ORGANIZATION FOR ECONOMIC COOPERATION AND DEVELOPMENT
-------------
O.E.C.D.
COMMITTEE FOR INFORMATION,
COMPUTER AND COMMUNICATIONS POLICY
HIGH LEVEL MEETING

THEME 1:

INFORMATION AND COMMUNICATIONS TECHNOLOGIES
FOR ECONOMIC DEVELOPMENT

Note prepared by Francois Bar
from various material by the
Berkeley Roundtable on the International Economy

May 1987

This document represents the background document for Theme I.
It is submitted to the Committee for comments and discussion.

INFORMATION AND COMMUNICATIONS TECHNOLOGIES
FOR ECONOMIC DEVELOPMENT

Information and communications technologies have clearly reached star-
status on every developed country's economic agenda. By the products and
industries they directly generate, through the structural transformation
they permit and provoke, electronics have become a powerful agent of
economic development. Clearly however, the current transformation, based
on the diffusion of electronics-based systems, will not rejuvenate all
economies automatically, nor will it affect all countries equally.

Differences in national industrial legacies, economic structures, and
governmental policies will matter greatly to the implementation of the
technologies' potential.

The purpose of this note is not to prescribe specific policies, but
rather to propose a framework and to outline issues for discussions that
can inform policy development. Rather than attempting to spell out all-
encompassing --and necessarily limited-- theories, we relie extensively
upon examples and sector studies. Their stories provide concrete evidence
of the constraints and implications of various policy choices.

The major economic development potential of information and
communications technologies lies in the diffusion of their products and
processes throughout the economic fabric. Ultimately, diffusion is a
market process. Therefore, it cannot be imposed, nor prescribed by
policies. Policies aimed at the diffusion of electronics will be one
element, not always a major one, of the environment within which firms make
decisions about production technologies, production processes, and product
design. Policy development must therefore procede from an understanding of
this environment and its internal dynamics.

If technology diffusion is to be the overarching policy goal,
governments have an important role to play in facilitating access to the
technology, in preparing the grounds for its diffusion, in stimulating its
implementation. We chose here to organize the discussion around five major
issues. Each one of these five issues is analyzed through a particular
segment of the electronics industry, or a particular feature of the
technologies.

The diversity and pervasiveness of electronics applications point at
the first issue. The range of electronics sectors is so wide that few
countries can control all of them. Choices have to be made about which
specific sectors to support, and how to promote them, based on policy
assessments of the strategic linkages that tie them together and to the
economy as a whole. Because the economies of individual countries
increasingly depend upon each other, national policy choices must be
evaluated in light of their international trade and security consequences.
The second issue centers around the importance of final demand for
electronic products. Market demand plays a critical role in guiding
technological development and providing the resources to foster and sustain
high technology sectors. The case of the semiconductor industry
illustrates this issue. It underscores the importance of market processes
in channelling technology diffusion.

The third issue is about compatibility and connectivity in a world
where information systems increasingly need to communicate. The story is
told through the case of the computer industry. It stresses the role
policy can play in helping to diffuse the benefits of computer technologies
by promoting the emergence of common standards.

The fourth issue follows the emergence of digital integrated
communications networks. These new networks constitute the infrastructure
of economies increasingly built around electronics-based information
technologies. National policies and regulations will powerfully shape the
development of this electronic infrastructure, and thereby affect the
diffusion of the products and services it supports.

The fifth issue is about the skills required to carry out the
electronics transformation. It is illustrated through the case of
manufacturing automation. Machines will not eliminate labor, but their
development, implementation, and operation will require new skills.

Technology diffusion will rest on the training and education provided not
only to a few scientists and engineers, but to the large population of
those who implement the transformation.

The choice of these five themes is not arbitrary. They correspond to
the fundamental dynamics driving the development of the electronics
sectors, and the diffusion of their products and processes. Neither is
each dynamic specific to the particular case we use as an illustration.

Final demand is as critical to the computer industry as it is to the
semiconductor sector, and skills matter as much in telecommunications as
in robotics. Indeed, each of the five issues runs through all facets of
the electronics transformation, and invariably underscores diffusion as
the ultimate policy goal.

I. THE DIVERSITY OF INFORMATION TECHNOLOGY
STRATEGIC LINKAGES AND POLICY CHOICES1

A. The Electronics Sector: from Microchips to Robots and Computers

1. Enabling technologies: Semiconductors and Software

The entire electronic industry --from computers to digital watches and
robots-- rests upon two basic technologies: the integrated circuit (IC),
and software. They constitute the essential building blocks of any
electronic system. Importantly, the two technologies are intricately
related and interdependent.

The IC is singlehandedly responsible for the dramatic cost decrease
and performance increase of electronic systems over the past decades:
following the law Gordon Moore spelled out in 1964, the number of
components (transistors) integrated in a single circuit is roughly doubling
every year. A memory chip capable of storing 4,000 bits of information
used to cost $4 in 1978 ; today one that can store 64 times more
information (256K dRAM) costs less than $2. Without the IC, many products
(such as radiotelephones, compact disc players or personal computers) would
simply not exist today.

But most importantly, ICs have added tremendous capabilities to
innumerable traditional products. Whether they regulate the performance of
an automobile engine, control the routing of a telephone call, fine-tune
the sails of an America's Cup yacht, or guide the work of a machine tool,
they underlie the single most important transformation of products and
processes in recent times.

Software is the other indispensable element. For one thing,
integrated circuits are useless without the programs and instructions that
guide and regulate their operations. However, the interdependence of ICs
and software goes far deeper. Without sophisticated Computer Aided Design
(CAD) programs, engineers could not design and test today's increasingly
complex microchips. In turn CAD programs can only run on powerful
computers built around the mighty chips they have helped to design.
Complex chips, such as microprocessors and some custom ICs, offer
perhaps the best illustration of this truly synergistic relationship
between software and silicon. To a large extent, the software that
instructs them is built into their circuit architecture, embodied within
their physical design. Furthermore the programs they are able to process
depend largely upon their logic and physical layout. Hardware and software
have become so inextricably entangled that success in electronics requires
both the manufacturing expertise to etch ever smaller lines on ever smaller
silicon chips, and the software skills to conceive the design, testing, and
application programs.

Beyond its intricate role in the IC industry, software is becoming an
increasingly critical enabling technology. Like ICs, software pervades all
sectors of the electronic industry. Whether it is an Operating System that
controls a computer's inner workings or an application program used for
word processing, the instructions that guide a robot arm or the complex
rules that manage public telephone switches and keep track of phone bills,
software is an integral part of all electronic products. The performance
of these products, their development costs, and their ability to answer
their users' needs depend critically on the software they use.

2. Systems and Applications

The possible combinations of these two basic ingredients, chips and
software, are endless. Most electronic products are basically built of ICs
cleverly arranged on printed circuit boards and stuffed into some kind of
box (today increasingly arranged in various types of packages), ruled by
a series of programming instructions. The number of different systems and
applications they represent is staggering. Following is a list of the
major segments of the electronics industry.

Consumer Electronics: An increasing number of everyday products
relies on electronics to operate. A short list includes: Video Cassette
Recorders (VCRs), electronic watches, kitchen appliances, pocket
calculators, burglar alarms, hi-fi systems, thermostats, television, etc.
Telecommunications: Electronic technologies have deeply transformed
the telecommunications industry. Public and private switches, the network
nodes that route phone calls and manage network operations, increasingly
relie on digital technologies. Transmission links, whether copper cables,
fiber optics or microwaves, integrate advanced ICs in their repeaters,
multiplexers and satellite transponders. The terminals attached to the
telecom networks, telephone handsets, facsimile machines or videotex
terminals, are now built around specially developped microchips.

Computers: The dramatic decrease of ICs' cost performance ratio has
fueled a double evolution of the computer industry, away from mainframes
which used to constitute its largest segment. At one end, mini- and micro-
computers have brought computers closer to their users, and allowed the
decentralization of processing capacity. At the other end, ever faster
chips allow supercomputers, such as the Cray XMP, to perform 200 Million
instructions per second (Mips). Computers and telecommunications together
form the basis of office automation technologies.

Production automation: Electronics now control the operation of
machine tools, and increasingly allow them to sense their environment.
Software can instruct a machine to change from one task to another, or to
work in a different way. Sensors and computer vision, coupled with
artificial intelligence applications, allow robot arms to adapt to
unpredictably changing tasks. Through local area networks, computer aided
design (CAD) becomes progressively integrated with computer aided
manufacturing (CAM) to automate the entire manufacturing process.

The distinction between these various electronic sectors is fading
rapidly. New words, like "telematics" or "mechatronics", attempt to grasp
the convergence of formerly distinct areas. Indeed, nothing fundamentally
distinguishes a telephone switch --a computer that manages a telephone
network-- from a computer that keeps track of a bank's accounts, or a
computer that controls a machine-tool. Communications networks now
routinely transform voices into streams of digital bits: people and
computers are made to speak the same digital language, and their messages
--voice, images or data-- travel alike throughout an increasingly
integrated and digitized network.

The convergence of formerly distinct electronics sectors goes even
further. Not only do the products look and work alike, they also
increasingly rely upon one another to operate. Today's computers, because
they have become smaller and decentralized, need communication networks to
work together. Indeed, the network itself is so much a part of the
computer's operations that is is no longer possible to tell where the
computer ends and where the telecom network begins. The same is true of
robots spread throughout a factory: they need a network that coordinates
their operations, transmits CAD designs from the research labs, or ties
them into the broader operations of the firm.

B. The economic impact of IT

The direct impact of electronic technologies on economic development
first derives from the sheer size of the electronic sector. By any
account, this is one of the largest industrial sectors in the developped
economies, one which now rivals the largest traditional sectors, such as
automobile. Such comparisons indeed provide a striking illustration of the
sector's size: in 1986, the shipments of the US automotive industry
totalled $162.4 billion ; that same year, the US electronics industry
shipped $198.3 billion worth of products2.

Although the precise numbers depend on the definition they adopt,
various sources estimate the global revenues of the world information
industries around $400 billion in 1986. Moreover, whatever the cyclical
variations of individual sub-sectors may be, electronics as a whole are
experiencing a spectacular growth rate. The trade press calls it a bad
year when, like in 1986, the global electronics markets only grow at an 8%
rate. Predictions see the world electronics markets reaching the trillion
dollar mark by 1990.

The economic impact of electronics reaches far beyond electronics-
based sectors such as the semiconductor, computer and telecommunications
industries. The diffusion of their products and the new production
processes they make possible holds even greater promise for economic
development. Electronic technologies and products increasingly pervade the
economy, to such extent that it becomes hard to distinguish between high-
tech and low-tech sectors. If microchip producers and those who build
computers around these microchips clearly belong to the electronics sector,
what about an automobile company which uses robots to make cars, implants
microchips inside its carburators, and spends money on research ranging
from programming langages for robots to the design of on-board computers?

For the entire economy, electronics are transformative technologies3:
electronic industries are developing products, production processes and
technologies that radically transform the structures and the organizations
of production and exchange activities. Indeed, despite the popularity of
home computers and video cassette recorders, most electronic products are
producer goods, not consumer goods. They are bought to be integrated in
the products of other industries (like microprocessors in autos,
appliances, airplanes or toys) or in the production process (like robots,
computers and lasers accross the range of manufacturing and services), or
both.

To remain competitive, that is, to survive, traditional industries
must assimilate the new technologies, design products that make use of
their possibilities, develop production processes that harness their
potential, and use electronics to create flexible organizations that can
swiftly detect and adapt to market changes. It is not simply a question of
placing new NC machine-tools in old factories, but rather of reorganizing
the production process around the possibilities electronics-based
technologies have opened. Beyond individual firms, the applications of
electronic technologies are transforming the organization of economic
activity. They help invent new ways to conduct business, tap resources,
access markets, coordinate workforces and equipments, or link various
organizations along the new infrastructure of information networks. Indeed
the most powerful economic impacts of electronics stem from such diffusion.

Economic development will go to the economies that are best at using
computers, and not necessarily to those that are best at making them.

C. A strategic Technology: control of and access to electronics

Competitiveness, the degree to which a nation can, under free and fair
market conditions, produce goods and services that meet the test of
international markets while expanding the real income of its citizens4,
rests on the capacity to diffuse electronic technologies. National and
economic borders matter critically to this ability, therefore to economic
development. In both the business and military sense, electronics is a
strategic technology: competitive advantage lies with those who have access
to and control over electronics technology, thus within nations which
posess a dynamic electronics sector. The need for interactions between
producers and users of electronics and the vulnerability that derives from
technological dependence both account for this strategic dimension.

1. User-Producer Synergies: the required interactions

The ability to apply electronic technologies, whether to a new product
or a new process, is tightly linked to the ability to develop and
manufacture new electronic products. The conception and design of IC-based
products requires an intimate knowledge of which ICs are available, and
what their capabilities are. Conversely, IC producers need a detailed
understanding of the potential applications in order to design circuits
that will answer their users' requirements and find markets. Similarly,
industrial robot manufacturers must collaborate tightly with their
customers to grasp end-user needs ; robot users must know precisely what
robots can and cannot do to integrate them efficiently within reorganized
production facilities.

Such collaboration between users and producers requires sustained
interaction. It rests upon a close knowledge, efficient communications,
and precise understanding of what each other's needs and constraints are,
all things more easily achieved within a single country. If no American
robot maker offers the machine a small US textile firms needs, it will have
to use a Japanese (or German) substitute, designed for and in collaboration
with different users. The needs of the American user probably differ from
those of foreign users who use the machine within a different industrial
organization, to achieve their own specific goals. Chances are the foreign
robot will not perfectly fit its application, hindering the robot user's
competitiveness.

Furthermore, borders can only slow down the diffusion of a new
technology. Users from its originating country will therefore be the first
to use it, and thus have a chance to create a competitive advantage over
foreigners who will only use it later. In the electronic sectors, where
everything changes rapidly, such a delay can make all the difference
between success and failure.

Clearly however, most countries cannot possibly cover the entire range
of electronics sub-sectors with national firms able to remain
internationally competitive at the leading edge of technology. The
difficulties some faced as they attempted to foster the development of a
complete electronics "filiere" underscores the danger of spreading too
thin; attempting to grab too much, countries risk to grasp too little.
The question then arises for policymakers to assess the importance of
specific linkages between certain segments of the electronics industry
and the rest of the economy, to choose where and how governmental action
should be focused.

Specifically, this raises three sets of policy questions. First,
under what circumstances and in which particular sectors does the
development and diffusion of technology requires such close ties between
users and producers? Second, to what extent are those ties required within
a single country? and third, to what extent is this a regional problem,
one for example that could be resolved through interacions between French
and German firms, but not between European and Japanese firms.

2. The Vulnerability of Dependence

To depend upon someone else's technology for one's own competitiveness
can generate vulnerability, or at least the fear of vulnerability. This is
true both for countries and companies, and has international as well as
domestic policy implications. The question is one of industrial structure:
firms unable to produce the critical components they need for their systems
will have to buy them either from merchant suppliers or from integrated
suppliers. Nationally, this can represent a difficulty when the best
components are produced by an integrated domestic firm that makes them for
its own use, and relies upon their superiority for its competitveness in
the final systems markets. Internationally, trade issues complicate the
problem: firms from countries unable to produce the critical components
they need for their systems will have to buy them elsewhere. Whether the
foreign suppliers are merchant firms or integrated producers, trade
policies often impose further restrictions, making it even more difficult
for user firms to obtain the latest generation components.

The two related questions of industrial mix and structure compound
each other both at the domestic and internatinal level, because most
electronic products are primarily intermediate products. Access to state-
of-the-art electronics critically determine the competitiveness of the
firms who ultimately use them in their products (like the microchips used
in a VCR) or their production processes (like the numerical controls of an
assembly line). Companies may be denied such access because of trade
restrictions when only foreign firms make the product they need ; or
because the domestic integrated firms making them also make the final
products that use them (IBM will of course not sell the proprietary ICs
that embody the advances of its latest Personel System/2 computer, even to
US firms) ; or because the supplier is both foreign and an ultimate
competitor (the same Japanese companies make VCRs and the components that
make their VCRs superior ; understandably, they won't sell those components
to european VCR producers). In all cases, products designed around
inferior components or manufactured with older processes will be at a
disadvantage in international competition.

To this purely economic logic, governments must also add an
inseparable military dimension. Today's weapons and intelligence systems
relie on technologies essentially similar to those of commercial
electronics5. The new COCOM rules on technology export imposed by the
United States in February 1985 stress the link, as they restrict exports
of the so-called "dual technologies". The US strictly controls sales of
primarily commercial electronic products which have potential military
applications, not only to Eastern Block countries but also to its allies.

Whether the reasons behind these restrictions are purely military or
combined with more commercial purposes, the consequences remain the same:
they underscore the strategic necessity to secure a reliable access to
advanced electronic technologies.

Yet among today's open economies, international sales must, and will,
occur. Foreign markets are not simply tempting opportunities, but generate
the necessary resources to sustain domestic growth. Foreign sources of
capital, technology and know-how have become indispensable to development
in all countries. Indeed, because national economies have grown so
increasingly interdependant, a balance needs to be struck between
commercial, strategic and trade imperatives. In this balancing act
however, it is important to understand the stakes.

Issues for discussion

* If single countries cannot support the entire range of
electronics sectors, how should they assess their relative
importance to select those they should promote?

* What are the trade-offs implied by specific national market
access restrictions and promotion strategies? In particular,
how do one country's choices affect the options of other
countries, and how can domestic policy decisions be reconciled
with the imperatives of an open trading system, and of national
security?

II.SEMICONDUCTORS
   THE IMPORTANCE OF FINAL DEMAND6

Studies of the electronics industry always make a special place for
the semiconductor industry. In large part, as pointed out earlier, this is
due to the critical importance of ICs as a fundamental enabling technology,
to the fact that ICs are the basic building blocks of any electronic
system. In other words, what products can be made and how efficiently they
can be produced is largely shaped by one's mastery of integrated circuits
technology.

Final demand has been one of the most important factors in the
evolution of the semiconductor industry: its volume determines the market
resources available to IC makers for research and development, and, most
importantly, its character has shaped dramatically distinct technological
and commercial strenghs in various countries. Success stories and
failures in the semiconductor industry highlight the determinants of
success and failure in electronics as a whole. Intersectoral similarities
stem from the fact that electronics products are intermediary goods.

Access to the best components determines the users competitiveness, and
they will strive to secure an adequate supply. Their success in doing so
will be largely constrained by the structure of the national merchant
industry, and by policy restrictions on international trade. Similarly,
integrated IC producers need to generate market revenues without giving up
their strategic technologies. Here again, domestic and international
policy issues interplay, industry structure and industry mix jointly shape
the impact of final demand through the market. The resulting dynamics are
not easy to unravel, but ultimately determine both a country's access to
the enabling microelectronics technologies, and its ability to diffuse
them.

From the inception of electronics up until very recently, the United
States was without contest the world leader in the semiconductor industry.
This superiority rested upon a solid foundation of advanced technology,
developed in response to strong demand for advanced circuits, first from
the Department of Defense, then from the computer industry. The size of
this considerable final demand was decisive in fostering the development
and growth of the US merchant semiconductor industry.

During these formative years, AT&T's Bell Labs also played a critical
role. Because of antitrust controls imposed on AT&T, the Bell Labs were
obligated to license cheaply all the technologies they discovered and
developed. Pulled by growing demand from DoD and the computer industry,
pushed by rapid technology improvements financed by the market or bought
cheaply from AT&T, the American IC industry made spectacular progress
in integration and cost reduction.

However, these very sucesses induced new problems. With the advent of
large and very large scale integration, new manufacturing technologies make
it possible to produce increasingly complex circuits at very low unit cost.

However, those circuits must be manufactured in ever larger quantities to
spread the growing costs of research and development . At the same time,
because circuits become more complex, they tend to become more specialized
and can fit fewer specific applications. IC producers risk being squeezed
between these two trends, having to produce ever larger quantities of chips
at ever smaller unit costs for ever narrower niche markets.

This vicious cycle fuels the double evolution of the IC industry,
where two related but distinct technological trends co-exist. The first
trend pushes the industry towards a more "mature" phase, where the game is
to produce large quantities --at low unit price-- of relatively simple and
standard components. By contrast the second trend emphasizes innovation,
characterized by the growing number of new niche markets for complex custom
ICs designed to answer the needs of specific users. Business strategies
and industrial policies in the IC industry must be developped around these
two trends. Although both avenues --standard and custom-- rest on tightly
related fundamental technologies, success in each of them requires
distinctly different sets of skills, organizations, research and
manufacturing decisions.

A. Maturity: Commodity products in an adult industry

Traditionally, memories have been the largest product segment of the
IC industry, accounting for over 20% of the world semiconductor market7.
This has made the memory markets important for two reasons. First, they
have traditionally generated the bulk of the industry's profits to be
reinvested in research and development. Second, the volume production
technology RAMs required, stimulated (and funded) the development of
advanced IC manufacturing technologies, that could in turn be applied to
the production of all other ICs.

As the memory industry matured, manufacturing expertise, the capacity
to produce large volumes at low unit-cost, and the commercial ability to
sell to a mass market became essential. In such a mature sector,
production strategy and capital investment matter more directly than
product innovation. This advantages large diversified industrial groups,
such as many Japanese companies, which can draw resources from other
divisions (consumer electronics for example) to invest in IC production.

They can therefore afford the highly automated production lines mass IC
production requires well before they control the mass markets that will
justify and support such expense. By contrast, most American merchant
semiconductor producers, who specialize in IC manufacturing, could not
afford such a strategy. The few integrated US firms, such as IBM, only
manufacture ICs for their own consumption, and therefore do not directly
intervene in those markets. Thus, Japanese producers gained control over
the most advanced mass manufacturing technology, that later enabled them
to produce larger quantities of standard circuits at lower unit cost than
their US competitors.

B. Innovation: Components become systems

Faced with declining margins on commodity ICs and the Japanese
penetration of the mass markets, the US response centered around new
complex components. The challenge was to find a way to satisfy the
specific needs of a range of niche markets, while producing large enough
quantities of each circuit to benefit from scale economies. Increasing
integration made it possible to place a growing number of a system's
components on a single silicon chip. A chip that used to be simply a
component within a larger system became capable of containing a complete
system itself.

The solution was to manufacture large quantities of complex circuits
which can in some way be adapted, customized to specific uses. Micro-
processors, EPROMs and EEPROMs*, as well as Programmable Logic Devices
(PLD) offer such a solution because they can be programmed after
manufacture to execute various tasks. Semi-custom, standard cell, gate
arrays and full-custom chips offer another type of solution, as they permit
customization at diverse stages of the manufacturing process.

Critically, the very nature of these complex circuits tends to
represent a considerable obstacle for foreign suppliers. Indeed, because
the components have become systems, their physical design increasingly
embodies information and concepts that are essential to the final systems
that will use them: a personal computer's capabilities are more or less
those of the microprocessor and supporting circuitry it centers around. It
then becomes dangerous to entrust their manufacture to a foreign company,
which could use the strategic information they contain to successfully
compete in the final systems markets.

C. Final Demand Shapes IC Strenghs

The respective strenghs of American and Japanese IC producers owe
little to chance, nor to genetic or cultural differences that would make
the Americans more inventive, and the Japanese better at technological
imitation and mass production. First and foremost, they reflect
differences in the nature of the final markets that induced and sustained
the development of each national industry. Early Japanese inroads into
semiconductor markets were unequivocally tied to final demand from the
consumer electronics sector: portable radios, color TVs, pocket
calculators, digital watches or VCRs provided markets and revenues for the
development of relatively simple microelectronic circuits. By contrast,
used to supplying complex circuits to their major client --the computer
industry-- the American IC makers were well positioned for the transition
towards complex, application specific circuits.

Obviously, the structure of final demand creates clear economic and
market incentives for the national IC industry to design and develop the
kinds of circuits that national system manufacturers will buy. Beyond this
however, the character of final demand further fashions the IC industry
through the sustained interaction it requires between IC users and
producers. Indeed the recent Japanese inroads in IC markets traditionally
controlled by US producers can be traced to the changing structure of
Japanese final demand for ICs, progressively resembling its American
counterpart. In 1980, consumer electronics used 58% of the semiconductors
sold in Japan, while industrial electronics, computers and communications
only used 42%. In 1985, the share of consumer electronics had dropped to
about 40% 9. Similarly, the European weakness in IC largely results from
the fact that European firms, from Swiss watchmakers to French TV producers
or German telecom equipment providers, generally ignored the potential of
microelectronics, and failed to generate a strong European demand for
integrated circuits and the revenues European IC makers would have needed
to invest.

When they tried to address these problems, European governments
typically subsidized the IC suppliers. Never did they explicitely try to
foster an independant European demand for ICs. In fact, they only promoted
IC demand through National Champions (such as Thomson, ICL or Siemens) who
had to buy preferentially from their country's IC manufucturers. Such
protection compounded the lack of demand pressure on European IC producers.
It further isolated European chip makers from international competition,
in striking opposition with their American and Japanese counterparts.

Issues for Discussion:

* How can governments stimulate final demand in ways that will both
stimulate development of the electronics industry and the
diffusion of their products?

* If choices have to be made, should governments support the users
or the suppliers of electronics, given their specific character
as intermediary goods?

III. COMPUTER INTERCONNECTION
     THE IMPORTANCE OF STANDARDS

Interconnectability and compatibility, based on common standards, are
now critical. As more powerful mini- and micro-computers decentralize
information processing, they increasingly need to be connected.

Electronic data processing systems are no longer stand-alone machines, but
intelligent networks linking decentralized processing capabilities. The
issue of standards is therefore of central importance, not only for the
development of the computer industry itself, but most importantly for the
diffusion of computer products and technologies to the entire economic
fabric.

This view considers computers as constituent elements of an integrated
system of production. Such network systems are characterized by network
externalities: the benefits derived by one user of the system increase with
the number of other users. Conversely, the users excluded from using a
system because it follows a different standard suffer a direct efficiency
loss ; such is the case of the owners of Apple micro-computers who are
denied access to the vast library of IBM-compatible software.

Indeed, users of computer technologies are largely at the mercy of the
social mechanisms entrusted with providing compatibility between the
various components of the systems they use. Moreover, analysts widely
recognize that markets left to their own devices usually result in an
insufficient degree of standardization, and induce losses of efficiency
from an overall economic point of view. Governments therefore have an
opportunity to affect the global welfare of the economies that produce and
use these computer systems by indirectly channelling the market-driven
processes that shape standards in emerging technologies, or by directly
specifying the characteristics of technological products. They can design
policies that promote cooperative standard setting among firms, or more
simply mandate compliance with government defined standards.

In these efforts however, policy makers face three dilemmas. The
first, to use Paul David's terminology10, is the "Narrow Policy Window
Paradox": policy intervention is most effective at the beginning of the
technological evolution, and this only during "narrow windows", or very
short periods. The second dilemma results from the "Blind Giant Quandary":
public agencies, those entrusted with developing the standards, are most
powerful when they know the least about the technology.

To escape from these first two dilemmas, policy makers can only strive
to keep the policy windows open as long as possible, while the blind giants
try to learn more about what will make a "good" standard. At this stage,
any government action which prevents the industry from locking in on a
particular standard will be beneficial, even though the early indecisive
period may result in short term inneficiencies. Usually however, one
standard will soon become dominant, and confront policy makers with a third
dilemma as they need to cope with "Angry Orphans", those who had selected
the now abandonned standard. So as to maintain credibility for future
policy, and not to compound the risks users face when they must choose
among various emerging technologies, governments should favor the
development of ex post facto integration technologies -- various types of
adapters and translators may help. As much as possible, standards should
be developed that do not completely exclude alternatives to the dominant
solution.

The evolution of the computer industry provides a good illustration of
this framework, and helps outline the policy options that exist to promote
interconnectability. To a large extent, computer compatibility and
interconnectability are software issues. They underlie IBM's dominance
over the computer industry. By introducing the 360 series architecture in
the 1960s, the company was the first to provide a range of compatible
computers, that could accomodate its clients' growing needs without forcing
them to re-write their programs or re-encode their data. A self-
reinforcing market process was then unleashed. More (IBM) computers were
able to run similar programs, creating enormous opportunities for software
engineers to write IBM-compatible programs. In turn, the growing number of
applications written to IBM's specifications made it compelling for users
to buy IBM machines, increasing the installed base of IBM computers, thus
the market for compatible software, and so on... Critically for the users,
the costs of "translating" their programs and files to another computer
standard, and the cost of renouncing access to the IBM-compatible world,
far outweigh the costs of remaining faithful to IBM, even when other
manufacturers offer more advanced machines. Thus, IBM keeps its clients
firmly "locked-in".

Under IBM's supremacy, other computer makers face a tough alternative:
resist, like the (former) BUNCH, NEC, and most European makers who chose
non-compatibility ; or surrender, like Amdhal, Fujitsu, Hitachi, and all
the others who produce IBM-compatibles. Both strategies entail major
risks. Those who chose not to follow the IBM de-facto standard renounce
access to the major part of the market, and must fall back on specialized
niches. By contrast, those who decide to be compatible loose the
initiative, and restrict their options to strategies that are merely
responsive, rather than aggressive.

Compatible equipment makers have repeatedly tried to force IBM to give
them free access to such interface information, through antitrust suits in
the US as well as in Europe. They were unsuccessful however, since in 1982
the US Justice Department ruled, in favor of IBM, that these were
legitimate business practices. In 1984, the EEC followed a similar route,
although for quite different reasons, as it negotiated an agreement with
IBM. In essence, for those governments, the narrow policy window had
already slammed shut. In their view, the benefits of going along with a
firmly established standard justified sacrificing the requests of the
dominated actors, and the interests of a few angry orphans.

Where legal means have failed, the technological evolution towards
small decentralized computers may well succeed: it brings increased
pressure for open systems, in reaction to IBM's traditionally closed
systems policy. Metaphorically, economic power is following computing
power as it becomes more decentralized to rest increasingly with the users.

Indeed, companies and countries can no longer afford to use a multitude of
electronic machines unable to talk to each others. The grounds gained by
the UNIX operating system, as well as the agreement first ratified by 12
major European computer makers to support Open System Interconnection
(OSI), push to unify the computer indutry behind common standards that
escape IBM's control --and manipulations. UNIX was designed as a portable
Operating System, so that programs written for one computer can easily be
transported onto another. OSI, championed by the European Community,
specifies interconnection standards that allow computers and peripherals
of different makes to communicate over standardized networks.

Importantly, OSI is conceived and developed in a way that does not
exclude any particular computer system. Rather, it takes an approach that
fits around all of them and thus leaves no angry orphans but those who
refuse to play by the common rule. As interconnection technologies
evolved, the standard setting organization was able to work closely enough
with various computer makers to keep the policy window open while it became
less blind a giant.

Of course, it would be naive to only retain the impression of perfect
harmony that results from this simplified framework. In particular, it
leaves out what might be called the "political economy of standards"11:
various agencies involved in the standard setting process, various
governments representing the interests of different computer producer and
user communities, undoubtedly need to pursue a range of goals, not all of
which are compatible with global network systems efficiency. Similarly,
industrial actors do not forget their own interests through this process
and for example, IBM is certainly taking steps to gain more control over
the definition and evolution of the OSI standards. Nevertheless, the main
thrust of the argument remains. It underscores the benefits of a standard
setting policy flexible enough to keep open the windows of opportunity
while learning more about the technology, wide embracing enough to include
those who otherwise would have become orphans and, rightly so, angry.

Issue for Discussion:

* How can governments promote the emergence of standards that will
allow interconnectivity and thus help the diffusion of computer-
based technologies, without precluding innovation and
technological development.

IV. TELECOMMUNICATIONS
    BUILDING THE INFRASTRUCTURE

A major transformation of the telecommunications infrastructures in
the OECD countries is underway. Suppliers of telecommunications services,
producers of communications equipment, and major users of both services and
equipment are fashioning radically new, digital telecommunications networks
in the industrialized countries. National phone networks are being
digitized ; digital overlay data, facsimile and video networks are being
built, expanded, and in some cases integrated into the existing public
telephone network to form integrated services digital networks (ISDN).

Large corporate and public sector users are also building private digital
networks either wholly under their control or by leasing circuits from
public service providers. Over both the public and private networks, new
generations of telecommunications services like videotex, electronic mail,
voice messaging, high-speed data transmission and videoconferencing are
emerging. Complementing these changes are enormous new market
opportunities for service providers and for suppliers of the
telecommunications equipment which comprises the new networks and controls
the delivery of new and old services.

The new telecommunications infrastructure emerging from the
convergence of data processing and telecom technologies acts as a powerful
agent of economic development because it opens a series of new
opportunities. First, the network itself has to be built. This will
induce significant growth within the equipment and service industries that
supply the parts, assemble, operate and maintain the new telecommunications
network. Second, the network generates a set of lucrative and expanding
markets for terminal equipment and new services that can be connected to,
or delivered over the new infrastructure. In 1986, telecommunications
services and equipment together accounted for at least $115 million or
about 5% of U.S. gross domestic product (GDP), about $65 billion in Europe,
and $25 billion in Japan, or about 3% of GDP in each. Because the
telecommunications sector is growing much faster than GDP, it is expected
to account for between 7 and 10% of GDP in the advanced countries in the
early 1990s. Third, and most importantly, the shape and characteristics
of the new networks --and the pace at which they are created-- will affect
business strategies in all economic sectors, structure opportunities for
profit and growth, and influence who can capture these opportunities12.

Telecommunications networks constitute the infrastructure of the
information economy, much like roads and railroads do for the goods
economy. Because such an infrastructure determines what can be transmitted
(or transported), in which conditions, between which points and at what
price, it sets a new basis for economic activities : production processes,
exchange mechanisms, institutions and organizations, business strategies
or location decisions have to be developed and evaluated within a new set
of constraints and opportunities. Because they define a new topography of
opportunities, the new telecommunications networks raise important policy
questions. They have the potential to profoundly transform the very
structure of the economy, and redefine the basis of competitiveness. Yet
the mechanisms through which this transformation will occur are still
largely unknown.

Current evolution suggests that the shape, character and functionality
of the new networks will substantially depend upon the national environment
within which they emerge. Variations among the national regulatory
arrangements, differences in the structures of national telecommunications
industries, or distinctive legacies set by the traditional domestic
telephone networks may result in significant differences between the
telecommunications infrastructures of various countries. Here again, this
is simply our presumption, based on a preliminary analysis of recent
changes in international telecommunications13. The presumption, however,
is strong enough, and its implications momentous enough, to warrant a
thorough discussion of the policy options.

In turn, the competitiveness of domestic firms, from all economic
sectors, will critically depend upon the shape and characteristics of the
communication networks and services they can mobilize. Conversely, as
firms strive to secure a competitive edge over their foreign (and domestic)
competitors, they will attempt to shape the networks' evolution to their
advantage. Indeed, businesses are clearly becoming one of the major forces
driving the development of the new information infrastructure, one of the
major influences upon its configuration and characteristics. These
competitive forces interact powerfully with national telecommunications
policy to shape the emerging network infrastructure --with critical
implications for both policy and business strategy. Telecommunication
policies need to be examined anew in the light of these changes. The
development of these policies requires an understanding of the specificity
of the telecommunications infrastructure.

The telecommunications network infrastructure is repeatedly compared
with the earlier transportation infrastructures, roads and railroads. The
analogy carries considerable suggestive power. Remembering how deeply
roads and railroads have transformed the economy and its "geography of
opportunities" suggests the profound transformation a new infrastructure
will yield. The analogy however, should often be left here -- a powerful
image. Attempts to evaluate telecommunications policies within a framework
directly inspired by transportation networks quickly run into major
problems. Indeed, the telecommunications infrastructure differs from the
previous transportation infrastructure in several important ways. These
differences matter to the role of policy in the process and dynamics of the
networks' emergence. We focus here on five of these differences14.

First, the influence of transportation technology on distance is
different in kind from that of telecommunications. Improvements in
transportations merely "stretched" geography : A rubber band on a map
showing the area within reach would be stretched as transportation
technology improves. Suddenly with telecommunications, the accessible area
so stretches that the rubber band breaks : it is exactly as fast, and soon
will be equally cheap, for somebody in Berkeley to transmit voice or data
to San Francisco, Paris or Tokyo. Transportation technologies yielded
incremental improvements in accessibility, but telecommunications provides
a quantum leap towards ubiquity. National telecommunications authorities
then loose a great deal of their control: a multinational dissatisfied with
the high tarriffs imposed through one country's telecom policy can demand
changes and threaten to transfer instantly its telecom traffic abroad.

Second, the transportation analogy implies a false idea of
homogeneity. It views telecommunications access as uniform, much like
access to a road. Once laid out, the road network provides essentially the
same service to anyone connected to it. Widening a two-lane country road
to a four-lane freeway may reduce traffic jams, but does not inherently
change transportation. By contrast, access to a wide-band optical fiber
link rather than a simple telephone line makes all the qualitative
difference between the ability to transmit video images and data, and being
restricted to voice communications. Basic network access no longer
suffices for firms seeking to take advantage of telecommunication's
competitive potential: to implement their strategies, a twisted pair will
not substitute for an optic fiber. Telecommunications policies, by
providing (or not) access to various types of network facilities, by
regulating who will have access and at what price, will therefore directly
affect industrial and economic development. Fundamentally, these are
society-wide, hence political decisions.

Third, in a transportation system, technological improvement is
primarily embodied in the vehicles, and therefore diffuses instantly and
uniformly over the entire network : faster, more efficient trucks
immediately improve the whole transportation system. By contrast,
technological capability in telecommunications and service applications is
embodied within the network itself (software-controlled digital switches,
wide band optical fibers, "intelligent" multiplexers) and benefits only
those who have access to the more advanced portions of the network. With
roads and rail, it was important to control the vehicles and the vehicle
technology. With telecom, competitive advantage and power rest with those
who control the infrastructure itself, whether they are private or public
entities.

Fourth, telecommunications is a "soft" infrastructure, one built with
software as much as hardware. Applications developed over the network
(electronic-mail, packet switching, VANS, video conferencing,...) are
critical parts of the infrastructure, and inseparable from its hardware.

While ubiquitous connectivity tends to make all locations more alike, the
services and applications available over the network introduce major
differences: having access to the right network hardware is not enough, one
needs access to the right applications. The case of American Airlines
illustrates the point. With SABRE, it was first to offer travel agents on-
line computer access to its data listing flight and reservations
information, for all airlines. Of course American's flights were
systematically displayed in a prominent position, placing competitor
airlines that did not have their own system at a competitive disadvantage.

Travel agents, therefore customers, were connected to all airlines, but
trapped in American's application. Here again, competitive advantage will
rest with those, private or public, who control the "soft" side of the
infrastructure.

Fifth, a telecommunications network is a non-standard infrastructure.
With minor restrictions, trucks and trains can technically travel anywhere
along a continent's roads and railroad tracks. Not so with
telecommunications, where standards often constitute major barriers to
network access. There is not really one telecommunications network, but a
series of juxtaposed sub-networks with various degrees of interconnection.

Importantly, standards affect both levels of the infrastructure: hardware
and software. As they decide --or let the market decide-- how to set such
standards, policies will have critical economic consequences.

Recognizing that telecommunications networks are not homogeneous,
universally available "public" goods, brings new questions to the
attention of policy makers: Who controls the design of the networks? Who
controls their construction? Who controls what applications they will
support? Who has access to which networks? And what motives guide each of
these actors? What objectives do they pursue through the construction and
use of the networks? More than anything else, the policy answers to these
questions determine the evolution of the telecommunications infrastructure,
its shape and characteristics. They will in turn largely affect the growth
of the economies that rely upon these infrastructures.

Telecommunications has traditionally been an important area of
government involvement. In all developped countries, until recently,
state-owned or state-regulated monopolies developed, built, controlled and
managed the telecommunications network. With a handful of domestic
equipment suppliers, the monopolies would agree on what technologies to
promote, what products to design, what to sell and at what price. This had
important consequences for the telecommunications infrastructure it
generated. The rationale guiding network development was often that of the
state administration rather than the users'. Network management practices
and service offerings reflected the concerns of the monopoly network
operator more often than those of clients who had nobody else to turn to.

Centrally imposed standards guaranteed uniform access and equipment
compatibility anywhere within a single national network, but also served
to limit foreign penetration of the domestic markets, making international
communications all the more complex.

This cozy relationship tying governments and the telecommunications
sector is now changing rapidly and dramatically. AT&T is no longer a
monopoly in the US ; British Telecom in the UK and NTT in Japan have been
privatized and are facing competitors for the first time ; other
traditional monopolies are today threatened. The barriers governments once
erected to protect their national telecommunications industries are
shattered, giving way to intense international competition. New
technologies and new regulatory frameworks enable private and public
organizations to build, control and manage their own communications
systems, reducing the share of the network under public control.

Yet, if the role of government is being re-defined, telecommunications
policies and regulations retain considerable implications for the shape and
characteristics of the emerging networks. Deregulation in the United
States means that by and large, the networks will be shaped by the needs
of large users. Developmental re-regulation in Japan accompanies a
deliberate policy decision to build an advanced integrated digital
network in anticipation of its use. The European PTTs, who retain the
tightest control over telecommunications, appear determined to insure
that the telecom network will offer equivalent service to all, including
residences and small businesses.

These policies obviously have very different implications for the
future of each national telecommunications infrastructure. The trade-off
they imply, in terms of who has access to the networks and who controls
their evolution, are equally obvious. Their long term consequences are
less clear, and it remains debatable which policy will best serve long run
eonomic development. It is clear however that advanced networks will be
critical and that policies should be developped around the goal of
providing economies with an infrastructure able to support their
development.

Issues for Discussion:

* How should governments re-define their role in the
telecommunications sector to foster the emergence of the new
network infrastructure?

* If the networks emerge nationally, shaped by primarily domestic
dynamics, what are the implications for international trade
relations?

V. MANUFACTURING AUTOMATION
   LABOR RELATIONS AND SKILLS FOR FLEXIBILITY15

The diffusion of information and communications technologies in the
developped economies remains at a very early stage. Potential uses of
existing, not to mention future, electronics technologies, have yet to be
fully explored. Earlier fantasies of automation had predicted sweeping
economic reorganization, massive job displacement, tremendous productivity
increases, radical and almost instantaneous lifestyle changes. Both the
reality and the consequences of the applications of electronics to economic
activities have emerged more slowly and differently from what was expected.

Indeed the diffusion of new technology cannot be simply deducted from its
purely technological potential. The technology itself merely defines a
domain of possibilities. Within this domain, labor relations, business
strategies, management practices, skill structures, financial channels and
military strategies interact with technology decisions and policies to
frame complex channels for technological diffusion, even affecting to a
large extent the path of future technology development itself and shaping
the technology frontier.

Production automation is perhaps the area where the transformation
seems the slowest to materialize. If it falls short of past futuristic
dreams, the reorganization of the production system is nonetheless complex
and powerful. The current diffusion patterns deserve careful study because
they create channels, methods and habits that will condition technology
diffusion in the future. We focus here on the role of labor organization
and skills in the diffusion of electronics-based manufacturing technology.

To understand the role of this "human element", and the broad potential of
production automation, we first need to step back and to grasp two distinct
elements. The first is the purpose that drives the adoption of the new
technology, summarized in one word: flexibility. The second is the
industrial tradition that pre-existed the introduction of the technology
and that constitutes a country's legacy in responding to the transition.

Flexibility has become the slogan and the goal of today's application
of electronic technologies throughout the factory, and the theme of a large
literature on production automation. Firms seek both static flexibility
(the ability at any time to adjust business operations to shifts in the
market), and dynamic flexibility (the ability to design production lines
that can quickly evolve in response to changes in either the product or
production technology). Production Automation (PA) is expected to allow
firms to adjust output levels and to produce several different products on
a single production line (static), and to "make rapid changes in production
technology to lower costs and thereby improve productivity" (dynamic)16 .

Programmable automation, and the flexibility it permits, has major
advantages. First, it increases the advantages of batch production over
mass production. Batch production becomes feasible in situations where
costs had previouly required the rigidities of mass production. "since
approximately 75% of all machined parts are produced in batches of fewer
than 50, the potential uses of mechanization are widespread"17. Because
the equipment is controlled by an electronic program, set up time and the
cost of shifting between uses are dramatically reduced, yielding economies
of scope along with economies of scale.

Second, machines can perform more sophisticated tasks than before
because more advanced sensory techniques are possible. Machines will also
be used in dramatically new ways. CAD speeds up and sophisticates product
design, and design testing ; it reduces the cost of design and speeds the
shift between design and manufacture. Introducing new products, or
designing a range of related products, becomes faster and cheaper.

The challenge is not simply to replace old equipment and labor by new
machines within existing production system. The new equipment is part of
introducing an entirely new production system. Greater benefits will be
captured if the new technology is not simply used to automate existing
practices, but to permit new ones. The benefits of a single PA machine
taken in isolation are nothing compared to the benefits from a new
production system organized to take advantage of the PA machine. Indeed,
the real potential of the new production equipment comes from its
integration: fully integrated production systems linking design with
manufacturing, permitting an automatic shift from one product to the next.

Because the objective of the new production organization is
flexibility, it necessarily results in lower scale economies, as the cost
of individual pieces of machinery rises while it can only be spread over
shorter series. The integration of production automation however will
result in economnies of a different kind, that we at BRIE have called
systems economies. These result directly from the flexibility of a
complete integrated manufacturing system, that minimizes the time lost
between each step of the production process, and during the reconfiguration
of this system. Importantly, the benefits of such a system cannot be
fragmented, and must be understood within the production process as a
whole. Such Computer Integrated Manufacturing (CIM) systems are still a
long way off. But manufacturing practice and the use of manufacturing
automation are evolving rapidly.

The current period of economic transition is a time when dynamic
flexibility is of predominant importance. the transformation doesn't
simply mean that a few "sunrise" manufacturing sectors, such as personal
computers, are assuming the importance once held by traditional
manufacturing sectors, such as automobile. Rather, computers and
microprocessors have begun to alter the production process throughout
industry. The transformation is occurring because the electronic
technologies are agents of change, sources of innovation, within the
traditional sectors. The critical question is how the new technologies
spread throughout the economy as a part of national and corporate responses
to changing competition. Neither markets nor technology will dictate the
decisions. Rather, political, economic and strategic choices will
determine the path of technological development. Manufacturing Matters18
builds an argument explaining why specific national development emerge out
of these choices, and the legacy they constitute. Similar notions are
expressed by Richard Nelson19 and Giovani Dosi20 as "technology
trajectory".

The technology trajectories of different nations, as evidenced by the
patterns their industries follow in adopting programmable automation,
differ widely. Let us simply contrast the cases of the United States and
Japan. Per capita expenditure on industrial automation is roughly similar
in the two countries, with an advantage for the U.S.: $10.9 in the U.S.
versus $8.1 in Japan (far ahead of Europe, with $3.2) 21. However, the
picture changes completely when one looks at the specific technologies
employed. Advanced automation, including CAD/CAE, manufacturing planning
and control systems (MP&CS), robots, and Flexible manufacturing systems
(FMS), accounts for only 12.3% of total manufacturig automation in the US
in 1985, up from 4.3% in 1980. By contrast in Japan, these advanced
systems account for 31%, and have represented a consistently high share of
total automation since 1980, when the ratio was 36.2%.

The technologies of advanced automation are precisely those that allow
manufacturing flexibility. The rest of manufacturing automation consists
essentially of stand-alone numerically controlled machines. Therefore the
figures highlight two very different automation trajectories: a trajectory
of "rigid" automation in the US, contrasted with a trajectory of flexible
automation in Japan. Critically, the two trajectories rest on very
different approaches to labor. While rigid automation aims at the
elimination of labor from the production process, flexible automation uses
the machinery towards the different goal of swift adaptability and requires
broadly skilled production workers for its implementation.

Production automation technologies can fit into a series of distinctly
different economic and social settings. More importantly, the technologies
will be shaped by the context in which they emerge. Automation is used to
solve market, management, and labor problems, in ways that differ in each
country. Therefore, policy, market structure and labor arrangements will
shape the development of the technology differently in each national
context. In these respects, there is a lot to learn from the contrast
between American and Japanese policy. The following remarks first address
government policies directly aimed at automation, then consider the
implications of industry structure for diffusion, and conclude with the
role of labor organization and skills.

American policy in programmable automation has been largely conducted
by the Defense Department and was aimed primarily at the manufacture of
sophisticated weapons , from aircraft to tanks. Japanese policy in
contrast, has been aimed at the development and diffusion of commercially
applicable technologies. The policy of diffusion established mechanisms to
ensure that small firms could learn about the new technologies, find and
develop machines appropriate to their needs, and lease them on favorable
terms. The consequences are quite clear. American machine tool
manufacturers dominate production of larger machines used for the most
complex purposes. Japanese producers dominate the market for smaller
machines used in the broadest range of industrial purposes, thereby
controlling the mass market. They now sell about half of the NC machine
tools used in the US. Not surprisingly, the Japanese control precisely
that portion of the market that their policy addressed22.

The market structure, the mix of large and small firms in industries,
will likewise shape the ways in which the new technologies are used and
consequently the way they evolve. If economies of scale created a
technological advantage for large firms, today's automated production
technologies should permit small firms to design and develop products that
can be sold in competition with large firms. But evidence suggests that
fixed costs in marketing, distribution and finance are often more important
obstacles to new producers than production economies.

Thus, institutional supports --public or private-- are necessary to
help small firms firms harness the new production technologies. First,
there must be manufacturers of PA equipment suited to small firms. Second,
there must be a network of service companies to maintain the equipment.

Whereas large companies can provide in-house service, small firms often are
not able to do so. Third, there must be marketing channels and access to
credit for small firms, as well as equipment producers aiming to meet their
needs. Japan's policy of financing the diffusion of programmable
automation equipment to small producers creates such an environment.

Similarly, studies of Italy's small producers show a particular
institutional fabric that supports small firms23.

The existing pattern of labor relations, the arrangements between
labor and management, and the skills of the workforce will also shape the
diffusion of electronics-based production technologies throughout the
industrial sectors. Management favors the development and introduction of
technologies that fit its vision of how work should be organized, of how
control --and whose control-- should be established. Which technologies
are applied, how they are applied, is in large part a strategic response
to skill availability and prices. American shopfloor organization largely
reflects production strategies based on notions of economies of scale, with
narrow job definitions serving a rigid mass manufacturing system. By
contrast labor organization in Japan, which defines job responsibilities
broadly, is better suited to the adoption of the new technologies.

Moreover in Japan, the labor force is being broadly educated to understand
both the technologies and their applications.

The ability to diffuse the new electronic technologies in traditional
sectors is as vital as the ability to develop them in the first place.

Although advanced technological development requires an elite of scientists
and engineers, the diffusion of advanced technologies rests upon a broadly
educated and skilled population. A skilled and involved workforce will
help firms create the "dynamic flexibilities" required to sustain
productivity increases. Crucially, automation strategies seeking the
elimination of skilled workers directly threaten the firms dynamic
flexibility: indeed, their own skilled workers, not their robots and
engineers, often have the experience and know-how necessary to continuously
develop, absorb, and apply new production technologies.

If the new equipment is used simply to strip labor out of production,
to substitute directly capital for labor in existing production
organization, then PA is likely to be ineffectively used and its potential
missed. As the low-skill functions become automated, higher skills become
necessary. Static flexibility --the ability to vary product mix-- demands
workers trained to perform a variety of tasks. Dynamic flexibility 
--the ability to fluidly introduce process innovation-- demands broadly
trained workers, sufficiently well- versed for example in the fundamental
principles of basic math and science that they can easily understand and
adapt to the new technological regimes.

Issue for Discussion:

* The availability of a skilled workforce is critical to the
adoption of innovative production strategies. How can
governments develop educational reforms --broadly understood to
include adult training and retraining-- that will help their
country to meet this challenge?

1 The arguments of this section were first developed by BRIE with the
Institut Francais de Relations Internationales (IFRI) in Rapport Annuel
Mondial sur le Systeme Economique et les Strategies 85/86, Chap 3.4:
"Nouvelles Technologies et Strategies Economiques", IFRI, Economica, Paris,
1985.

2 The auto industry is defined to include automotive stamping (SIC 3465),
motor vehicles and car bodies (SIC 3711), truck and bus bodies (SIC 3713),
parts and accessories (SIC 3714), truck trailers (SIC 3715), and motor
homes (SIC 3716). The electronics sector includes computing equipment (SIC
3573), communication equipment (SIC 3651, 3661, 1662) and electronic
components (SIC 3671, 3674-79). Data from the US Industrial Outlook, 1986.

3 This view that the full impact of electronics technologies stems from the
structural economic transformation they provoke constitutes a central theme
of our work at BRIE. The particulars of the transformation are analyzed in
details in Manufacturing Matters, S. Cohen & J. Zysman, Basic Books,
New York, 1987.

4 This meaning of "competitiveness" is the one adopted by the President's
Commission, as defined in Volume III of the Report of the President's
Commission on Industrial Competitiveness, by Stephen Cohen, David Teece,
Laura Tyson and John Zysman, Washington D.C. 1985.

5 For a detailed analysis of the interrelations between civil and military
technologies, and their policy implications, see : Jay Stowsky, Beating our
Plowshares into Double-Edged Swords: The Impact of Pentagon Policies on the
Commercialization of Advanced Technologies, BRIE Working Paper, April 1986.

6 The arguments developed in this section are taken from Michael Borrus,
Responses to the Japanese Challenge in High Technology: Innovation,
Maturity, and US-Japanese Competition in Electronics, 1983, and Reversing
Attrition: A Strategic Response to the Erosion of US Leadership in
Microelectronics, 1985, BRIE working Papers. Additional sources are cited
specifically throughout the text.

7 combined markets for dRAMs, sRAMS, and ROMs, in Electronics,
"1987 Market Report" 1/8/87 and 1/22/87.

8 Erasable and Programmable Read Only Memories, and Electrically Erasable
and Programmable Read Only Memories.

9 Data compiled by Michael Borrus (Reversing Attrition...) from various
sources: Electronic Industry Association of Japan (EIAJ), Electronics
Components Manufacturers Association, and Dataquest.

10 In "Narrow Windows, Blind Giants, and Angry Orphans: The Dynamics of
Systems Rivalries and Dilemmas of Technology Policy" (paper presented to
the International Conference on the Diffusion of Innovations, Venice,
Italy, March 17-22, 1986), Paul David outlines a framework to analyze these
dynamics. This section draws largely upon his model.

11 Paul David, op. cit. p. 24.

12 Michael Borrus & John Zysman, "The New Media, Telecommunications,
and Development" BRIE working paper, 8/84.

13 For a more detailed analysis of these current developments, see: Borrus
M., Bar F., and Warde I., The Impact of Divestiture and Deregulation:
Infrastructural Change, Manufacturing Transition and Competition in the US
Telecommunications Industry, BRIE Working Paper, 1984, and Borrus M. et al.
Telecommunications Development in Comparative Perspective: The New
Telecommunications in Europe, Japan and the US, BRIE Working Paper, 1985.

14 The original analysis is presented in: Francois Bar, "Business Users and
the Emergence of the New Telecommunications Infrastructure", OECD-BRIE
Telecom User Group Project, unpublished, BRIE, March 1987.

15 This section is excerpted in substantial parts from S. Cohen and J.
Zysman, Manufacturing Matters: the Myth of a Post-Industrial Economy,
Basic Books, New York, 1987.

16 Burton H. Klein, "Dynamic Competition and Productivity Advances", in
Landau and Rosenberg, eds. Positive-Sum Stragegies. Cited in Manufacturig
Matters, p.131.

17 Carol Parsons, The Diffusion of New Manufacturing Technologies in US
Industry, BRIE Working Paper, Berkeley, 1985.
18 Cohen and Zysman, op. cit.

19 Nelson R. R., and Winter S. G., An Evolutionary Theory of Economic
Change, The Belknop Press of Harvard University Press, Cambridge MA, 1982.
20 Dosi G., Technical Change and Industrial Transformation, Mac Millan,
London, 1984.

21 These figures and the following are from: Arcangeli F., Dosi G., and
Moggi M., "Patterns of Diffusion of Electronics Technologies", paper
prepared for the Conference on Programmable Automation and New Work Modes,
GERTTD, Paris, 2-4 April 1987. Data from various sources, including
Electronics, Motorola, CBEMA, OECD (Industrial Structure Statistics), US
Dept of Commerce, Reseau (Milan).

22 Manufacturing Matters, pp. 172-173

23 Manufacturing Matters, pp. 173-174

Back to Top