificial Intelligence, Oxford University 

Press.



Woodward, F., (1965), Structure of Industrial Research 

Associations, Organisation for Economic Cooperation and 

Development.