Innovation Systems and European Integration (ISE) A research project funded by the Targeted Socio-Economic Research (TSER) program of the European Commission (DG XII) under the Fourth Framework Program, European Commission (Contract no. SOE1-CT95-1004, DG XII SOLS), coordinated by Professor Charles Edquist of the Systems of Innovation Research Program (SIRP) at Linköping University (Sweden). Sub-Project 3.5.2: Technological Diversification Vs. New Innovators The Dynamics of High Tech Industry: Swedish Firms Developing Mobile Telecommunication Systems Submitted to the Commission : March, 1998 Maureen McKelvey, François Texier, Håkan Alm Systems of Innovation Research Program (SIRP) Department of Technology and Social Change University of Linköping TEL:+46 28 10 00 FAX:+46 28 44 61 E-mail:[email protected]Abstract: The history of mobile telecommunication in Sweden is to a large extend the history of the firm Ericsson and its relationship to the Swedish PTTs as well as to other small firms. Mobile telecommunication industry in Sweden emerged at the end of the 1970s, beginning of 1980s, under the impulsion of the Nordic PTTs which, through the creation of the NMT standard, opened a field where new technologies and a new market could develop. During the 70s Ericsson was developing, in cooperation with Televerket, through a joint-venture called Ellemtel, a new switch for the fixed network, the AXE system. However in Svenska Radio AB (SRA), a firm jointly owned by Ericsson and the British company Marconi, a small team was developing, in collaboration with Televerket's Radio Labs, technologies for the mobile network. When the NMT standard was finally operational in 1982, it was not clear that the public telecommunication side of Ericsson should provide the technologies for the radio communication side. The vision of the top management of the firm was the "paperless office". Integrating telecommunication and computer technologies would revolutionize office work. Therefore, Ericsson acquired in 1982 small Swedish firms and tried to access the US market. However this vision of the integrated office had come too early in time and Ericsson failed in its venture. Ericsson Information System was sold to the Finnish firm Nokia in 1988. Ericsson Radio System (ERA) was created in 1982. In order to access lacking competencies in radio communication, small specialized firms were purchased. The first, Magnetic, specialized military radio and television equipment, was bought up in 1983. The second firm, Radiosystem, was created in 1978 by three engineers from Magnetic to develop specific components of radio-base stations for the NMT standard. This firm was bought up by Ericsson in 1988. In the mean time analogue standards were made operational, research had started on digital technologies in order to improve the quality of the system. This implied new collaborations with universities. Joint research started between the main technical universities in Sweden and Ericsson as early as 1974. Moreover, Ericsson entered on all three major standards, in the US, Japan and Europe (GSM). From 1990, the policy of Ericsson became firmly oriented towards high investments in research and development and a focus on mobile telecommunication systems. Key-words: diversification, new innovators, mobile telecommunication, Ericsson, NMT, GSM, Radio base stations, Switches.
72
Embed
Linköping University (Sweden). Sub-Project 3.5.2: Technological … · 2015-07-28 · 10.2 Envisioning and Acting upon Opportunities ... develop competencies to move into new technological
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Innovation Systems and European Integration (ISE)
A research project funded by the Targeted Socio-Economic Research (TSER) programof the European Commission (DG XII) under the Fourth Framework Program,European Commission (Contract no. SOE1-CT95-1004, DG XII SOLS), coordinated byProfessor Charles Edquist of the Systems of Innovation Research Program (SIRP) atLinköping University (Sweden).
Sub-Project 3.5.2: Technological Diversification Vs. New Innovators
The Dynamics of High Tech Industry:Swedish Firms Developing Mobile
Telecommunication SystemsSubmitted to the Commission : March, 1998
Maureen McKelvey, François Texier, Håkan Alm
Systems of Innovation Research Program (SIRP)Department of Technology and Social Change
University of LinköpingTEL:+46 28 10 00FAX:+46 28 44 61
Abstract:The history of mobile telecommunication in Sweden is to a large extend the history of the firm Ericsson and itsrelationship to the Swedish PTTs as well as to other small firms. Mobile telecommunication industry in Swedenemerged at the end of the 1970s, beginning of 1980s, under the impulsion of the Nordic PTTs which, through thecreation of the NMT standard, opened a field where new technologies and a new market could develop.During the 70s Ericsson was developing, in cooperation with Televerket, through a joint-venture called Ellemtel, a newswitch for the fixed network, the AXE system. However in Svenska Radio AB (SRA), a firm jointly owned by Ericssonand the British company Marconi, a small team was developing, in collaboration with Televerket's Radio Labs,technologies for the mobile network. When the NMT standard was finally operational in 1982, it was not clear that thepublic telecommunication side of Ericsson should provide the technologies for the radio communication side. The visionof the top management of the firm was the "paperless office". Integrating telecommunication and computer technologieswould revolutionize office work. Therefore, Ericsson acquired in 1982 small Swedish firms and tried to access the USmarket. However this vision of the integrated office had come too early in time and Ericsson failed in its venture.Ericsson Information System was sold to the Finnish firm Nokia in 1988.Ericsson Radio System (ERA) was created in 1982. In order to access lacking competencies in radio communication,small specialized firms were purchased. The first, Magnetic, specialized military radio and television equipment, wasbought up in 1983. The second firm, Radiosystem, was created in 1978 by three engineers from Magnetic to developspecific components of radio-base stations for the NMT standard. This firm was bought up by Ericsson in 1988. In themean time analogue standards were made operational, research had started on digital technologies in order to improvethe quality of the system. This implied new collaborations with universities. Joint research started between the maintechnical universities in Sweden and Ericsson as early as 1974. Moreover, Ericsson entered on all three major standards,in the US, Japan and Europe (GSM). From 1990, the policy of Ericsson became firmly oriented towards highinvestments in research and development and a focus on mobile telecommunication systems.
Key-words: diversification, new innovators, mobile telecommunication, Ericsson, NMT, GSM, Radio base stations,Switches.
1.1 Research Questions and Case Studies .................................................................................................. 52. Mobile Telephony Overview................................................................................................................... 10
2.1 International Market Overview........................................................................................................... 102.2 Dividing the Technologies .................................................................................................................. 122.3 Technical Overview ............................................................................................................................ 13
3. Overview of Relevant Swedish Actors................................................................................................... 184. Switches for Cellular Telephony—Early History in Late 1970s ......................................................... 23
4.1 Televerket, Ellemtel and the AXE-switch ............................................................................................ 234.2 AXE and the Nordic NMT systems...................................................................................................... 254.3 First Operational NMT-system—Not Sweden but Saudi Arabia......................................................... 284.4 Changing Responsibility—SRA and Systems ...................................................................................... 29
5. Parallel Early History of Small Firms in Radio Communication, to 1980......................................... 305.2 Ericsson and SRA................................................................................................................................ 305.2 Magnetic and Radiosystem ................................................................................................................. 315.3 Cooperation and Competition............................................................................................................. 32
6. Ericsson as Mobile TelecommunicationSystem Provider from 1981.................................................. 336.1 Holland 1982, a Turning Point........................................................................................................... 346.2 Implications of, and Strategies for, Becoming a System Provider...................................................... 35
7. Ericsson’s Vision for the 1980s—Information Systems, not Mobile Telephony ................................ 377.1 Risks and Possibilities......................................................................................................................... 377.2 Failure ............................................................................................................................................ 39
8. Early Telecommunication Deregulation to 1985—Impact on Firms.................................................. 408.1 American Market ................................................................................................................................ 418.2 European Markets............................................................................................................................... 428.3 Nordic NMT 900 and Other Firms ..................................................................................................... 43
9. The Shift to Digital Standards—Late 1980s to 1990s........................................................................... 459.1 Sweden and GSM Standard, 1987/88 ................................................................................................. 469.2 Swedish Competence Build-up............................................................................................................ 499.3 Ericsson Gains Competencies: Collaboration and Doing it Alone .................................................... 519.4 Since Then and Forward..................................................................................................................... 53
10. Conclusions ............................................................................................................................................ 5410.1 Diversification vs. New Innovators................................................................................................... 5510.2 Envisioning and Acting upon Opportunities..................................................................................... 5810.3 Implications ...................................................................................................................................... 60
11. References .............................................................................................................................................. 66List of Abreviation ...................................................................................................................................... 71Appendix A: Summary of developers and suppliers for different standards implemented in Sweden. ........ 72
List of Tables
Table 1: Market shares for mobile telecommunication systems by manufacturer. World (W) sales for1987, 1990; Installed lines for 1996.............................................................................................................. 11Table 2: Market shares for cellular phones by manufacturer. World shares (W) for (1984, 87, 91, 91)and for US market (1989, all; 1995 digital; 1996, digital). ........................................................................... 12Table 3: Ericsson concern: % of turnover by business area (1985, 1990, 1992, 1996, 1997) ...................... 19Table 4: Data about Ericsson Information Systems ...................................................................................... 39Table 5: Relations in SI at key innovation opportunities .............................................................................. 59
List of Figures
Figure 1 : Simplified diagram of a small cell concept................................................................................... 15Figure 2 : Simplified structure of radio base station for NMT and GSM systems........................................ 17
3
PrefaceThis research has been financed by the European Union in a Targeted Socio-
Economic Research (TSER) project. This TSER project is ’Innovation Systems and
European Integration (ISE)’, DG XII, European Commission, Contract no. SOE1-
CT95-1004 (DG XII-SOLS).
Within the subproject ’Diversification vs. New Innovators’, our research on
diversification of Ericsson versus new innovators in the Swedish mobile telephone
infrastructure industry will be compared with a parallel case study of Nokia in Finland,
but the most substantial part of the subproject involving patent statistics is carried out
by Bocconi University in collaboration with Fraunhofer-Institut (Malerba et al 1997).
The main research questions of the subproject are whether established firms or new
innovators (new firms or new to a technical area) develop different types of
technologies, and whether the pattern is the same or different in different countries. As
elaborated below, these case studies of Swedish and Finnish firms instead address
specific questions about how and why new innovators vs. diversifying firms can
develop competencies to move into new technological areas.
This project has been carried out at the Department of Technology and Social
Change (Tema T), Linköping University, Linköping, Sweden in the ’Systems of
Innovation Research Program’ (SIRP). We wish to thank members of SIRP, of the ISE
project, and participants in the January 1997 workshop ’Mobile Phones : Success for
Scandinavian Firms?’ for some early comments (see McKelvey 1997b). In addition, we
extend our thanks to all the persons who have given us their time in interviews.
4
1. Introduction
This report addresses a fundamental research question, namely, the dynamics of
high tech industry, specifically why, and how, firms can identify and act upon the
innovation opportunities inherent in a new technology with market potential.1 The
purposes of this report are to use a dynamic view of systems of innovation in order to
explore the discontinuities, challenges, and innovation opportunities introduced by new
technologies and new high tech products. We do so by examining how Swedish firms
have developed and sold equipment for mobile telecommunication systems.
Two more specific research questions about the dynamics of high tech industry are
addressed. The first is in relation to what types of firms innovate. In general, the
discussion about industrial dynamics caused by the emergence of radical new technical
change is linked to a discussion about whether large or small firms innovate under
different circumstances. Are small firms those which introduce innovations due to their
greater flexibility and growth potentials? Or do large firms have more appropriate
resources and competencies to innovate and to bring initial ideas to market? Many
argue it is the new firms which can introduce a new technology, particularly if it is
based on a different type of technological competence (Tushman and Andersson 1986).
A basic dichotomy is thus whether dynamics in the economy due to technical change
can be explained by the emergence of new firms or by the diversification of large,
existing firms.
The second question is whether firms innovate in isolation or whether, and when, a
firm’s innovative activities are made possible by relationships with other actors.2 The
dynamics of interaction between a firm which is innovating and other actors like firms
and government agencies and institutions relevant for innovation—or a system of
innovation (SI)—are thus also of interest.
These research questions are interesting not only because we know so little about
the dynamics of innovation processes in terms of the turbulence of firm entry and exit
from an industry but also because we know so little about cases where existing firms
and existing industries manage to bridge discontinuities introduced by new
technologies. Sometimes, existing large firms can gain radically new competencies
which challenge their traditional way of working, as in the case of the introduction of
biotechnology into the pharmaceutical industry, where small biotech firms live in
symbiosis with large pharmaceutical firms (McKelvey 1996b). Two ways to examine
1Following the OECD (1993) definitions, the concept ’high tech’ is operationalized as research and
development (R&D) intensive industry. The division of high - medium - and low tech industries is basedon an average for industries for all OECD, and high tech means a high percentage of sales is spent onR&D in manufacturing. For services, the definition is more complicated due to lack of formal R&Dspending, but can be equated with employing skilled persons.
2See also Ehrnberg and Jacobsson (1997).
5
these questions spring to mind. One is to count entry and exit of all firms in an industry
over time and relate that to the emergence of new technological areas. We take a second
approach, which is to present such material, but mainly focus on how and why different
types of firms moving into a new technological area can identify and act upon
innovation opportunities. The next section specifies the research questions and case
studies of Swedish firms making mobile telecommunication systems.
Sections 2 and 3 give relevant information for the case studies. The first gives an
overview of market and technical aspects and the second, an introduction to the main
Swedish firms. Sections 4 to 9 examine the turbulence and inertia in the Swedish
mobile telephone infrastructure industry from its beginning in the mid-1970s to the
present and constitute the main body of the text. The main focus is on transitions to new
technologies and new markets. The concluding section 10 addresses two things. Firstly,
it discusses whether these cases of technological change are driven by diversifying
firms or by new innovators and also the dynamic relationships between an innovating
firm and others in a SI. Secondly, it addresses what the history of Swedish mobile
telecommunication can tell us about the dynamics and inertia of industry, in order to say
something of relevance for firm about innovation processes as well as for government
policy-making for economic growth and employment.
1.1 Research Questions and Case Studies
Our specific approach and research questions have been developed by confronting
theoretical literature on industrial dynamics, specifically diversification versus entrants,
and on systems of innovation. Here, we are only concerned with this question in cases
of radical technological change, which involve a new and changing set of technology
and knowledge.
First of all, there is the definitional question about which firms can be defined as
diversifying, as new innovators, and/or as new firms. Much of the research in industrial
organization defines diversification as a move to a new industrial sector, thus basing the
distinctions on broad product groups. Our definition of diversification differs because
we start from the idea that learning and innovation can be difficult for firms, especially
when new technologies are involved.3
Diversification is therefore here defined as the process by which a firm enters a new
market, with a new product line for the firm. As we are interested in radical change, we
are only interested in cases where new technology is involved. Thus, while development
of existing product and market rely on the development of existing competencies within
the firm, diversification necessitates new competencies, both technological and
3This assumption is based on an evolutionary view of economic and technical change, which is used
not only in economics in, for example, Dosi et al (eds.) (1988) but also in business studies with thecompetence-based theory of the firm. See Teece et al (1994) and Nelson (1991).
6
economic. This does not mean that the firm has to leave its existing
market/product/technology for the new one, but more that the firm broadens its scope of
activities. As the process of diversification implies changes at technological and
commercial levels, new facilities and organizational changes may also be necessary.
Following inspiration from Malerba et al (1997) work on analyzing technological
diversification, we therefore define diversification as moving into a new technological
field and involving a market new to the firm.
In defining diversification and new innovators, we distinguish firm diversification
from other types of firm growth. Other growth can come through market development
or through development of existing products. For example, a firm can enter a new
market with an existing product, or move from its present product line to a nearby one.
Ansoff (1957:114) argues that ’diversification calls for a simultaneous departure from
the present product line and the present market structure’.
The other side of the dichotomy of diversification is new innovators and new firms.
New innovators are new firms, first of their kind, since by nature these firms are started
to take advantage of the emergence of a new high tech product. For example, these
firms might be start-ups from university research or spin-offs from existing firms. More
importantly, these firms may be characterized by their fast growth potential as well as
their likelihood to take risks and innovate. Although there may be reasons for optimism
about the ability of small firms to introduce dynamics as in the American case, recent
research on small Swedish technology-based firms has shown their minor importance
for a number of variables like production, employment, etc. (Jacobsson and Rikne
1996). If these variables are taken as indicators of turbulence, then Swedish small firms,
or new entrants, are not very important in the Swedish economy.
A related question is how firms diversify, i.e. what strategy they use to move into a
new area. Malerba et al (1997) reduce the diversification strategy to the acquisition of
firms. Although purchasing existing firms is one means, there are other strategies to
diversify which either involve relations to others to obtain information or development
of in-house competencies. Two ways of obtaining information from the outside are
outsourcing of R&D projects and licensing existing products/knowledge from others.
In-house R&D is an internal source of knowledge, but existing R&D often needs to be
redirected, and often re-organized, in order to move into new areas of knowledge as the
pharmaceutical industry shows (Hendersson 1994). Building up significant in-house
competencies in new areas implies a commitment of the firm management, suggesting
that new high tech products can have difficulties in being accepted.
The basic problem is that at the firm level, arguments can be made both for waiting
to see what happens and for investing early into uncertain, new technologies. Waiting
means that an existing firm can see whether the new technology has the promised
growth potential but moving in early means the possibility of being a technological
7
leader and capturing early, and hopefully durable, market share (McKelvey 1996a, ch
9). Both strategies may seem equally rational to different persons in one firm or to
different firms at one point in time. This implies that assumptions should be made not
only of uncertainty but also that decision-making takes places within limited bounds
(Nelson 1991).
Our research questions must therefore move us beyond the firm per se to see the
firm in relation to institutions, firms, other organizations and individuals which make
innovations possible, that is, the system of innovation. Systems of innovation is an
approach which Edquist (1997:15) argues has a number of common features such as
inclusion of institutions, holistic perspective, emphasis on dynamics and the importance
of innovation to the economy.
Systems of innovation are here defined in terms of how and why relations with
others help, or hinder, firms to identify and act upon innovation opportunities
(McKelvey 1997a). Relations may, for example, lead to the transfer of technical
information relevant for a new technological area or specify a market demand of which
the firm otherwise might not be aware. Firms are active in their own right in such a
definition, but the system is important to the extent that relationships and institutions
influence the firm’s perceptions of opportunities, either positively or negatively. This
definition is based on evolutionary economics. It is thus explicitly dynamic in that it
focuses on innovation processes and on the changing importance of the system. SI so
defined focuses on knowledge production, exchange and selection.
A few words are therefore appropriate about the challenges and implications of
looking at dynamic SIs—rather than describing existing ones—in relation to our
research questions. Firstly, a dynamic perspective has two dimensions, first, in the sense
of looking at qualitative change over time and secondly in the sense of examining when
a SI is important to the firm, and when it is not. Over time within a technological area,
there are probably qualitative changes in how innovation processes are organized as
well as in the rate and direction of technical change. Innovation processes themselves
are filled with uncertainties in the sense that they are based on novel knowledge and
technologies, which may or may not work as envisioned, either technically or
commercially. By extension, then, it is reasonable to assume that under some
circumstances, the firm will rely on a SI to innovate, but under others, it innovates alone
as much as possible.
Secondly, dynamics must be explicitly taken into account when the focus in on
innovation processes over time. Bringing in time automatically introduces dynamics in
that the composition, strength and importance of a SI can probably vary over time, due
to actors identifying and acting upon new tensions and new opportunities. Thus, actors
are assumed in our perspective to be changing the conditions of competition, rather than
accepting given and externally determined factors. Qualitative changes over time are
8
not only caused by changing in technical change but also by actors’ own actions. For
example, the relative influence of SI elements and of firm own action is assumed to
vary over time, and so we want to know when and under what circumstances they vary.
That is why we try to identify when and why four elements of a SI—university,
government, competing firms, small firms—have each been crucial for technical and
economic innovation, and when not as well as when and why internal firm factors have
been more important.
This report addresses these larger research questions in relation to one specific case,
namely by analyzing the development of the mobile telephone infrastructure industry in
Sweden.4 We define the mobile telephone infrastructure industry as including products
for the network (radio base stations and switches) as well as the cellular phones
themselves. The current study is thus limited to goods products.5 Mobile telephony is
that part of the telecommunications industry which has globally expanded at an
extremely high rate, particularly during the 1990s. For example, the OECD market for
mobile communication services doubled between 1993 and 1995 (to USD $65 billion)
as have the number of subscribers (users) in most OECD countries (OECD 1997:37-41,
49-51).
Since mobile telecommunication is part of the telecommunication industry,
industrial organization researchers could argue that existing firms which move from
fixed to mobile equipment cannot be seen as diversifying as they are in the same
product group. However, as argued more extensively in section 2 and in McKelvey and
Texier (forthcoming), although some of the basic knowledge and technology are the
same, there are important differences. Much of mobile telecommunication relies on
quite different, and often quite rapidly changing, bodies of knowledge and technology
and markets differ in structure and type of characteristics demanded. Due to some
significant differences, moving to mobile telephone technology may therefore be a case
of technological diversification, and so section 10 will consider whether diversification
occurred, and whether diversifying firms or new entrants innovated, in relation to the
three component technologies—switches, radio base stations and cellular phones. While
the infrastructure elements of switches and radio base stations are necessary elements of
4We use the term "mobile phones" to denote what is also called cellular phones, digital phones, car
phones, cellphone, etc. Note that this does not include wireless phones which you can use around yourhouse but which are part of the traditional system. What used to be called ’phones’ are part of what hasvariously been called the public, fixed, or wired phone network.
5One reason is that so far at least, the hardware is most significant for exports, and exports are oftentaken as a measure of the specialization of a country, relative to other industrial countries. Especially inthe so-called Basalla index. However, in a much larger study than the current one, it would be extremelyinteresting to analyze the dynamic relationships between services and goods in mobile phones. After all,Edquist, Hommen and McKelvey (1997) postulate that R&D intensive areas with a high interactionbetween services and goods provision are those with the greatest potential for employment growth. Thus,one could study the dynamics of the relationship between operators or providers of the service and thoseproviding the hardware.
9
mobile telephony, they are not as visible to the casual eye as the phones, but they are
very significant products for some firm and for economic change more generally.
On the one hand, this Swedish case could be read as a history of the large
telecommunication firm Ericsson. Ericsson has spent one of the highest percentage of
sales on R&D of Swedish firms during the 1990s (Ny Teknik 1997b, 19961; Wolmesjö
1992:16), mainly because they have invested heavily in R&D for mobile
telecommunication since the early 1990s. Ericsson’s mobile telephony business is
particularly noteworthy as a successful radical innovation, especially when you
contemplate the fact that Ericsson’s export in 1996 was about 10 billion crowns, or as
much as a traditional Swedish industrial specialization, the paper and pulp industry,
exports (Ny Teknik 1996b). Or that mobile communication has moved from being 2% of
Ericsson’s sales in 1975 to 70% in 1996 (McKelvey 1997b:2). The large firm Ericsson
is a Swedish international success story, not only for mobile phones but for renewal in
the Swedish economy (see Edquist and McKelvey forthcoming; and McKelvey and
Edquist 1997). Ericsson is a particularly interesting case because it, together with the
pharmaceutical firms Astra and Upjohn & Pharmacia, are the exceptions to the Swedish
decline in high tech products. These are profitable firms which have expanded in recent
years.
Yet on the other hand, the Swedish story also involves small firms, as well as
Ericsson’s relationships with other actors and with institutions relevant for innovations.
As our focus is on which firms can innovate, how and why, the small firms included are
those (small) Swedish firms which are competitors in one of the major product
groups—switches, radio base stations, and telephones. This point can be made stronger
by stating what our case is not. The case is not of all the small firms which are
suppliers; nor the filère, development block, or larger network or relations relevant to
production;6 nor of the total industrial sector; nor of Ericsson alone. As our purpose is
to understand the dynamics of high tech from a system of innovation perspective, only
those organizations and institutions relevant for innovation processes for mobile
telecommunication are included. As Ericsson's top management initially considered it
as a minor niche, we also discuss their vision of the future in the early 1980s, namely
information systems for the paperless office - which was a spectacular failure.
Thus, this report analyzes Ericsson’s ability to move into mobile telecommunication
in relation to small firms; to being able to access competencies externally and develop
them internally; to exceptional individual entrepreneurs within Ericsson and within
small firms; to the Swedish PTT; to international buyers; etc. So the Swedish case
history tells of dynamic changes in a system of innovation and the points at which
6For the concepts, see, respectively, (Perroux 1969), (Dahmen 1988), (Håkansson 1987).
10
relationships in the SI are important to the firm to innovate. To analyze relationships in
SI, the following five sets of questions have been specified:
1) Where did ideas and visions come from? How were they implemented in the
firms?
2) How were technical and market competencies acquired or developed by firms?
What was the role and importance of in-house R&D? External sources?
3) How important was national or international university research?
4) What was the role of government policy or government agencies?
5) What were the relationships between large and small firms? Why?
These have been chosen both because they enable us to focus on the location where
knowledge important to the innovations has been generated as well as on the relative
importance of external actors to the firms at specific points. The choices have been
informed by our theoretical framework.
2. Mobile Telecommunication Overview
As mentioned in the introduction, the mobile telephone infrastructure industry is
here defined as including the infrastructure system, which can be divided into radio base
stations and switches, and cellular phones. Mobile phone networks can be contrasted
with the fixed, or public, network. On the one hand, mobile telephones fulfill the same
functions as ’normal’ phones literally connected to a public telecommunication network
while on the other hand, they allow a new freedom of movement and work on
somewhat different technical principles.
Mobile telephony is quite interesting because no one expected it to become such a
growth sector. Moreover, mobile and radio communications were mainly developed for
military purposes, especially communication over a long distance on land, by air or by
sea. Cellular telephony was initially seen as a niche civilian market for specialized
groups like doctors on call or fireman. There has been explosion of use and of diffusion
to new user groups beyond all predictions, and today, mobile telephony is not only
widely spread as an alternative way of communicating but also a direct replacement for
some fixed network calls. This section therefore gives some basic information about the
international market development and about technical change in order to provide a
framework for understanding later empirical sections and for the conclusions.
2.1 International Market Overview
The market for mobile telecommunications has expanded dramatically, particularly
during the 1990s. At the beginning of 1996, 82 million people were connected to
mobile telephone networks around the world, whereas by the end of that same year, 137
million were connected (LME 1997a). Access to digital standards were responsible for
65% of the growth in 1996, with the three major digital standards—GSM, D-AMPS and
11
PDC—responsible for 98% of the net increase of digital subscribers. Only three years
earlier in 1993, there were only 16 million worldwide (Hultén and Mölleryd 1993:2). In
1985, mobile phones were installed in about a half million cars, and Ericsson projected
a growth to about 9 million (in cars) by 1989 (Affärsvärlden 1985a:18, 23).
Correspondingly over time, the price of the phones has dropped dramatically. The
price of the phone has dropped from around USD $4 000 to USD $100 over this period,
and in some markets or under a temporary marketing campaign, the operators subsidize
phone purchase in order to attract new subscribers, so the price of the phone can be
close to zero. The cost of making calls differs depending on the market, but generally,
there are more differentiated tariffs now so that that market can be segmented between
business and personal calls (OECD 1997).
As to the firms selling mobile telecommunication systems as well as selling the
phones, it is mainly large, multinational firms selling products, and they have been
involved since early on. The information given below are approximation based on
different sources, as market estimates are hotly contested between competing firms. It
gives, however, a fairly good idea of market shares as the disputes are often over a few
percentages. For major competing firms, Table 1 shows the international market share
of network systems for major competitors for 1987, 1990 and 1996 while Table 2 shows
the international market share of phones the years 1984, 1987, and 1991 and the share
of the American market for 1989, 1995 and 1996.
Table 1: Market shares for mobile telecommunication systems by manufacturer. World
(W) sales for 1987, 1990; Installed lines for 1996.
1987 Wa 1990 Wb 1996 Wc
Ericsson 45% 40 40
Lucent (AT&T) —* (close to Motorola) 16
Motorola — 30 15
NEC — — 12
Nokia — — 5
Nortel
(Northern Telecom)
— — 7
Siemens — — 5
Sources:a (Affärsvärlden 1987) b (International Management 1994:27; Ekwall
1991:38) c Calculations based on (LME 1996b). *’—’ means not listed in that source.
12
Table 2: Market shares for cellular phones by manufacturer. World shares (W) for
(1984, 87, 91, 91) and for US market (1989, all; 1995 digital; 1996, digital).
'84 Wa '87 Wa '91 Wa '91 Wb '96 Wc '89 US,
alld'95 US,
digitale'96 US,
digitale
Ericsson 9-10% 4 —* — 20-25 4 34 55.7
Ericsson/
GE**
— — 7 4 — — — —
Hitachi 6-8 7 5 — — — — —
Mobira/
Nokia
13-15 14 6 16 — 10-15 28.8 33
Mitsubishi — — — 8 — — — —
Motorola 10-12 13 30 25 20-25 10-15 36.3 8
NEC 10-12 11 12 8 20-25 10-15 — —
Panasonic 7-9 8 7 8 — — — —
Toshiba 5-7 10 7 4 — — — —
All others 27-40 33 26 — — — 0.8 3.3
Sources:a (Oskarsson and Sjöberg 1994: Table 1) b (Andersson 1992:23) c (Froste
1997) d (Nylander 1990:30) e (Dataquest 1997)
*’—’ means not listed in that source. **Ericsson and GE had a joint company for
selling cellular phones between 1989 and 1994.
Tables 1 and 2 show that Ericsson has been a major competitor for mobile
telecommunication systems since the beginning and that from the mid-1990s, it has also
become one for cellular phones. The market for systems has been more dominated by a
few firms whereas the market for phones has seen a growing concentration over time.
Initially, the phone market was quite fragmented among different producers.
2.2 Dividing the Technologies
Our article separately analyzes the three different key components of the mobile
telephone infrastructure industry—switches, radio base stations (RBS), and telephones.
It is very important for our research questions to differentiate these three components as
well as industrial products from the provision of services, even though most of the
previous Swedish studies lump them all together (Mölleryd 1996, Lindmark 1995). We
do so for the following five reasons:
1. Each requires its own core set of knowledge and technology which differs from
the other two, even though some of these core sets are shared across them. For example,
13
both the phones and the RBS require knowledge about radio communication and
signals.
2. The markets differ between RBS and switches on the one hand and telephones on
the other. RBS and switches are sold to network operators whereas in each geographical
market, the phones have first been a status symbol for certain user groups but then
become more and more a mass consumer good. Although the users for
telecommunication network equipment have traditionally been state-owned PTTs, the
markets for RBS and mobile switches opened up during deregulation, meaning that
PTTs as well as independent and competing operators have been the new customers.
3. The bottlenecks, challenges, and advances of innovation processes for the three
components have differed somewhat. The reasons why a firm had competence in an
area which turned out to be important for a component have differed. For example,
previous military orders for mobile communication have been crucial for RBS whereas
switches for public telecommunications have been important for switches for cellular
networks.
4. Internationally, even the large firms generally specialize in different parts of the
mobile telephone infrastructure industry. For example, although Ericsson has been
successful with its phones in recent years, its strength has lain in providing the mobile
telecommunication system, where Ericsson equipment represents about 40% of world
sales as well as of existing network equipment installed (Ny Teknik 1997a).
Internationally, some firms only sell some of the components of the infrastructure
whereas others try to sell the whole concept.
5. Because of the reasons listed as 1 to 4 above, there have been pressures on the
firm to change organizationally. For example, Ericsson has had to reorganize in order to
position the mobile phone area against their traditional telecommunication products as
well as to find ways to integrate both R&D and sales of these separate components to
become a system provider.
These five reasons demonstrate the importance of dividing the mobile
telecommunication industry into three components—as well as how they are
integrated—in order to understand how firms identified and acted upon innovation
opportunities as well as the relative importance of a SI at different points.
2.3 Technical Overview
A brief comparison between fixed phones / telecommunication networks and
mobile ones will lead us into a discussion of some relevant technical details. When
calling in either network, the switch makes the connection between the line of the
person calling and the line of the person called. Switches thus transfer calls between
different users, and both types of networks rely on switches. From having been large
electromechanical devices which initially required an operator and then became
14
automatic, switches are now based on electronics and computer programs. This has also
enabled them to provide different types of services to users like call waiting.
The major technical differences between the networks arises from the obvious
difference, namely wires versus radio signals. In the fixed network, signals between
those communicating travel through wires, where the signals are transformed in
switches, and then further transmitted. In mobile networks, the phone of the person
calling sends a radio signal to the nearest RBS. The RBS converts the signal into an
electric signal and sends it to a switch in the public telephone network which can handle
these signals via wires.
Because mobile telephony relies on radio communication rather than wires,
different types of knowledge and technologies are required to allow RBS and phones to
communicate. Electronics and signal processing are crucial. The voice or data is
converted into an electric signal, which then is converted into a radio signal and
transmitted. The reverse process occurs when the radio signal arrives to the receiver:
radio, electric signal, sound. Radio communication technology therefore involves a
number of sub-technologies like transmitters, receivers, amplifiers, filters.
Another difference is that the fixed phones are basically ’dumb’ and only for calls
whereas mobile phones are based on microelectronics, which allow multiple functions
like all the features (that few users know how to program or use). The mobile telephone
also includes advanced radio communication technology in order to communicate with
RBSs.
Radio communication between RBSs and phones places some demands both on the
organization of the (physical) network and on the necessary functions. Two major
problems with radio signals are that they tend to lose energy and hence their intensity
while traveling in air, and that there are a limited number of frequencies available. The
small cell concept for the infrastructure was developed to manage these problems.
Instead of having one radio base station sending a strong signal over a large area,
several smaller stations are used. Each station is situated at the center of a cell, which
like bee cells, together cover a large area.
15
Figure 1: Simplified diagram of a small cell concept
With this design, the radio signals do not need to be so strong, and the same
frequencies can be used several times in different areas. Designing infrastructures to
meet the needs of network operators, and hence subscribers, has continued to be
challenging for manufacturing firms (Jismalm and Lejdal 1990). This physical
infrastructure also has to support certain functions in order to be useful on a larger
geographical scale, and two important functions are roaming and handover. Roaming
means locating the mobile phone of a person called, and handover means the transfer of
a radio connection from one cell to the next. These two functions are taken care of by
Public
telecommunication
Radio Base Station
AXE
AXE
16
the radio base station controller (which is a computer-based switches in current Ericsson
equipment).
Our overview would not be complete without mentioning standards. If we start with
the Nordic and broader European case, the first (civilian) standard for modern cellular
telephony began being specified in 1970, and was called NMT 450, that is Nordic
Mobile Telephony standard based on the 450 megahertz bandwidth.7 Two demands
which were novel were that it be fully automatic and have roaming within the Nordic
countries. The standard was initiated by the Nordic PTTs, and a working group of PTTs
in Finland, Norway, Denmark and Sweden, designed the technical specifications in
discussions between 1975 and 1978, although the Swedish PTT took a leading role
(Mäkitalo 1997, Kågström 1997).
In 1978 implementation of the project was launched, and the Nordic PTTs started to
look for suppliers of the different component technologies, i.e., radio base station and
switches. This NMT standard set very specific technical specifications, which meant
that a network operator had the possibility of buying components from different
producers and putting them together themselves. NMT 450 was implemented in
October 1980 in Sweden, and at the beginning of 1981 in Denmark, Finland and
Norway. However, the very first implementation occurred in Saudi Arabia in August
1980, which also turned out to be an important order for Ericsson (see section 4.3).
The NMT 450 was more successful than ever expected. It was initially forecasted to
have 50 000 subscribers by 1990, whereas by 1992, it had approximately 250 000 and
its replacement system had more than 350 000 subscribers (Hultén and Mölleryd
1993:4, 7). Because more subscribers were joining than the standard could handle, the
Nordic PTTs then went to the NMT 900 (megahertz) standard in 1986. The NMT 900
system was developed as an intermediary system, between the NMT 450 and the future
European digital standard, which was later agreed to be GSM (Mölleryd 1996,
Lindmark 1995, Meurling and Jeans 1994). Both NMT standards handle signals in an
analogue way.
The first Swedish discussions to introduce a digital standard started in 1982, again
under the lead of Televerket, the Swedish PTT. In 1992, the GSM, or Global System for
Mobile Communication,8 standard was introduced in the Nordic countries and in
Europe. One advantage of GSM is that a much larger number of subscribers (telephone
users/calls) can be handled, but implementing the system required a new infrastructure
to be built up. GSM is an open standard where producing firms can configure how
7See Meurling (1994:7-14) for further technical details. See also Lindmark (1995) for information
about standards. Previous to NMT 450, The Swedish PTTs had developed several mobile telephonesystems, MTA, MTB, MTC and MTD. Those were either small scale systems (MTA, MTB and MTD) orresearch project (MTC). However, experience with them was one important to several companies(Mölleryd, 1996, Lindmark 1995).
8Previously Groupe Spécial Mobile
17
communications between the components should be formed, and so responsibility for
configuring the system shifts from operator to a producing firm. The consequence is
that the firm either has to be a system provider or have alliances with others. Most of
Europe now has the GSM standard, although some countries still have the NMT 900
standard as well. The GSM standard is also used in other countries like Australia, New
Zealand, and China. In fact, GSM was used in at least 105 countries by 1996 (America’s
Network 1996).
The transition from NMT to GSM involved some dramatic changes in the
knowledge and technology used to process and control the radio signals, since the GSM
standard is based on digital signals requiring more data processing.
Figure 2 : Simplified structure of radio base station for NMT and GSM systems
Source: Kägström 1997
Figure 2 compares a RBS for NMT and for GSM. In an NMT system, the core
knowledge in the RBS has to do with the radio equipment. The control unit used is
relatively simple and does not require large amounts of computer programming. In a
GSM system, the control unit is proportionally much more important, and therefore new
competencies were needed to develop a functioning system. The radio equipment was
comparatively much similar to the one of a GSM.
The USA and Japan have each gone their own way. The USA first had the AMPS
(Advanced Mobile Phone System) and then a digitized version of AMPS (D-AMPS).
Japan has had the PDC (Personal Digital Cellular) standard but while it has been late in
coming into the mobile phone networks, Japan is currently trying to lead-frog today’s
standards in order to have one more appropriate for mobile computer communications
more generally. (America’s Network 1996).
RADIOCONTROLUNIT
CONTROL UNIT
NMT GSM
RADIO
18
3. Overview of Relevant Swedish Actors
This section gives a brief overview of the Swedish actors discussed in the following
sections, specifically of the large firm Ericsson, small competing firms, and the Swedish
PTT Televerket. It presents information which is relevant in relation to our research
questions about the size of firms, where they came from and went to, and how they
acquired technical competencies.
Ericsson
The firm LM Ericsson is the dominant firm in Swedish telecommunication industry,
including cellular telephony. It has always been involved in telecommunications. It was
founded in 1876 by Lars Magnus Ericsson. Using technologies developed by Bell, but
not patented in Sweden, Ericsson developed Sweden's first telephone in 1878. Rapidly,
the firm expanded and produced both telephones, electromechanical switches as well as
military radio equipment.
Ericsson consisted of two major divisions by the 1970s. The main division was the
switching division, or internally called BX (Business Exchange). BX produced switches
for public telephone networks, which was the largest product area, and its customer
were mainly public telecommunication operators (PTTs). Ericsson mainly sold to third
world countries and could be considered a second rank company in the international
competition (Ny Teknik 1997a). The second division was much smaller and called
business radio communication, or BR. It consisted in production of military radio
equipment and consumer electronics (radio and television). Much of the important
activities of BR relevant for mobile telephony were carried out in the separate, jointly
owned firm SRA. At the beginning of the 1980s, a third business area was the S
division (or System Division) working with computer systems (see section 7).
The mobile phone business has gone from being 2% of sales of Ericsson sales in
1975 (mainly to military communication) to 70% in 1996 (McKelvey 1997b:2). The
overwhelming importance of mobile telephony to Ericsson was reflected in a new
organizational structure as of January 1997. Today the structure of Ericsson has three
main business areas: 1) Mobile Phone and Terminals, for consumer products, 2)
Infocom Systems, providing operators with networks and services, and 3) Mobile
Systems, producing mobile communication systems. Not only has mobile telephony
increased its share, the concern as a whole has expanded rapidly, especially during the
1990s. Table 3 shows the dramatic shifts between business areas as well as the overall
growth in sales (current prices).
19
Table 3: Ericsson concern: % of turnover by business area (1985, 1990, 1992, 1996,
1997)
1985a 1990b 1992c 1996d 1996d (after
restructuration)
PublicTelecommunication
33,8% 45% 43% 18,6%
Ericsson InformationSystems
31,8% — —
Business Networksand Components
19.8% 14% 14% 15,4%
Microwave Systems 6,7% 4% 4% 2,5%
RadioCommunication
8% 25% 27% 63,2%
Mobile Systems 43%
Infocom systems 30%
Mobile telephonesand terminals
17%
Other 2% 10%
Total Turnover,current prices (USD)
$ 5,4 bill $ 7,5 bill $ 7.5 bill $ 20,7 bill $ 20,7 bill
1992:30). d(Ericsson Annual Report 1996). e(LME1997). (Conversion from SEK to
USD has been based on 6 SEK = 1 USD)
From the mid-1980s through the early 1990s, the growth of mobile telephony
internationally and as Ericsson products was to surpass all expectations. It moved from
being a small percentage of the concern to being the dominant one, and along with its
expansion, came an explosive expansion in the overall size of the firm. In connection
with the growth in total sales and turnover, Ericsson has increased the number of
employees internationally from about 70 000 in 1985 to about 90 000 at the end of 1996
(Månsson 1985:24, America’s Network 1996).9 The international aspects of business
have also decreased the importance of Sweden as a market. In 1990, sales in Sweden
represented 12% of total sales in 1990 whereas in 1996, Swedish sales represented only
3% of the total but 60% of R&D were still in Sweden (LME 1990:4; Ny Teknik
1997a:31). Thus, Ericsson has not only been expanding rapidly in mobile
telecommunications and totally, it has also becoming increasing international.
9In mid-1997, Ericsson employed 44,000 persons in Sweden, an increase of 14,000 employees in
Sweden between 1992 and 1996 (Ny Teknik 1997a, 25).
20
Relevant Ericsson collaboration in Sweden includes two cases of jointly owned
firms, SRA and Ellemtel. SRA was a jointly owned company with Marconi for
production (and development) whereas Ellemtel was purely a jointly owned R&D
company with the Swedish PTT Televerket. In addition, Ericsson has had much
international cooperation, but those cooperation have been with other large companies
such as Cisco for internet standards or Matra to sell to the French market. The
collaboration in Sweden has been with small companies, which often end up being
bought up, or else with Televerket.
SRA
Swedish Radio AB (SRA) was created in 1919 by the three Swedish firms, ASEA,
AGA and LM Ericsson to produce radio equipment for the Swedish Navy and the
Swedish PTT. In 1921, Marconi Wireless Telegraph Company ltd., a British firm, took
a 43% share, while Ericsson increased its share to 57% by buying the shares of ASEA
and AGA in 1927. In 1962, Ericsson then bought some of the Marconi shares,
increasing its ownership in SRA to 71%. Finally in January 1983, Ericsson became the
sole owner of SRA and changed its name to Ericsson Radio Systems (ERA) (Bevenius
1995:1). The internal name for this business line was BR (Business Radio).
The main competencies of SRA were traditionally in military, land radio
communication. Much of the radio communication knowledge initially came mainly
from Marconi, and in fact, letting Marconi buy share in SRA also gave them access to
Marconi's patents and thereby facilitated development of the company (Mölleryd
1996:81, Lundquist 1997). However, SRA also produced radio equipment for the
consumer market, under the brand name Radiola. This part of the firm was sold in 1964
to AGA (Bevenius 1995:3). Over time, however, Marconi’s interest and role changed
dramatically, and during the 1960s, Marconi stopped its activities in mobile
telecommunication. Marconi led some projects at SRA, but in total, few projects were
done in cooperation with them (Lundquist 1997).
SRA mainly developed land mobile radio equipment for the Swedish army and for
other specific usage, such as naval communications or police equipment. Although a
majority of work was oriented towards the military market, part of SRA was involved in
the development of mobile radio for private use in cooperation with Televerket
(Mölleryd 1996:34). The top management of SRA had an explicit policy to shift their
business from military to civilian areas by the mid-1960s (Lundquist 1997). Ivar
Ahlgren, CEO of SRA, realized that changes in the international situation would lead to
cuts in military expenditures, and that therefore SRA had to change its orientation. It is
out of this shift in policy that Ericsson’s nascent mobile telephony business was born,
and the technology used in the radio base stations was, initially at least, directly based
on previous military technologies (Lundqvist 1997). See section 5.
Ellemtel
21
Ellemtel was started as a joint R&D company between Ericsson and Televerket. In
1968/69 the two actors began to discuss closer form of co-operation than their previous
collaboration in order to develop computer controlled switches for the public
telecommunication network (Vedin 1992:98; Vedin 1994). In May 1970, Ellemtel was
formed as a jointly owned R&D company, with approximately 400 employees to
develop a new switch, called AX (later known as AXE). The initial specifications for it
were ready in 1971, and by 1972 the AX system was clearly defined and ready to be
developed. Televerket and Ellemtel wanted to start the R&D project as soon as possible,
but Ericsson was having second thoughts. Ericsson took the decision in February 1972
that they would participate in development of the AX switching system and that
Ellemtel would do the majority of the R&D work in cooperation with the two owners.
See section 4.
Televerket
As the joint collaboration with Ericsson over computer controlled switches indicates
as well as its key role in developing the NMT standard, Televerket had its own,
significant technical competencies. It used these not only in order to use R&D to
monitor trends but also to develop and produce its own products, particularly switches
for urban areas. Televerket has had its own R&D labs, Radiolaboratorium, through
much of its history. Deregulation in the 1980s and 1990s has meant a gradual reduction
of their technical strength, to instead concentrate on being a network operator. Due to
the strong technical position of Televerket, Ericsson has had a discussion partner for
technical matters through much of its history.
Note that Televerket manufactured its own AXE-switches for the fixed Swedish
network up until 1994, so Ericsson never had a captive home market. Televerket
manufactured the switches in a company called Teli AB, and they did so until Ericsson
bought the company in January 1994. At that time Teli had about 1300 employees and
the acquisition ended Televerket’s era as a manufacturer of telecommunication
equipment.10 See section 4.
Magnetic
The firm Magnetic was started in 1952. Its activities were mainly in military
technologies for radio communication; in measuring systems for the navy and airforce;
as well as in radio equipment for the airforce (Kågström 1997). In the 1960s, Magnetic
had one production unit in Bromma and several development units, where one of these
units was developing transmitters for television. In 1968, Televerket needed television
transmitters for the new channel TV2. The frequency of emission had been decided to
10Why did Televerket stop making switches for the domestic market? According to Östen Mäkitalo
(1997), Televerket had to concentrate their efforts to their core business, which is to function as anetwork-operator. The manufacturing was considered as a relatively small-scale business, and thereforeunnecessary to continue.
22
be in the range of 400 to 800 MHz (the so-called UHF band), and the contract for
development and production went to Magnetic.
Torbjörn Johnson, an engineer there, took the responsibility for the development
along with Leif Kågström and Tommy Moberg. It was a small project, in that a total of
10 persons were involved in development and manufacturing of the transmitters for
TV2. Unexpectedly and unintendedly, it turned out that development of these TV
transmitters would turn out to be an important technical competence for RBS for mobile
telephony (see further section 6).
In 1975, as did other producers of radio equipment, Magnetic received an offer from
Televerket to produce radio base stations for the manual mobile telephone system
MTD, which was to be an interim system before NMT 450 got started. The frequency
of emission was the same as in the television transmitters where the firm had
accumulated some competencies, namely 450 MHz band. Magnetic was acquired by
Ericsson in 1982. See sections 5 and 8.
Radiosystem
The relationships between Magnetic and Radiosystem are more complicated than
just being small firms and competitors on the Swedish market. Radiosystem was a spin-
off company from Magnetic, a spin-off arising from conflicts and entrepreneurial
visions in 1978, the same year as the Televerket bidding for NMT.
At Magnetic, Torbjörn Johnson had been manager of the department designing the
radio base station. According to Mölleryd (1996:54), Johnson had an agreement with
Magnetic that they would pay him royalties on a product he had developed. When
Magnetic received an important order in 1978 but refused to pay him royalties, Johnson
decided to quit and start his own company. In June 1978 he convinced two other
Magnetic engineers to start a new company, Radiosystem Utveckling AB, or
Radiosystem. The first months of operation of Radiosystem were difficult since the
three engineers had to finish on-going projects at Magnetic as well as with the
development of own products. They therefore first worked as consultants for Magnetic
on the production of a new generation of television transmitters. In parallel,
Radiosystem bid for antennas for the NMT system to Televerket. After 10 years of
successful operation they were bought up by Ericsson in 1988 (Kågström 1997). See
sections 5 and 8.
Allgon
The firm Allgon was created in 1947 to develop and produce radio-antennas for car.
This activity expanded to antennas for the TV network and to military radio-
communication. In 1982, the management decided to start producing equipment for
RBS, mainly antennas and combiners, for different standards. The firm exported 90% of
its production in 1995 (Mölleryd 1996:57).
23
Spectronic
Spectronic entered the market of mobile telecommunication in 1985, and is now
developing and producing advanced mobile phones for the NMT 900 standard. Its
previous activity was in radio communication equipment.
The Operators
Televerket has been the operator of civilian mobile communication since its
inception from the mid-1950s. They were instrumental in testing and installing the first
three Mobile Telephone Systems (MTA, MTB, and MTD) in use in Sweden before
NMT11 (Hultén and Mölleryd 1993:2-3). Televerket was also instrumental in designing
and testing the NMT and GSM standards. See sections 4, 8 and 9.
Comvik and Europolitan are two private operators. Comvik is a firm belonging the
Swedish holding company Kinnevik. It was created in 1981 to start a new
telecommunication network and initially had an NMT standard but did not attract a lot
of users. Comvik used a proprietary system developed by E.F. johnsson with their own
mobile telephones. In 1992 it started operating a GSM network in Sweden. The operator
Europolitan was created in 1989 by two engineers previously from Ericsson Radio. It
was first called NordicTel, and had financing from four large Swedish firms—Custodia
AB, SAS, Spectra Physics and AB Volvo (Mölleryd 1996:65-71).
4. Switches for Cellular Telephony—Early History in Late 1970s
This section first describes the history of Ericsson’s traditional competence with
switches for fixed networks in order to set the stage for understanding why and how it
could move into switches for cellular networks. One thing that is particularly interesting
is that Ericsson’s previous knowledge, experience and even products (switches) for
public telecommunication networks ended up being the foundation for switches for
mobile telecommunication. Cellular telephony was initially seen as a very small niche
application area, and people in the public switches area laughed at the thought of
Ericsson becoming dominantly a mobile telecommunication firm.
4.1 Televerket, Ellemtel and the AXE-switch
Ericsson has had a long cooperation with the Swedish PTT Televerket (now Telia),
and the history of switches for cellular telephony really begins with computer controlled
switches for fixed (or public) telecommunication networks.12 Televerket and Ericsson
had co-operated during the sixties to develop computer controlled switches but those
11 The NMT standard was based on the test system MTC (Mölleryd 1996).12This section is mainly based on Vedin (1992, 1994) and Meurling and Jeans (1995).
24
were developed for different uses. Televerket was developing a switch, A 210, that
could handle many subscribers in big-city areas, whereas Ericsson was concentrating its
efforts on a smaller switch, E 12, that could be used in smaller townships and rural
networks. Televerket and Ericsson co-operated in some basic technologies such as
memories which were common to both. Cooperation was administrated by
”Elektroniknämnden”, a council jointly set up by the two actors. The aim of the
cooperation was to co-ordinate technology development, but the co-operation did not
work out very well (Meurling and Jeans 1995:27). They collaborated technically, but
Ericsson had no captive home market and so sold internationally.
In 1966 Ericsson decided to develop the AKE 12 even further and to build a more
powerful switch, the AKE 13. This lead them closer to a competing product for
Televerket’s switches. Televerket and Ericsson soon realized that the development of
switches was consuming more and more resources, and that the numbers of engineers
available in the home country was limited. They therefore decided to extend their
cooperation by starting up the joint R&D company Ellemtel to develop a computer
controlled switch. Allowing Ellemtel to proceed with its first R&D project, the AX
project meant that Ericsson’s own project to develop the AKE system had to be put on
ice. On the one hand, the AKE 13 would take much less time and resources to develop
compared to the AX, which would cost four times as much, but on the other hand the
AX-switch, if everything worked out, would be a superior product. Having taken the
decision that Ellemtel should proceed, Ericsson stopped its AKE development, and
allocated its resources for computer controlled switches to the AX project (Meurling
and Jeans 1995:43).
The first prototype version of AXE was ready in 1976 and was fully analogue. The
development time started in 1970, when Ericsson and Televerket agreed to develop the
switch under initial specifications, and ended in 1978 when the first switch was
installed.13 There were many questions about the AXE-system and whether it would
function, thus there were much uncertainties during the development process 1970-77.
The initial specifications and analogue solutions meant it was not an ideal construction.
Ericsson was ahead of its competitors by that time but had to digitize all parts of the
AXE soon if they were to keep their technological lead.
Before going further into specific projects and areas of knowledge which turned out
to be crucial for Ericsson’s change, we wish to say a few words about development of
the AXE switch because it may help explain some of the organizational and technical
changes necessary for its adaptation for cellular telephony. In the decade between 1976
and 1986, the AXE switch underwent large changes. The hardware and software were
13System specifications were ready in 1974; hardware design in the beginning of 1976; and software
at the end of the same year.
25
replaced completely in that ten year period. They are still continuously changing, but
the AXE architecture, or system platform, is still the same. The AXE-switch is
upgraded continuously, but it is the software which changes the most. Erik Öhrnulf
(1997), Vice President ERA, claims that the entire software is rewritten every third
year, due to new applications, standards and general improvements of performance.
The AXE-switch is a system platform for switches that has to be supported by
hardwares (standard microelectronics elements and microcircuits) and software
(programming systems) to function. Hardwares which are not strategically vital are
purchased on the open market whereas parts considered strategically important are
manufactured by supplier companies with close and stable relations to Ericsson. The
programming systems has been to a large extent assured by purchases from the open
market (Meurling and Jeans 1995:214), although Ericsson itself has largely become a
software programming firm too.14
Adapting the AXE switch for different markets and customers, or what is known as
development of the application design, is mainly done in-house in Ericsson, or
sometimes by joint venture companies in global markets. During the 1980s, Ericsson
had joint ventures with the following companies for this: Thorn-EMI in the UK,
Atlantic Richfield in the USA , Matra in France, OPC in Korea, and Telefonica in
Spain. Ericsson has today joint ventures with companies in several countries such as
Japan, China, Malaysia and Germany (Meurling and Jeans 1995:210). Moreover, as it is
important to be close to different markets, so Ericsson pursues AXE-related research
and development all over the world. For example, the Ericsson laboratories in Aachen,
Germany does research on mobile AXE-switches for the GSM-system while the
laboratory in Richardson, Texas, was initiated in 1984 to adapt AXE and the office
switch MD 110 to the American market (Smith 1984).
4.2 AXE and the Nordic NMT Systems
At the same time that the AXE-switch was being developed for public
telecommunication, cooperation began to develop around a common Nordic automatic
mobile telephony (NMT) standard in January 1970. In 1981 the NMT 450 system was
ready to function in commercial use. It thus took an enormous amount of time, ten
years, to specify the standard and get it functioning.
In 1971 the NMT-group gathered potential suppliers for the mobile telephony
system, and around forty different national and international companies received
preliminary specifications from the group composed of PTTs of Norway, Sweden and
Denmark. Televerket’s radio labs (Radiolaboratorium) then tested the system and its
14The system platform is a completely in-house business (together with Televerket and Ellemtel),
and the basic structure is the same today as it was in 1976 (Öhrnulf 1997).
26
specifications for another two and a half years, that is, 1975 to 1977/78. They tested
switches, radio base stations and terminals in order to make them handle the demands of
the specifications.
The Nordic PTTs considered it important to test the system substantially before
introduction for commercial purposes. Moreover, it was considered important that
testings of the system should not be done by firms in the industry in order to avoid
future problems of competition. To give all firms equal access and none inside
information, Televerket’s R&D division Radiolaboratorium did the testing of the
experimental types (Mäkitalo 1997). The PTTs would be the network operators.
In 1977 the NMT-group opened bidding for supplying switches to a number of
companies. Bidding was international, but Ericsson won the order to deliver switches to
Sweden, Denmark, Norway and Finland. Ericsson’s main competitor was NEC.
The various histories of Ericsson’s involvement in switches for mobile telephone
networks tell slightly different stories about which switch Ericsson offered for the initial
NMT order. Mölleryd (1996:47) claims that Ericsson at first intended to deliver the
AKE-13 switch, which had been developed in the early 1970s whereas Televerket
thought that it was not sophisticated enough and therefore demanded AXE instead.
Meurling and Jeans (1994:50) expand this to argue that the Ericsson switching group,
BX, felt that the AXE-system was not fully developed so wanted to sell the AKE-13
switch instead (a computer-controlled switch with electromechanical switch elements).
However, Televerket wanted the AXE switch because the digital version had been
tested in commercial use. Other sources stress that Ericsson really had no choice but to
deliver AXE. According to Åke Persson (1997) who was then engineer at BX, there
were no alternative to AXE for the mobile telecommunication application. The AKE-13
project to develop a transit station system was just in the process of being shut down.
According to Östen Mäkitalo (1997), who was head of Televerket
Radiolaboratorium at that time, Ericsson offered the AKE-13 switch during the bidding.
Televerket then made it clear to Ericsson that if they did not offer the AXE, Televerket
would choose the digital switch from NEC. NEC had a similar type of digital switch as
Ericsson’s. What were Televerket’s reasons for wanting a digital switch from Ericsson?
Not only had they done experimental testing on one but Televerket also wanted to be
able to introduce services in the mobile network that demanded a modern digital switch.
Moreover a digital switch was seen as important for handling the functioning of
roaming, and Nordic roaming was one demand of the NMT specifications (Mäkitalo
1997).
Televerket felt that one important reason for choosing AXE over NEC’s digital
switch was that it is natural to have the same switches in all their telecommunication
networks (Mäkitalo 1997). Televerket would be operator for both networks. Moreover
Televerket would have considered it a failure if it could not have used the switch
27
jointly developed with Ericsson! Mäkitalo (1997) emphasized that choosing Ericsson
over NEC had nothing to do with national industrial policy, such as supporting national
champion firms.
There was thus a very initial discussion between Televerket and Ericsson over
which switch would be appropriate. An important question was whether AXE was an
appropriate switch for the NMT- network, and more generally, what size of switches
would be appropriate.15
Åke Lundqvist, then managing director at SRA, was the first within Ericsson who
tried to convince the switching section of Ericsson (BX) to deliver the AXE-switch for
the NMT system. At this stage Ericsson had a division of labor within the company
between BR working on RBS (including the jointly owned firm SRA) and BX, working
on switches (See section 5). However, switches for the NMT 450 were low on BX's
priority list.
This was probably one of the reasons why SRA tried to build a switch of their own
for mobile networks. Although SRA had worked on such a switch and had some
prototype version at the time, BX definitely felt that SRA did not understand the
complexity of the public telecommunication network nor of switching. SRA simply did
not have the competence to bring it off. Åke Persson, engineer within BX, insisted that
it would be better to use the AXE-system instead of the one developed by SRA (Persson
1997).
Nonetheless, the order for switches for the NMT 450 standard lead to a resource
allocation problem between SRA and BX due to the fact that Ericsson already had
problems delivering the AXE-switch to customers in public telecommunications
(Meurling 1997, Meurling and Jeans 1994:51, Persson 1997). They also had to adapt
AXE for the cellular network. The main adjustment was not hardware related but was
rewriting software for mobile communication. The AXE software for NMT 450
application was written by BX, and accounted for about 40% of the development costs
for use in mobile telecommunication (Meurling and Jeans 1994:52). There were only
about 50 men who worked with the development of the AXE-switch for the mobile use
(Persson 1997).
When AXE was being developed, the main use envisioned was as a switch for local,
public (fixed) networks. Ericsson did not see mobile telecommunication as a future and
prospective market, and so consequently, they thought that the application of the AXE
switch for the mobile telephone network would be limited. However, the internal
15Questions arose because the AXE switch was more expensive than the AKE switch and because
the mobile telephone market was very limited at the time. Ericsson's AKE-switches were smaller and lessexpensive than their competitors. Ericsson and Televerket would thus save time and resources becausethe AKE-switch was fully established and functioning, whereas the AXE switch was still not fully provenon the field. At the time (1978), there were only 2-3 AXE switches functioning in the publictelecommunication network (Persson 1997).
28
discussion over who had the resources and especially interest to develop a switch for
mobile network quickly became a discussion about which division had the technical
competencies.
4.3 First Operational NMT-system—Not Sweden but Saudi Arabia
Although NMT 450 was a technical standard developed by the Nordic PTTs, the
system was first used in Saudi Arabia in 1979. This was an important event in the sense
of trying out the NMT system, but it is also a bit of history important for Ericsson and
AXE for both fixed and mobile networks. Ericsson teamed up with one of its major
competitors in telecommunication for part of the order.
Saudi Arabia had decided to build a completely new (fixed) public
telecommunication network at the end of the 1970s; Ericsson and Philips joined forces
to have a better chance of getting that order; and in 1978 the joint venture Ericsson-
Phillips won the order. Ericsson first tried to convince the Saudi Ministry to choose
their ARE-switch for the fixed telecommunication, but the Saudis wanted and got the
AXE instead. The contract was the biggest ever signed in the telecom business at the
time. The contract with Saudi Arabia was very important because it ended a period of
uncertainty for the firm concerning the competitiveness of the AXE-switch (Meurling
and Jeans 1994:66).
In 1979 Saudi Arabia wanted to do further investments in switches and transmission
for their public telecommunication network again in order to expand it. Åke Lundqvist
(1997) suggested to the Saudi Minister of Communications that a cellular network
should also be installed.
One day in 1997, Åke Lundquist, then CEO of Ericsson’s little radio company SRA,went into the room of the Ericsson CEO concern Björn Lundvall...’Can’t we try to sella mobile telephone system to Saudi Arabia?’ Lundquist wondered. ’They want thelatest of everything else.’The proposal was unusual. Ericsson didn’t have a mobile telephone system to sell.’But if we just get an order, surely we can pull together the pieces to deliver one’,Lundquist reasoned (Palmgren 1992:52).
The Saudis agreed, but it was not clear at first that Ericsson would deliver the NMT
system because the Ericsson project in Saudi Arabia involved cooperation with Philips
and Bell. The first idea was to use the Philips system, which was designed to operate in
the 160 MHz band, but that turned out to be impossible because that frequency was
already occupied by the Saudi Arabian military (Meurling and Jeans 1994:55). Ericsson
had by then started to deliver switches and parts of radio base stations to the NMT-450
system in the Nordic countries, so Per Åkerberg, head of land mobile radio and
telephony systems at SRA, and Håkan Ledin, head of BX, suggested that Ericsson
should deliver a complete NMT 450 system instead. Philips was not very fond of
handing it over to Ericsson, but a few months later it was agreed that Ericsson would be
29
the sole supplier of a NMT 450 system to Saudi Arabia. However, the system was not
functioning technically when the order was signed, so the development of the NMT 450
had to be accelerated within Ericsson (Lundqvist 1997, Meurling and Jeans 1994:67).
This deal turned out to be very important for Ericsson, not only for seeing that the
system worked but also for financing the whole cellular telecom development for quite
some time (Lundqvist 1997). In other words, the deal was very profitable!
4.4 Changing Responsibility—SRA and Systems
More and more PTTs in especially Europe became interested in developing cellular
networks in the late 1970s, and they started to send out requests for proposals to the
manufacturers. This increasing demand lead to organizational changes within Ericsson
as well.
Svenska Radio Aktiebolaget (SRA) put together Ericsson’s proposals at first, but as
the mobile telephony business grew bigger there had to be a closer co-operation with
the public telecom division (BX). The first organizational change was to set up a joint
working group to do the proposals, but soon Ericsson’s top management decided that
SRA should have the overall responsibility for mobile telephone systems, so in late
1981, SRA was given full business responsibility (Lundquist 1997). Björn Svedberg,
then CEO for the whole Ericsson concern, was responsible for lifting the full business
responsibility for mobile telephony to SRA and hence to Åke Lundquist, head SRA.
SRA became a fully owned part of the Ericsson concern in 1983, called Ericsson Radio
System (ERA).
However, BX still had responsibility for developing the AXE switch for all
applications, including mobile networks. By 1977, BX had already been fully involved
in the work with AXE and taken over the R&D work from Ellemtel, which was a
complicated transfer. To transfer the system design and R&D at Ellemtel into
production at Ericsson, BX demanded extensive and thorough documentation as well as
an extensive transfer of engineers from Ellemtel to Ericsson (Mäkitalo 1997). A
division and balance of power between BX and BR was thus organized, but at the same
time they could support each other technically and in market questions and hence could
(hopefully) act together (Meurling and Jeans 1994:40).
In summary, the history of Swedish business in switches for mobile
telecommunication is almost exclusively a history of Ericsson. Within Ericsson,
switches for cellular networks were initially seen as a very small, specialized
application, and there was a certain juggling of resources and viewpoints between
different divisions, including SRA.
The exceptions to seeing it as dominance by Ericsson is the crucial role played by
Televerket. There is Ericsson’s long cooperation with the Swedish PTT Televerket;
Televerket’s crucial role in specifying the NMT 450 standard; and AXE being
30
developed in the joint R&D company Ellemtel. Televerket was the driving force when
it came to specifying and testing the NMT-450 system, whereas the other Nordic PTTs
agreed with the system but did not have resources to become a major part of the
development process. Nonetheless, the telecommunication companies were consulted
continuously during its development, so that the suppliers could simultaneously develop
components and equipment for the mobile telecommunication system (Mölleryd
1997:26).
However, this is the only point of connection with the SI of crucial importance to
innovation. Universities and other science institutes were not involved in the NMT 450
work at all. Other than Ellemtel, there are basically no small companies involved
directly in the market and technical developments. Thus, there are no new firms in the
switching area and over time, and Televerket reduced their in-house technical
competence and eliminated their in-house production of switches. This has left Ericsson
as the dominant Swedish actor in this technological area.
5. Parallel Early History of Small Firms in Radio Communication, to1980
Unlike the switching field where the company Ericsson has been the dominating
figure in the Swedish story and where the technology and products lay close to previous
Ericsson ones, the history of Swedish firms in radio base stations for mobile network is
more complex. Small competing firms, including some which were owned by Ericsson,
are the main actors. The core of Ericsson’s competencies developed out of the small and
long-standing jointly owned firm with Marconi, namely Svenska Radio AB. SRA had
been heavily specialized in military applications. Later on in the 1970s, 1980s and
1990s, other small and often competing firms either moved into civilian communication
or were started up but bought by Ericsson. This section will address the dynamics of
these small firms competing over radio communication, namely SRA, Magnetic and
Radiosystem.
5.1 Ericsson and SRA
Ericsson’s initial competence in radio communication for radio base stations (RBS)
and phones can be traced back to SRA. Since the early years of mobile telephony and
trials from the 1950s on-wards, SRA had been involved in designing RBS, or
components of them as well as the phones. These were initially for military purposes
although some niche civilian markets were also identified, and Televerket had a number
of systems.
According to persons within SRA, the fact that Marconi owned 21% of SRA was
initially important:
31
SRA had an independence in technology and culture, which allowed them to try thingswhich did not find room in Ericsson as a whole. Also, they keep their profits (ratherthan subsidizing other parts of the larger concern) and plowed that money into moreR&D. Being independent and small allowed them to start a small project on mobilephones in SRA which central Ericsson management saw as not very important. Due tofinancial independence, SRA could continue pursuing these R&D projects over time,without immediate threats of cancellation, when they kept using money, year afteryear, without any profits. Once bought completely by Ericsson, SRA lost much of thisindependence (McKelvey 1997b:5).
The importance of SRA's independence at an early stage of mobile telephony can be
summarized as that there was lots of money, long-term strategy for technical
development, independence, and room for brave ideas (Lundquist 1997).
When the Nordic PTTs called for bids for the different components of the NMT 450
networks in 1977, several firms met with a bid, among others, SRA, Magnetic,
Mitsubishi, Philips and Nokia. SRA replied but could only provide control units for the
RBS. Compared to contemporary systems, the control unit was a major part of the
system, and an important novelty as the system was fully automatic rather than going
through human operators (Kågström 1997).
There were several reasons why SRA was not ready with the development of RBS.
The first is that SRA was involved in the American market. SRA was early to work in
the American civilian market, and although this strategy gave knowledge about the
early American mobile telephony specifications, the firm did not concentrate on the
Nordic market and so was lacking knowledge about the Swedish and Nordic technical
specification.16 In this case, access to Swedish specifications can thus be argued to be
important for the firm's abilities to develop component products.
A second reason was that in the Swedish market, SRA was focused on mobile
telephones themselves rather than RBSs. To expand its competencies in the domain of
mobile telephones, SRA acquired Sonab, a Swedish manufacturer of radio equipment in
1978. However, this acquisition also turned out to be important for RBS for the Saudi
order as they included components from the independent company Radiosystem as well
as ones previously designed by Sonab. The sole component developed within the
previous SRA (i.e. SRA minus Sonab) was the control unit (Meurling 1995:53). Thus,
SRA acquired another small Swedish firm in order to quickly gain competence in a
missing area of technical competence, namely phones.
5.2 Magnetic and Radiosystem
In 1978 Televerket bought the main parts of the RBS (transmitters, receivers,
antennas, amplifiers and filters) for NMT 450 from other firms than SRA, with half of
the order going to the Swedish firm Magnetic, and the other half to the Japanese firm
16Remedying this problem was one of the major reason for buying their competing firm Magnetic in
1982 (Meurling and Jeans 1994:51, Kågström 1997).
32
Mitsubishi and the Swedish firm Radiosystem (Meurling 1995:51, Kågström 1997).
However, for both systems, SRA produced the control unit for the RBS.
The competing firm Magnetic had mainly developed military technologies for radio
communication, including radio equipment for the airforce, and measuring systems for
the Navy and the Air Force. They came into RBS for cellular networks through their
development of TV transmitters in the 1970s as well as the MTD system.17 For the
Swedish NMT 450 order, Magnetic produced transmitters, receivers and filters,
antennas, amplifiers and combiners.
The third competing firm in RBS was Radiosystem, which was a spin-off from
Magnetic in June 1978. The three engineers who moved had all been very active in the
development of the MTD system at Magnetic and could hence transfer technical
competence in filters, combiners, antennas, amplifiers, transmitters and receivers as
well as on the NMT technical specifications to the new firm.
Kågström (1997) argues that this was important information as the first
specifications of NMT were only available for firms which were known from their
previous involvement in the production of radio equipment, hence giving inadvertent
advantages to existing firms over start-ups.18
Thus, the first order Radiosystem received was from Televerket, where Mitsubishi
produced transmitters and receivers while Radiosystem produced filters, antennas,
amplifiers and combiners. This was very important for financing, because it meant they
would receive 30% of the cost of the project in advance to buy components. As this was
a small start-up firm, financing for growth was a crucial question.19 Although the firm
quickly won this order from Televerket, its first years involved difficulties finding
financing and competent (technical) personnel. Therefore risks of going bankrupt were
high.
5.3 Cooperation and Competition
These three firms cooperated as well as competed. Radiosystem initially consulted
for the competitor and firm it had started at, Magnetic. In fact, Radiosystem also had a
working relationship with SRA between 1979 and 1985. Like Magnetic, SRA was
involved in production of early mobile phone stations (that is, telephones mounted in
cars or trucks) for the Swedish MTD system (Kågström 1997). In 1979, SRA contacted
Radiosystem to develop and produce antennas and amplifiers for the NMT system to be
installed in Saudi Arabia. Moreover the first tests on the NMT system to be sold in
17That is, the intermediary mobile communication system used before NMT could be installed.18However this argument is not followed by Meurling (1997) who claims that the specification
about the NMT standard was widely available so that any firm could bid.19When the firm started in June 1978, the capital was 5000 Swedish crowns ($1000 USD at the
time) and each of the three founders took bank loans, using their houses as security. Radiosystem laterreceived a credit check of 125 000 Swedish crowns (25000 US$) (Mölleryd 1996:54).
33
Saudi Arabia were done in the laboratories of Radiosystem (Kågström 1997). There
were some relationships between the small Swedish firms competing over radio
communication in terms of consulting and/or developing sub-components of RBS.
Thus, the early history of firms in radio communication in Sweden is much more a
story of small competing firms—SRA, Sonab, Magnetic and Radiosystem—than of the
large concern Ericsson. There were initially three small Swedish firms involved in radio
communication, the key technology for radio base stations and cellular phones. Ericsson
owned a significant share of SRA, but they acted like a small independent firm. The
first three were small firms which had existed for a number of years, which had
significant military business, and which had competence in civilian radio
communication. SRA had been involved in the development of the Mobile Telephone
system A and B (MTA and MTB) in the 1950s and 1960s while Magnetic produced the
RBS for the MTD in 1975. Sonab produced phones. Although military contracts gave
resources, and reasons, to develop radio communication competencies over a longer
time, development of TV transmitters seems to be a linchpin technology for Magnetic's
move into cellular networks. Radiosystem differs from the others in that it was started
relatively late, in 1978, as a spin-off by an entrepreneur. Nonetheless, their initial
technical staff moved directly from Magnetic and hence took competencies with them.
6. Ericsson as Mobile Telecommunication System Provider from 1981
Ericsson engaged in organizational change when it began to see itself as mobile
telecommunication system provider, although these were initially fairly minor changes.
Although SRA was a jointly owned company, it was given full responsibility for this
business area from late 1981. BX, however, would continue to develop software and to
develop and manufacture the AXE-switch. The exception to SRA’s overall system
responsibility was the Nordic market, where the customer PTTs would still purchase
their radio base stations and switches to BX and SRA/ERA separately because they
took responsibility for the overall system design and implementation (Meurling and
Jeans 1994). In other countries, PTTs and other operators began to want package deals.
The important point with this organizational change is thus that Ericsson began
delivering systems (other than in Scandinavia) rather than delivering separate
components for mobile networks. However, as we shall soon see, this change in
organization, but especially in the concept of what Ericsson was doing and selling, was
not an obvious change but one which had to be argued for. Again, the in-house
entrepreneur Åke Lundquist at SRA was instrumental in trying to develop the
innovative potential of this new technology.
PTTs in Europe became increasingly interested in mobile telephone networks from
the early 1980s and on-wards, and an order from Holland turned out to be a turning
point for Ericsson as a mobile telecommunication system provider. Holland was
34
important because it became a turning point in perceptions internal and external to the
company about Ericsson being a supplier of complete mobile telephone infrastructures,
and not just a supplier of bits and pieces.
6.1 Holland 1982, a Turning Point
In 1982, the PTT of Holland decided to implement a NMT 450 network. Selling the
previous package deal (with Philips) to Saudi Arabia apparently went smoothly because
the buyer had been looking for a system solution, whereas the Dutch PTT was willing
to create its own system. The deal therefore required more active arguments inside and
outside Ericsson. The Dutch wanted to buy AXE switches from Ericsson and RBS from
Motorola. Not only were the components and communication among them so well
specified that an operator could mix and choose among (identical) options, moreover,
the Dutch—and others—could use the NMT specifications freely and paid no
reimbursement to the Nordic PTT for their long-term development and testing of this
standard.
Once again, it was the CEO of SRA, Åke Lundquist, who argued that Ericsson
should be a system provider. Lundquist had previously done some active marketing to
convince the Saudis that the system would work (Meurling and Jeans 1994:55), and
now he argued both within and without the firm that Ericsson sold systems, not
components. Åke Lundquist argued to Björn Svedberg, CEO of the Ericsson concern,
that they had to sell the infrastructure, i.e. switches and radio base stations, as a package
deal. The problem with Lundquist's argument that Ericsson should sell the whole
system, or nothing, was that Ericsson could lose the whole order, especially at a time
when switches were the main product of the firm. Switches for mobile networks were
only a special and small application area. However, Åke Lundquist at SRA did enjoy
support of some managers at BX, especially Hans Flink and some others. Together but
especially due to Lundquist's visions and drive, they convinced the top Ericsson concern
management that Ericsson had to sell a mobile network as a system. (Lundquist 1997,
Meurling 1997).
Convincing the Dutch PTT required showmanship. Lundquist (1997) said, ’This
was the only time I slammed my fist on the table while arguing with a customer’. One
argument they used to persuade the PTT that a package deal was a good idea was that if
Ericsson delivered a whole system, Ericsson would also take the responsibility for
configuring, designing, and hence the initial functioning of the mobile network.
At that time, the competencies of SRA in RBS were in fact very limited, and SRA
did not master the small-cell concept that was required. The management mentality at
SRA was, however, that problems were there to be solved, so rather than going home
and first developing in-house competencies and then making a bid on another system in
a few years, SRA did what they could to convince the Dutch that they knew what they
35
were talking about and then patch together a solution. Lundquist contacted two
American engineers who had been working as consultants for SRA—Chandos Rypinski
and Jan Jubon—to help SRA design a mobile cellular system (Meurling and Jeans
1994:59).20 The time for having the specification ready for the Dutch PTTs was very
short—one day. So the following day, the two of them as well as Åke Lundquist and
Åke Persson, an engineer originally from BX and newly employed at BR to adapt AXE
system to mobile telephony, took a plane to Holland and tried to convince them they
knew what they were talking about.
Åke Lundquist and Åke Persson had completely different interpretations of the
situation and risks involved in getting a system to work in such a short time. Åke
Lundquist said the following about selling a whole system to Holland or nothing:
We did not have so much to lose, really...There were many decision then which were takenwith great risks...That's why we were called horse-dealers by the BX... Without us therewould never have been anything... BX saw this as a minor part of the AXE-system (ÅkeLundquist 1997, author's translation).
Åke Persson had a different memory of the first meeting with Lundquist, Rypinski
and Jubon just before the bid to Holland:
That is a thriller, it was terrible...Then I was here at SRA and I met Åke Lundquist for the firsttime...(and Mats Ljunggren) and they were absolutely lost...They did not really understand,they made base stations and there was not much to it. It was Televerket Radio who had theknowledge about basic theories,21 I think...I sat in a conference room and listening to loadAmericans and wondered, Can this be true? (Åke Persson 1997, authors translation).
Somehow, though, they won the Dutch bid for a whole system and opened for this
new Ericsson strategy of selling systems.
6.2 Implications of, and Strategies for, Becoming a System Provider
This change in strategy implied a number of things. For the first, it implied a new
organization to handle the sales of mobile telecommunication systems, rather than of
components. It also signaled a change in understanding of what the firm was selling.
Rather than selling hardware components, they would take over responsibility for
designing and implementing each system they sold, hence they were also selling a
service. Network operators would have correspondingly less responsibility and need
less technical competence.
Parallel with this change in overall strategy for mobile telephony, Ericsson was
finding additional reasons to buy all the shares in SRA. In the 1970s and early 1980s,
20The impact of those two engineers would be even greater since they later helped convince Åke
Lundquist to enter the American market in mobile telecommunications (Meurling and Jeans 1994:60,Lundquist 1997).
21Answering a direct question Persson said that basic theories were on basic radio transmition andreception, as well as theories about cellular networks.
36
some conflicts had occurred between the BX (switching) division at Ericsson and
Marconi. One of which was that Marconi was partly owned by GEC Communication, a
direct competitor to Ericsson in switching technologies (Lundquist 1997). Moreover,
Ericsson had some financial reasons for owning all of SRA, namely that they could then
access all the retained earnings and profits. The part of SRA concerned with
development of mobile telephony started to be profitable (Lundquist 1997). In order to
control the activities of SRA, Ericsson would have had to owe more than 90% of the
share. Ericsson bought them out in January 1983, at which time SRA was integrated
into the Ericsson concern and renamed Ericsson Radio System (ERA). Turnover more
than doubled each year between 1983 and 1985, and was the most profitable business
area (Affärsvärlden 1985a:17).22
In addition to becoming a mobile system provider, Ericsson was also selling the
phones (also known as handsets). The first phones were for the NMT 450 system and
hence introduced in 1981. ’Mobile’, ’take-along’ or ’car’ phones were at that time seen
as something to be permanently installed in vehicles. They were about the size of
suitcases. By 1984, Ericsson had the lightest of the portable phones at 6.1 kilos
(Mannberg 1984:64), but the vision of use was hardly what it is today of using
miniature cellular, or mobile, phones everywhere. According to a Bell Labs cellular
expert in 1983 : ’ We asked businessmen if they could use a portable, and the typical
response was, “Where would I use it? Would I stop in a corridor to answer a call? Or
answer a portable at a meeting?”’ (Free 1983:28). The visions of projected use were
thus quite limited, often tied to specific instrumental uses like emergencies, and so the
forecasted market was quite low.
For Ericsson, the phone side of the business was apparently initially seen as
something necessary to provide along with the mobile network but nothing with a large
future as an independent product. The competence to design and manufacture phones
was the same as that for RBS, as the core technology in both rely on radio
communications. Thus, the Swedish firms Sonab and SRA, both owned or partly owned
by Ericsson by the mid-1980s, had had previous experience making phones both for
military uses and for the civilian radio communication systems starting from the mid-
1950s in Sweden (Hultén and Mölleryd 1993:2). Moreover, Ericsson had been
manufacturing phones for fixed networks essentially from its beginning in the late 19th
century, but this was a relatively small product, not even an independent business area
by 1983 (LME 1995). So mobile phones are in some ways a continuation of a previous
Ericsson product line, but they are based on a different technology (radio
communication) and in the 1980s at least, being sold to a niche rather than mass market.
22In the same period, the number of employees increased from 400 to 1 800.
37
Even with these initial and profitable orders in Saudi Arabia and Holland in the
early 1980s, the top management of Ericsson—who was in good company with many
other firms—did not believe very much in the potentials of mobile telephony. It was
seen as a quite small, specialized market and not expected to grow very rapidly; it was
indeed a very small percentage of sales, overwhelmingly outsized by BX and even
Ericsson Information Systems (EIS). In 1986, radio communication was still only 8% of
the concern (See section 3).
The Ericsson concern management, led by the CEO Svedberg, had their visions
fixed on another technology and vision of innovation. So, the ownership structure and
business culture of SRA allowed for independence, which, together with an explicit
strategy to become less dependent on military business, lead to what was initially a very
small project on mobile communication. If top management had not allowed room—for
whatever reasons whether conscious or a mistake—for mobile telecommunication to
develop there and also given resources for complementary products, then Ericsson
would be a firm in crisis today, not a darling of stock market analysts.
7. Ericsson’s Vision for the 1980s—Information Systems, not MobileTelephony
The top management of Ericsson, led by CEO Björn Svedberg, had other visions of
innovation opportunities than those particularly Lundquist and SRA/ERA were
identifying and acting upon. For Svedberg, the vision of the Ericsson future was in the
integration of telecommunication and office work in the ’office of the future’ or
’paperless office’. This would be Ericsson Information Systems (EIS). The prospects of
growth were estimated at 15% per year, while telecommunication would only grow by
6%. Moreover the fact that sales of public telecommunication systems depended on
national political decisions was considered more and more as a burden. Going into this
new market would permit Ericsson to have customers in the private sector, mainly
firms, and not mainly national PTTs (Affärsvärlden 1981:25).
7.1 Risks and Possibilities
Investing in information technology was a clear change to a new sector for
Ericsson. It entailed changes in the type of customers and technical specialization.
Moreover, as in radio communication, buying firms was a new strategy in how the firm
would gain competencies.
Despite significant differences in technologies and product areas, some
complementarities with traditional areas of Ericsson competencies could be identified.
(Månsson 1981:31). The development of AXE had given Ericsson competencies in
system development and also clearly shown that the future of telecommunication was
no longer in electromechanical skills but in computer programming. Ericsson therefore
38
had a stock of computer programmers who could be transferred to the new activity.
Moreover, Ericsson had faced problems with internal data communication and so had
developed its own in-house integrated communication systems which could be useful.
The early prognoses on EIS was that it would be about 20% of the turnover of the firm
within few years (Månsson 1981:32). Thus the prospects of expected growth were high,
and AXE acted as a cash cow for other projects, inducing optimism (Sundquist
1988:12).
For Information Systems, Ericsson had to buy other firms to rapidly have access to
technologies. In late 1980, Ericsson bought the firm Datasaab, which was the computer
division of Saab and then Facit, a producer of mechanical-based office machinery
turned computer manufacturer, in October 1981. Through the purchase of Datasaab,
Ericsson gained knowledge of computer terminals, microcomputers, network equipment
as well as of marketing of products for the private sector, especially firms. Buying
Datasaab in 1980 provided the firm with terminals, minicomputers and network
equipment. The purchase of Facit brought computer printers and other office equipment
(Sundquist 1988:12). Ericsson started to produce IBM compatible personal computer in
1984. As the PC was seen as the main node of a ’paperless office’, it would be
important for Ericsson to have a strong position there. The goal of Ericsson was to
become a giant in the various parts of computer technology.
Ericsson’s decision to diversify into computer communication was regarded as a
very risky investment, by observers both in Sweden and the USA. The main risks were:
* Ericsson had to enter a market where firms like IBM or ITT had entered a couple
of years earlier. They would be followers, imitators, rather than leaders and have to
battle for market share (Meurling 1997, Affärsvärlden 1985a:19)
* Moreover, it was not clear how to integrate the new division within Ericsson,
hence implying a very large restructuring (Affärsvärlden 1985b:26)
* Ericsson did not know how to manage sales of consumer electronics as personal
computers (Affärsvärlden 1985b:26). It meant diversifying into a new product area
where the firm had no previous experience.
* Finally, competencies were still lacking, both in technologies and in marketing.
Buying firms was thought to be one way to obtain them but particularly the purchase of
Facit was not being successful. According to Håkan Ledin, former CEO of EIS, the
decision to buy Facit was controversial and forced by Marcus Wallenberg (Källen
1989:19).
Ericsson management accepted these risks, although recognized it would take at
least two to three years before the investment would give returns. EIS would have four
divisions initially: 1) Communication Systems (Business switch MD110 and office
network communications), 2) Alfaskop Terminals (Computer terminals based on
Datasaab), 3) Business Systems (Datasaab bank teller terminals and minicomputers)
39
and 4) General Terminals (Voice, computer and text systems; Facit equipment). (Skole
1982). Ericsson products were successful on the European market, with a sixth of the
3000 largest European companies equipped with Ericsson's technology Alfaskop in
1985 (Månsson 1986:25). However the market for PCs in the US had been
overestimated.
7.2 Failure
Table 4 shows the tremendous expansion of EIS, especially in terms of persons
employed, with a high of 21,700 in 1984. As a percentage of the Ericsson concern, it
was approximately 30% between 1983 and 1987. The profits, however, were dismal.
Table 4: Data about Ericsson Information Systems
1982 1983 1984 1985 1986 1987
Sales (mkr) 4400 7500 9300 10600 9600 9700
%of Ericsson concern 23 28 30 30 29 30
Profits/losses (mkr) - 237 -217 -806 -284 0
Number employed 13240 19150 21700 20800 16800 14600
Source: Affärsvärlden (1988a:56)
EIS thus represented an enormous investment by Ericsson but an unprofitable one.
In fact, the only year with profits was 1983. The restructuring of Ericsson Information
System started in fall 1985, and continued during the three following years without
signs of improvement. In 1988, EIS was bought by the Finnish Nokia which also had a
computer division (Affärsvärlden 1988:4). Ericsson refocused its activities to
telecommunication.
There are many reasons for the EIS failure, involving both market and technical
aspects. One market reason was that the investments were based on a vision which was
over-optimistic about the promises of the ’paperless office’. Ericsson was not the only
firm which failed to integrate telecommunication and information technology. Neither
did the market expand at the rate and in the directions expected. Other reasons can be
found internally in the firm. Despite Ericsson’s initial emphasis on office systems to be
sold to firms, computers represented a mass market, where the competition was much
tougher and based more on price than technical finesses. This differed from Ericsson’s
stronghold products.
Moreover, Ericsson products had too many technical faults. An example is the
switch MD 110, which was to connect all components of the office information system
(and was based on technologies developed for AXE). More than 3 000 programming
bugs were detected after more than 300 systems had been sold (Datornytt 1985:37,
40
Affärsvärlden 1985a:17). To these technical problems were added logistics difficulties
in receiving some components in time. Moreover, the decision to buy Facit to exploit
technical competencies was regarded as a mistake. Facit was not at the technological
edge and was losing money when Ericsson bought it (Källen 1989). Finally,
restructuring to integrate the small firms within the larger organization was not
successful.
Thus, the idea of EIS which would integrate telecommunication and computer
communications had the strong backing of top Ericsson management and received the
resources and manpower to expand rapidly. It would be a diversification into a new
product area. EIS was also a failure. Vision, management support and resources were
not enough to overcome the market and technical problems. EIS was sold off, and
Ericsson was to become a telecommunications firm again in 1988.
8. Early Telecommunication Deregulation to 1985—Impact on Firms
In parallel with the rise and fall of EIS during the 1980s, there were a number of
important developments in mobile telecommunication, specifically expanding market,
deregulation of telecommunication generally, technical advances, and increasing firm
experiences. These were changing Ericsson into a telecommunication firm specializing
in cellular telephony, with a side business in public switches. The same trends were
affecting the other Swedish firms as well. Within Ericsson and especially due to the in-
house entrepreneur Åke Lundquist, the vision of Ericsson as a system provider for
mobile infrastructures was becoming increasingly accepted during this period. It also
lead to organizational changes, despite its very marginal size in the company, relative
both to public switches (BX) and to Ericsson Information System. In other words, the
high tech vision held by the top management in the 1980s was proven incorrect and a
business area considered minor by them but which was seen as the vision of the future
by other in-house entrepreneurs at SRA was shown viable, and profitable. What was
soon clear, however, was that the international wave to deregulate telecommunication in
the 1980s was also impacting the growth of mobile networks.
The orders to Saudi Arabia, the Nordic PTTs and the Dutch PTT were all based on
the NMT 450 standard, and so the very detailed specifications of this standard were
important to the firms’ abilities to deliver very specific products. The NMT standard
allowed equipment from different manufacturers to be used since it specified the
interfaces between the different sub-systems of the infrastructure. Competing standards
internationally in combination with deregulation meant, however, that the firms had to
decide strategy for which technical alternatives within cellular telephony to pursue.
Decisions about whether to sell mobile infrastructures both in Europe and in the USA,
which were the countries initially investing, were related to questions about the best
business strategy. Would it be better to develop technologies for all standards or to
41
concentrate on one? Would it be better to try to be a technical leader or a (fast) imitator?
The choices were not obvious, and competing firms made various combinations of
decisions.
8.1 American Market
In the beginning of the eighties it became clear that the American
telecommunication market would be deregulated. There were disagreements within
Ericsson about whether it was worth the market and technical challenges to try to get
into the American market, where Ericsson had never had a presence and would have to
fight the giants AT&T and Western Electric. Ericsson’s initial move in 1981 was a joint
venture with Atlantic Richfield to sell cables and which they bought out in 1985
(Dagens Nyheter 1989). Already by 1982, the US was projected as the major market for
EIS products, particularly the office switch MD110 as well as computers (Skole 1982).
In 1984, the American market was deregulated, AT&T was broken up into ’Baby Bells’
regional companies, and the market for telecommunication equipment opened for
foreign firms, which induced Ericsson to open a head office and R&D center to modify
AXE outside Dallas, Texas, and to try to get into the market (Dagens Nyheter 1989). By
1987, the American successes and failures were approximately: EIS sales were far
lower than expected; in mobile telephony, they had 35% of deregulated market; were
the third largest cable producer and had contracts with all seven regional Bells to sell
AXE (Ersman 1987:39).
Mobile telecommunication was pulled along with the overall Ericsson strategy to
sell on the American market. One of their incentives to try was that the cellular market
was increasing so fast so there was lots of money to be made. According to Lundquist
(1997) the potential losses in case of a failure in the USA were small. In order to
deregulate cellular networks, the American Federal Communications Commission
(FCC) allowed companies to apply for licenses to build cellular networks in given
geographical areas. The first filing of, or bidding for, licenses covered the 30 largest
cities in the USA and took place in June 1982. Although then a minor actor in the
American market, Ericsson decided to give it a try by being supplier to the network
operators which were bidding. Out of 140 applications submitted to FCC, forty were
specified with Ericsson equipment (Meurling and Jeans 1994:66). SRA and BX jointly
put together proposals for the various network operators which received licenses and
which had specified Ericsson equipment. In May 1983, Ericsson signed their first
contract for a cellular phone infrastructure in Buffalo, New York (Meurling and Jeans
1994:77). This was an important order, a foot inside the door.
In Buffalo, the AXE switch was used, which led to some problems in introducing a
new digital switch in the American context. The American standards were quite
different from the European one and therefore extensive investments were demanded in
42
R&D to adapt AXE.23 In order to handle the increased work load and to meet
specifications for cellular as well as fixed switches, Ericsson increased its US
organization and transferred a substantial amount of personnel. Moreover, co-operation
between BR and BX in Sweden was intensified. On April the 2nd 1984 the Ericsson
infrastructure was functioning in Buffalo, and Ericsson gradually became established as
a system supplier in networks installed in the USA (Meurling and Jeans 1995:179).
A vital difference between the American market and Ericsson’s previous experience
with NMT was that NMT standards provided given technical specifications while the
American AMPS instead had open interfaces and so specifications for interactions had
to be designed by producer firm(s). For example, there was no or only limited
possibilities to mix a RBS from Motorola and a switch from Ericsson unless extensive
work was done to get them to communicate. Actually, in stark contrast to the argument
that technical specifications set by governments are completely positive for producer
companies, this open interface is one of the major explanations for Ericsson’s success
on the American market. Moreover, using the AXE switch turned out to be an
advantage compared to competitors’ products (Persson 1997). When demand of
subscribers and use far outstripped the projected demand, AXE had the technical
capability to handle the load.
8.2 European Markets
Telecommunication deregulation in Europe in the early 1980s similarly opened
those markets to foreign companies, and once again, the AXE switch was key for
Ericsson’s successful bidding for important orders.
Vodafone, a private British fixed network operator, wanted a mobile network. As it
did not have its own technical competencies in that area, Vodafone did not just want the
system delivered but instead also wanted the supplier to install the whole network and
get it functioning. The competitors for the Vodafone order in the last round were
Motorola, AT&T and Ericsson.
Ericsson got the order mainly because they could supply the AXE switch to
Vodafone (Meurling and Jeans 1994:92). Vodafone wanted a powerful switch,
especially for the larger cities, and Ericsson’s AXE-switch was far more powerful than
those of AT&T and Motorola, which had developed smaller switches more specialized
to a perception of limited mobile phone use. Moreover the AXE switch could cope with
Vodafone's requirement for roaming. The contract with Vodafone was important for
Ericsson because Ericsson had been a minor actor on the British telecom scene but now
moved up to being a mid-size one. Ericsson (ERA) installed the system, and completed
it on March 15, 1985. Moreover, with this order, Ericsson became a supplier of the
23The first AXE contract for the public telecommunication in USA was signed in 1986.
43
three standards contemporary in use internationally, namely NMT, AMPS and TACS
(Meurling and Jeans 1994, 1995).
As perhaps in many countries, Swedish deregulation of cellular networks did not go
smoothly.24 Televerket’s monopoly position as operator of NMT started to be
challenged in 1981 by a second operator, Comvik.25 Comvik had provided mobile
communication services previously, and wanted to move into NMT but Televerket
opposed this. Ericsson’s CEO Svedberg supported Televerket by arguing that only a
publicly funded network would give full geographical coverage because private
operators would only concentrate on the cities, the profitable areas. Svedberg argued
that rapid expansion of the NMT (450) network was important for Ericsson’s future
international success, as they wanted to show a functioning, large scale system.
Comvik, however, won the right to be a NMT 450 network operator, and although they
never became a serious contender to Televerket there, they later emerged as one of three
Swedish operators given permission to establish GSM systems, along with Europolitan,
in 1991.26
Although jumping ahead of our story a bit, the close connection between Ericsson
and Televerket against a competitor is substantiated in a later case as well. When
Comvik offered tenders for bids for the GSM system in 1989, Ericsson again sided with
Televerket.27 The then new CEO, Lars Ramqvist, explained that they could not supply
GSM equipment because AXE had been developed with Televerket in Ellemtel, and
Televerket had a monopoly to market AXE switches in Sweden. Televerket replied that
there was no formal restriction for GSM systems, but that competition would be
improved if Comvik chose another supplier. Comvik finally chose a GSM infrastructure
with components from Motorola, Digital Microwave Corporation, and Siemens. Thus,
the long standing cooperation of Ericsson and Televerket also affected much later
competition over provision of infrastructures to operators.
8.3 Nordic NMT 900 and Other Firms
In 1985, bids were given for the NMT 900 infrastructure. The Nordic PTTs had
decided to develop the intermediary analogue standard NMT 900 between NMT 450
and a digital system because of the saturation of users in NMT 450. They wanted to
know what firms would be capable of producing, so that a discussion had ended in the
publication of the so-called Green Book of Mobile Telecommunication, where the
characteristics of the new NMT 900 system were specified.
24 Paragraph based on Karlsson forthcoming, chapter on ’Infrastructure Competition’.25Comvik is part of a larger concern, Kinnevik, which also distributes TV channels.26Two other operators also tried to establish GSM networks but their requests were denied. One was
the British company Racal Vodafone, and the other was NordicTel PCN. NordicTel was owned by,among others, Volvo and SAS, and managed by two former Ericsson Radio Systems employees.
27Paragraph based on Karlsson (forthcoming).
44
Both Åke Lundquist and Lars Ramqvist had tried to convince Televerket to buy the
American system (AMPS) or the English system (TACS), since Ericsson had already
developed the elements of the system. However, Televerket refused and the proposals
were eventually given up (Mölleryd 1996:45).
The Green Book resulted in a call to the main actors to offer products. The suppliers
of RBSs were: Finland — Nokia; Denmark — Philips; Sweden — Magnetic (owned by
Ericsson), Nokia, Radiosystem and Mitsubushi; and Norway — Radiosystem
(Kågström 1997). All bought switches from Ericsson (Hultén and Mölleryd 1993:4).
Radiosystem thus received a large part of the Nordic market for RBSs, and
especially the Norwegian market, which was considered as the ’impossible order’
(Kågström 1997). Their success in winning the orders not only put a strain on the
capacities of the small company, then employing 43 persons (Radiosystem 1985:12),
but also placed them as a serious competitor to Ericsson Radio (previously SRA). The
previous collaboration between the two firms stopped. When Ericsson Radio found out
that Radiosystem was putting in a competing offer, they decided to change suppliers,
also partly because Ericsson had bought Magnetic to develop this product. Another
reason might have been that Ericsson did not trust the capacity of Radiosystem to be
able to produce both RBS and the other products (Kågström 1997). Radiosystem still
lacked competencies in the domain of data processing and control units. In order to gain
access to those competencies, two consultants were hired by the firm to develop the
technologies.
Also in 1985, Radiosystem went public on the Stockholm stock exchange. When
they did so, Radiosystem had about 20 customers, including not only Nordic PTTs but
also firms in the business of radio communication. The three largest customers were
SRA, Mobira (Nokia) and Philips, and each accounted of 25% each of the sales of
Radiosystem (Radiosystem 1985:8).
Actually, the relative market success of Radiosystem is interesting because despite
the transfer of technical competencies with the founders, the main goal of the firm was
to develop and produce as fast as possible for a quickly increasing market. Radiosystem
was to mainly produce components, and no time or budget were assigned to explicit
formal research. Instead, the main way of getting new information was through
international fairs and informal contacts with other engineers from other firms.28
Another important way of getting new competencies was through recruiting personal
from other firms in the business of radio communication. The recruitment policy was
that ’engineers or radio-amateurs’ were welcome, as a recruitment advertisement
28The patenting policy of the firm was also very simple: there was no patenting at all. Here again the
same reason as for the research policy prevail. The firm was small and all the personel was busy with thedevelopment of new products or production. Another reason was the cost of patenting. The main goal ofthe firm was to produce in order to survive (Kågström 1997).
45
suggested (Kågström 1997). Radiosystem was thus turning into a serious competitor to
Ericsson for RBS in the Nordic market, despite its small size.
Through the mid-1980s, Ericsson also continued improving their mobile phones for
the NMT standard, introducing new ones for both the 450 and 900 standards in
1986/87. The first ’Hot Line’ product family was introduced, which were half as large
as the previous ones. Interestingly, Ericsson divided their product market into segments,
implicitly based on a ratio between price on the one hand and size and weight on the
other. They divided the products into ’mobile radio systems’ (for vehicles), ’portable’
(the same size but can be taken along), and ’pocket phone’.29
Thus, in the early to mid-1980s, Ericsson was increasingly successful in moving
into European and American markets for mobile infrastructures. Deregulation was an
important factor, enabling them to compete for PTT bids with national firms for each
country as well as for new, private operators. Ericsson initially sided with Televerket in
the Swedish case, as they wanted rapid expansion of the NMT systems. For Ericsson,
moving abroad also meant moving away from the NMT standards and thus necessitated
in-house competence to deal with the new technologies.
In-house did not, however, mean R&D in Sweden, as Ericsson opened new centers
internationally both to be close to markets and to access competent persons. Although
internationally they were selling systems, at home ERA and the Ericsson owned
Magnetic were facing competition from Radiosystem, as exemplified in the Nordic
NMT 900 orders. In phones, Ericsson was introducing new product families were price
was related to size and weight.
9. The Shift to Digital Standards—Late 1980s to 1990s
The shift from analogue to digital standards has been a major technological
discontinuity, not only because much of the core knowledge is different but also
because that knowledge has been under rapid development (see section 2). This shift to
digital has occurred internationally in the late 1980s and 1990s, but different digital
standards exist, with three major groupings being those developed and accepted in the
USA, Europe and Japan. Future standards are to be more designed for computer
communication than voice communication, and they are also under negotiation, with
divergent views on which technologies solutions should be chosen (America’s Network
1996).
The move to digital has consolidated Ericsson’s position as the main Swedish
competitor in the industry, but it has also been a period where actors other than firms
have actively developed Swedish competencies in relevant knowledge areas.
29The most important technical developments to reduce size and weight so rapidly were
miniaturization of components, new production methods, and surface mounting (Jimalm and Rydbeck1987:141, 146).
46
Nonetheless, Ericsson invested heavily to develop their own in-house competencies
both in Sweden and internationally and feel they rely heavily on their own resources
and knowledge.
9.1 Sweden and GSM Standard, 1987/88
The GSM standard is the European standard for digital mobile telecommunication
which was decided in 1987/88. It was launched in 1991, and by September 1996, GSM
had more than 21 million subscribers in 133 networks operating in more than 70
countries, with 50,000 subscribers signing on each day’ (America’s Network 1996).
GSM was an outcome of a working group for digital standard of the Conference on
European Postal and Telecommunication Administrations (CEPT), which started
working in the early 1980s and made decisions in 1987/88. The specifications of the
GSM standard were contained in more than five thousand pages.
Within Sweden, discussions among interested actors were also on-going. Already at
the end of the 1970s and beginning of the 1980s, a working group was established in
order to discuss future developments in technologies for mobile telecommunication.
The aim of this group was to be a reference group for university research regarding
digital standards, and included different actors from universities (especially Lund and
Chalmers Institutes of Technology), the military (the military research body, FOA),
Ericsson, and Televerket (later Telia). There has been exchange of competence within
the Nordic countries and especially between Sweden and Finland, and more recent
reference groups have included Ericsson, Nokia, Telia and Nutek (Öhrvik 1997).
According to Östen Mäkitalo (1997), head of the research department at Televerket, the
role of the military—through their research branch FOA— was limited, and decreased
over time.30 Other than firms, Televerket was again a driving actor. Televerket did
research concerning the narrow band TDMA-system, and in 1985 Ericsson joined this
research. Televerket also co-operated with universities to a large extent when they
developed their new TDMA-system (called 8-PSK) (Mäkitalo 1997). Ericsson was also
developing its own competencies (see section 9.3).
A relevant question to ask in this context is, Why was Televerket a system-builder
in the mobile telecommunication business when they were not set to produce equipment
for the system? After all, although Televerket had played a crucial role in the NMT
standard and benefited from being an operator, neither they nor the other Nordic PTTs
benefited from the international use of that standard, which can lead us to ask why they
wanted to develop the new standards. According to Mäkitalo (1997), Televerket saw a
30ERA and the military (FOA) co-operated for some time before this, but the competence of FOA
declined when they moved from Stockholm to Linköping. Radar technology has been important formobile telecom system because the military initially had relevant competence in array antennas (digitalsignal processing).
47
future market as an operator, but felt that if they would not develop the system no one
else would either. Televerket’s research department (Radiolaboratorium) had strong
support from the top management, because they thought that mobile
telecommunications would be a major business in the future. This belief did not exist in
PTTs in other countries.
In connection with asking why Televerket has been interested in developing
standards, we can also ask how extensive their investment has been for different types
of standards. About 10-20 persons worked with the specification of the NMT 450
system in Televerket radiolaboratorium during a ten year period. For GSM, there were
approximately 100 persons during a similar period. The development costs of the GSM-
standard for Televerket were slightly less than 1 billion Swedish crowns (Mäkitalo
1997).31
In 1986, CEPT tested eight competing alternatives in Paris to decide upon a new
digital European standard. Of the eight alternatives, there were four Nordic ones while
the other were German, French and Dutch. The Continental alternatives were broadband
solutions, which was the logical technical solution for large countries, while the Nordic
ones were based on narrow broadband solutions, more adapted to the geography of the
Nordic countries (Lindmark 1995:111). The alternatives from the Nordic countries were
from Ericsson, Televerket, Nokia and Elab (a Norwegian company) (Mäkitalo 1997),
and the four continental European alternatives were from consortia.32 Thus, Televerket
was the only PTT offering an alternative; all the rest were firms.
Just before the test, Televerket realized that their alternative would not be ready.
They therefore spread the information they had to Nokia and Ericsson as well and
started all over with a new solution. In fact, Televerket developed their 8-PSK system in
just 8 weeks before the Paris test, but they did not acknowledge that publicly. They
were afraid that they would then not have been considered as a serious contestant
(Mäkitalo 1997). The purpose of the information exchange between Nordic actors was
to stop the Continental broad-band solutions. It did not matter to Televerket which of
the Nordic alternatives won, as long as one of them won. If a broadband solution would
have been chosen as the new European digital standard, the Nordic actors would had
been left way behind in the technological development. Televerket’s solution, together
with the Norwegian one, performed very well in the test (Östen Mäkitalo 1997).
31Today there are about 20-30 people working on the next generation of mobile telecommunication
standards, but Telia does not have the competence nor the strength to lead the development of the nextstandard. However Telia management thinks that they could mobilize that strength and competence ifnecessary (Mäkitalo 1997). Of importance for future developments in Sweden is the semi-independentresearch foundation, ’Strategic Foundation’ (Strategiska Stiftelsen), of which the research program’Personal computing and communication’ supports relevant university research. Within this program, 26Ph.D. students are financed at Lund, Chalmers and KTH from 1997.
32The four consortia were: 1) ART, SAT, SEL, AEG, Italtel. 2)Philips and TRT. 3) LCT and 4)ANT, Bosch and Teletra (Lindmark 1995:111).
48
Overall, the tests indicated that the Nordic group of TDMA technology was superior,
although there were of course many discussions and disagreements.
In 1987 in Madeira, the European body decided to build a new European digital
standard based on TDMA, called GSM.33 Although the French and German companies
reserved against the decision, they did not vote against it (Mäkitalo 1997). The decision
was to their disadvantage because they had bet on a different technique, and were now
behind their Nordic competitors. In fact, for technical reasons, it turned out that the
GSM-specifications made pocket phones (rather than car phones) to be a real possibility
much quicker.
The signaling of a European change to digital standards in 1986 also affected
Radiosystem, the then last remaining small Swedish, competing firm. This in turn
affected in-house competencies at Ericsson. With the GSM decision, Radiosystem
became skeptical about its own future because they would either have to provide a
complete infrastructure or limit production to analogue technology for new, less
technically advanced markets. Their competence in computer programming was very
limited, which would make any attempts to design complete RBS for GSM very
hazardous indeed in terms of functionality and reliability. However some research
projects were started in order to prepare the transition from analogue to digital.34
Radiosystem's problems with GSM had an effect on Ericsson in that although
cooperation had been stopped in 1985, new discussions were started at the beginning of
1988. As a result, the negotiations led to an offer from Ericsson to buy Radiosystem at a
price three times as high as sales (Affärsvärlden 1988b:27).35 In 1989, Ericsson bought
Radiosystem, but did not integrated it with Magnetic, which Ericsson had bought in
1983. The name was later changed to Ericsson Radio Access AB in 1992.
Several hypothesis can be given for the acquisition of Radiosystem by Ericsson.
* Magnetic was not performing as expected in the NMT 900 sector and Ericsson
needed to get stronger in this field (Kågström 1997).
* Buying Radiosystem would suppress a national, and Nordic, competitor for
Ericsson. Radiosystem then had almost half the market of radio base station for the
NMT 900 market, with the rest going to Nokia. Magnetic, the second Ericsson flagship
for RBS after SRA, had succeeded to sell only in the Faeroe Islands and in the south of
Sweden. As Åke Lundqvist (1997) put it, ’we needed to....tidy up the market’.
33GSM is based on a new switching platform, which is open and so allows different components to
be connected together in a variety of ways (Emmett 1996:47).34The budget for the research was of 5 million crowns in 1987-88 and the team consisted in 6
engineers (the turnover of the firm at that time was 150 million crowns).35The purchase was for 465 million crowns, at a time when Radiosystem's own capital was 90-100
million Swedish crowns; had 147 employees; sales of 147 million Swedish crowns; and profits of 20million Swedish crowns.
49
* Other firms were interested in buying Radiosystem. This might explain the high
cost of the purchase, relative to estimated value (Kågström 1997, Meurling 1997).
According to Meurling (1997), Motorola was interested by Radiosystem. However,
Kågström (1997), then engineer at Radiosystem, doesn’t recall this but argues that
Nokia started to be interested after they got to know that Ericsson was interested. The
reason was that Nokia was purchasing some parts of its RBS from Radiosystem and
therefore were probably not eager to see Radiosystem in the hands of Ericsson, their
direct competitor.
* Ericsson wanted to access the competencies of the few key persons behind
Radiosystem's short, but very successful, existence (Affärsvärlden 1988).
After the purchase of Radiosystem by Ericsson in 1988, a team of engineers was set
up between Ericsson and Radiosystem to develop competencies in digital
communication. This team of about 60 persons allowed an in-house knowledge
acquisition on digital technology while the rest of Radiosystem of more than 140
persons continued developing and producing radio base station for its previous markets
and tried to find new applications for the existing technologies (Kågström 1997).36
However, the entrepreneur which had started Radiosystem, Torbjörn Johnson, then
went on to start another firm, Radiodesign, in 1994.
The first major orders for GSM were in areas with relatively high populations, like
Germany, the United Kingdom and France. The largest was the D2 network operated by
Mannesmann Mobilfunk in Germany. In fact, of the initial 14 European networks
ordered by 1992, Ericsson had won orders to supply equipment in ten of them
(Communications International 1992:30). See further section 9.3.
9.2 Swedish Competence Build-up
Here we will discuss the development of competence in Sweden which lead to the
building up of trained persons, knowledge and thereby the ability to solve additional
problems. There were different forms of relationships between Televerket (Telia),
Ericsson, Nutek, and the universities / institutes of technology.
Work directly on digital standards was mainly united as part, albeit a small part, of
the Swedish national initiatives for Information Technology, the so-called IT4 Program.
Telecommunication generally was budgeted to be around a quarter of the total. The
overriding goal of IT4 was to develop technologies which would strengthen Sweden’s
economic competitiveness. IT4 started in 1987 and ran up to 1990, and had a total
financing of about $182 million USD. Work was to be financed 50% by the state and
50% by industry; and involved financing from government agencies, specifically the
36An example of a new market for existing technology is the Radio in a Local Loop, where analogue
radio system is used to provide a public network, especially in developing countries, lackingtelecommunication infrastructures.
50
military (FMV), STU (National Board of Technical Development, later merged in
Nutek, National Board of Industrial and Technical Development), and Televerket.37 IT4
included both national research or R&D projects as well as participation in EU-projects
like RACE for broadband technology. Over time, universities were increasingly drawn
into these projects.
In one IT4-project, Televerket and Ericsson together built a prototype GSM-system.
The first phase started already in 1988 to test NMT 900 and GSM; the second went on
to build a test system; and then the third phase included development of critical
components, mainly at Ericsson Mobile Communications in Lund (IT4 1991b). By
1989, Ericsson could simulate a whole GSM system, based on computer aided
engineering tools, at their R&D center in Lund. That they could simulate and test a
system was, according to one source, ’one of the reasons why the Swedish system
proposals from Ericsson and Televerket (which were very similar) ended up being the
foundation of the European GSM system’ (Lilliesköld 1989:35). This IT4 test system
together with the extensive TDMA research in Sweden gave Ericsson an advantage in
accessing competencies in Sweden, according to Televerket’s research director
(Mäkitalo 1997).
As digital standards were a technological discontinuity requiring significant new
knowledge and techniques to function in practice, university research in related fields
was sometimes important. General Swedish research in areas relevant for the new
digital, mobile telecommunication in the late 1980s was funded through government
research councils. University research was important to develop knowledge relevant for
some parts of the GSM-system; for instance, Lund and Chalmers Institutes of
Technology were the main forces regarding modulation methods (Mäkitalo 1997).
Moreover, Televerket and Ericsson put resources in the same field during the 1980s
(Mäkitalo 1997, Örhvik 1997). Televerket has also tried to influence the direction of
university education by participation in board of faculties (‘fakultetsnämnder’) and
(‘institutionsstyrelser’), and moreover Televerket has supported university institutions
and financed professors in specific areas (Mäkitalo 1997). Ericsson has had similar
influence at relevant universities.
SRA/ERA has similarly influenced academic basic training at the Bachelor of
Science level, and although the extent of wielding influence has remained stable since
the 1970s, Ericsson has become more efficient in using it. As another example of
influence, SRA (later ERA) helped universities build test equipment, where known
technologies were used. There are still today a very close form of co-operation between
ERA and the Swedish universities (Öhrvik 1997). Co-operation between the industry
and university research can also be found in Nutek programs, also financed 50-50 with
37See (Arnold and Guy 1989), (Technopolis and SPRU 1992), (IT4 1990, 1991).
51
industry. For example, between 1993 and 1996, Nutek had a Research Program on
Telecommunication and Service, but which was only about $7.4 million USD over
three years (NUTEK 1997).
Ericsson also placed its R&D center for mobile telephones in Lund, across the street
from the Institute of Technology. This has been very important for its success with
phones (Skidé 1994, Lundqvist 1997).
Moreover, the actors have had some direct research cooperation relevant for digital
cellular communication. To be useful, it seems important that the research co-operation
between industry and university be very close and integrated. For example, Professor
Örvik (1997) from Lund Institute of Technology (later at Ericsson) argues that there
were much valuable transfer of knowledge between SRA and the universities through
close cooperation on Ph.D. projects. The basic problem, however, is to balance the long
term research of universities with the building up of knowledge areas more directly
relevant for research by firms.
In summary, Sweden as a country has built up a competence in knowledge and
techniques relevant for digital technologies. In addition to Ericsson’s own in-house
activities, Televerket has traditionally given strong support and had technical
competence as well as government funding of industrially-relevant research and
research cooperation between universities and firms. Building up such competencies
seems to be particularly important for the recruitment of engineers, which has been a
critical need for Ericsson during the 1990s at their fast pace of expansion. Currently,
Ericsson hires almost 900 Master of Science engineers per year in Sweden, which is
about the same number as graduate with a MS in electronics or computer science (Ny
Teknik 1997a). That is not, however, the same thing as saying that the Swedish
competence base was necessary for Ericsson to succeed in digital communications.
Ericsson has become increasingly a truly multinational company.
9.3 Ericsson Gains Competencies: Collaboration and Doing it Alone
Although GSM had been decided as a European standard by the end of the 1980s,
the USA and Japan each decided to develop there own digital standard. The problem for
Ericsson was firstly, whether to pursue all three major standards and secondly, how to
find resources to do so. The task fell to Lars Ramqvist, who became president and CEO
of Ericsson in 1990. Ramqvist’s had previously been head of Ericsson Radio System
and seen the explosive growth of mobile communications, at about 40% per year (Skidé
1994:13). His strategy was to pursue it and to go for all the international markets and
hence develop equipment for the three major standards:
’I had to explain to share-holders that I was proposing to increase R&D costs veryconsiderably, possibly by as much as 50% a year over two years,’ Ramqvist says. ’Wewould be spending 15 times as much on R&D as on dividends at a time when recessionhad hit the industry and we were faced with falling profits.’....
52
Ramqvist’s first priority was to focus his efforts. ’We had to realize that we could not doeverything,’ he says. ’Five or 10 years ago, Ericsson even produced its own nuts andbolts. Now we were going to spend more money on development than anyone else in theindustry, but we had to concentrate that effort on our primary goals in telecoms systems,and in particular on mobile communications.’We also needed partners in areas outside our core competence. So we establishedalliances with Texas Instruments in microelectronics, General Electric in a marketingcompany in the US, and Matra in France.’ (International Management 1994:27).
Thus, Ericsson’s new strategy to be a major player in mobile telecommunication in
the 1990s was both to dramatically increase in-house R&D, refocus activities, and also
develop alliances to gain different types of competencies or access to marketing.
For example, in the French case of developing a GSM network in 1987, the French
government chose AXE-switches from Ericsson, and radio base stations from Matra-
Ericsson Telecommunications SA or MET. This collaboration took the form of a jointly
owned company MET, which sold systems with AXE switches from Ericsson and radio
base stations from Matra. The co-operation mainly concentrated on the integration of
components. However, as in the Siemens case, there were disagreements over which
radio base station controller should be chosen. Ericsson had the most powerful BSC
suitable for high-traffic areas and thereby the Ericsson’s BSC was chosen in most areas.
MET had only limited success in the French market, and in 1992 their co-operation was
terminated (Meurling and Jeans 1994:124). Moving in sometimes required cooperation
with competing companies, but most were not long lasting.
One particularly important GSM order for Ericsson was in Germany, where they
had had a longer term strategy to enter. Before telecommunication deregulation,
however, this was a closed market, dominated by very large companies. In Germany,
Siemens and SEL (Standard Elektrik Lorenz) had long dominated the public
telecommunication business and also for mobile telephony. Ericsson tried to get a foot
into the market by cooperating with Siemens in 1986 to develop base station controllers
(BSC) for the GSM mobile telephone standard. One reason the collaboration stopped
was that Ericsson’s work in the area led to the discovery of the benefits of using AXE
as a BSC in the GSM-system, i.e. to use the powerful AXE-logic to control a number of
radio base stations. Siemens was not fond of the idea of giving over more power to—
and hence become dependent upon—an Ericsson component (Meurling and Jeans
1994:117).38
Ericsson then succeeded to sign a major contract in Germany in 1990 to provide a
GSM system. The network operator, Mannesmann, was impressed with what the
Ericsson system had already accomplished in the USA. Of particular importance was
that Ericsson could show that they had switches which could manage high mobile
38This development work was done under the leadership of Håkan Jansson, who became head of the
Mobile Telephone Division at ERA in 1987. Jansson came from BX and one of his duties was tocontinue to transfer AXE knowledge from BX to ERA.
53
telephone traffic in a rapid growing market (Meurling and Jeans 1994:120). As this was
the largest of the initial European networks, it was a very important order to prove
Ericsson’s products.
Thus, during the 1990s, Ericsson has followed Ramqvist’s three-prong strategy of
high R&D costs, focusing on only keeping core activities in-house, and alliances and
collaboration, particularly with firms having other technological and market
competencies like microelectronics, software, etc. This has meant pursuing the major
standards, continuing to be a mobile network provider, but also moving more strongly
into mobile phones. The data given in section 3 indicates Ericsson’s success.
9.4 Since Then and Forward
It is useful to compare Ericsson with one of their competitors, Motorola. Motorola
was seen as a strong company coming into the 1990s, lead by its CEO George Fisher.
Motorola was like a gigantic SRA, with competence in radio base stations but without
particular competence in switching. Their switches were way behind Ericsson’s in
performance. In fact, by 1992, Motorola had stopped producing switches but were
instead cooperating with Northern Telecom, Siemens and NEC. Fisher was skeptical of
Ramqvist’s strategies for the 1990s: Ericsson was spending too much money without
demanding enough profitability; couldn’t continue to do everything themselves, etc.
(Ahlbom 1992). This was at a time when Motorola was losing American market shares
in network systems while Ericsson was gaining; then, however, Motorola had a high
market share in phones. Since then, however, Motorola has also been losing market
share in phones, while Ericsson has gained enormously.
Lundqvist (1997), among others at Ericsson, argues that having the whole system
in-house has been crucial for success. For example, there have been movements of
individuals and thereby competencies from BX to BR. SRA could not have been
successful if these competence movements had not taken place. It was important that
Ericsson had most of the competencies and knowledge concerning switches in-house
(Persson 1997). Thus, it has both been important that Ericsson had core competencies in
all three major technologies areas but also in order to integrate a system and identify
problems and opportunities, it has been important that competencies have been spread
within the company. As seen in the area of switches, Ericsson mainly relied on in-house
R&D, with joint R&D being important. Ericsson employees often stressed in interviews
the importance of relying on in-house competencies and solutions
Whether Ericsson will be able to retain its position are unclear. There are two
difficult challenges ahead, namely future techniques and standards and the integration
of telecommunication with computer communications and IT more generally.
Firstly, mobile telecommunications itself is a fast moving area, and it is not certain
that companies which are successful today will be able to keep up with the very rapid
54
technical and market developments. As in the past, current standards will soon be
challenged by potential overload in the number of subscribers, warranting new
bandwidths, standards and technical solutions to handle the load. Moreover, new
techniques for signal processing are currently being developed, where Ericsson has
strongly supported one side and has worked with Japanese actors on it while many
American companies have bet on a different one (American’s Network 1996).
Moreover, there are a couple of trends which could change the whole mobile
telephony infrastructure industry as we discuss it here. Firstly, there are discussions and
preliminary attempts to build a worldwide communication network based on satellite.
One advantage would be that the same phone and/or modem would work anywhere in
the world. Secondly, telecommunications is really becoming telecommunication and
computer communications. Provision of information services and data communication
by TV networks, electricity companies, computer software companies etc. is
challenging not only the traditional providers (operators) of telecommunication services
but also the producing firms, as they must meet the needs of new types of users, with
new demands (McKelvey forthcoming). This synthesis telecommunication and
computer communication and the real threat to traditional telecommunication firms is
clearly illustrated by Microsoft’s move into building worldwide communication
networks for both telecommunication and computer communications.
10. Conclusions
The history told here of Swedish actors in the mobile telecommunication system
industry tells of firms in a dynamic relationships to others, with both national Swedish
and international contacts in the SI. It involves private firms and state actors; small
firms and large firms; standards bodies and universities; competing firms and
competing technical standards; the high end of the market consumers as well as mass
markets; local, national and international contacts. The obvious success story in terms
of sales, employee expansion and profitability has clearly been Ericsson, which has also
bought up the small firms. Moreover, their AXE-based infrastructure system has been
adapted to support all the major analogue standards (NMT, E-TACS and AMPS) and all
the major digital standards (GSM, D-AMPS, North American ADC and Japanese JDC)
for mobile telephony.
The two questions to be answered in this section are: 1) Were diversifying firms or
new innovators involved? and 2) How and why were Swedish firms were able to
identify and act upon innovation opportunities, in situations of rapidly changing
technologies?
55
10.1 Diversification vs. New Innovators
The first question about diversification versus new innovators will be answered in
relation to the three component technologies—switches, radio base stations, and cellular
phones.39
Switches
Ericsson has been the firm on the Swedish scene, but its move into switches for
cellular networks cannot be classified as diversification. The technology was a
modification of the then recently developed computer controlled AXE switch, and
initially the buyers were similar, namely PTTs. AXE was the result of joint R&D
collaboration with the Swedish PTT Televerket, in the dedicated R&D firm Ellemtel.
Ericsson had long experience with switches for public telecommunication networks,
and mobile networks were seen as one small special application area of existing
knowledge and products.
Because Ericsson had never had a large and secure home market, Ericsson had to
develop a switch that easily could be changed to meet different specifications and
needs. AXE was designed for flexibility, and had to function as a local exchange,
tandem exchange, rural exchange and international exchange. Just by replacing
hardware and software but using the same basic architecture, Ericsson could adapt and
hence sell AXE for many areas, instead of having a dedicated switch for each. In the
same way, AXE for mobile phone networks could be fairly easily adapted for different
standards and specifications.
In addition to flexibility, AXE is powerful in terms of computer programming and
handling data, and with it, Ericsson was in fact one of the technical pioneers into digital
switching (generally), although they had previously been behind competitors. This also
turned out to be very important for winning later contracts for mobile networks, as it
allowed them to offer desirable features like roaming and hand-over as well as handle
large numbers of calls.40 Note, however, that initial development of this software was
seen as a small niche market, and pushed by SRA, which was developing radio base
stations. The internal Ericsson division on switches, BX, kept its visions focused on
public telecommunication networks for a long time and laughed at the idea of Ericsson
becoming a mobile telecommunication company.
39A summary of the major actors involved in technical development can be found in Appendix A.40The AXE software could handle large volumes of data by using digital signaling. In a mobile
system, the hand-over function requires much data processing when an AXE switch has to decide fromwhich radio base station the telephone is receiving the strongest signal. If the telephone is in motion, thenthe AXE switch must be able to redirect communication links between phone and another radio basestation, where the signal is stronger. For similar reasons, roaming to locate a mobile phone outside thearea to which the handset has subscribed. The NMT standard and installed networks allows roaming innorthern Europe while the digital GSM-system allows roaming in Europe and in the forty countriesoutside Europe that have chosen the GSM-system.
56
When expected mobile phone use were predicted to be quite low, using AXE
seemed like a disadvantage because it was costly and large (up to 60,000 subscribers)
compared to competitors’ products. This apparent disadvantage proved, over time, to be
an advantage because as mobile telecommunication has grown, powerful switches have
proven necessary to handle services and the expansive growth in subscribers. The PTTs
demanded it fairly quickly.
As Ericsson had never had a large and secure home market, it had to compete on the
international scene. Not only is Sweden a small country in terms of population although
large geographically, the Swedish PTT had throughout its history manufactured most of
its own switching equipment.41 It could not rely on the strategy of some its competitor
companies, namely being coddled national champions propped up by state subsidies.
This meant that when a wave of telecommunication deregulation came in the 1980s,
Ericsson was used to competitiveness conditions and selling internationally, albeit most
successfully in countries like Asia and South America (Meurling and Jeans 1994, Ny
Teknik 1997a).
The buyers were initially essentially the same for mobile as for fixed switches,
namely national PTTs. The wave of deregulation starting in the 1980s not only opened
up previously closed monopoly markets in the USA and Europe, it also meant that other
private network operators also bought systems to operate. Moreover, competition has
become more and more on price, and the buyers have had less and less own in-house
technical competencies, meaning they rely more and more on supplier firms like
Ericsson.
Ericsson’s world market share of switches for the cellular mobile telephone market
is now 40%. The same figure concerning the GSM-system and Europe is 50% (Ericsson
1997a). Before 1990 the total amount of sales of public switches was bigger than the
correspondent amount for mobile switches. After 1990 however, the mobile switches
have been responsible for a bigger share of the sales of switches. Moreover, the
profitability of mobile switches has been higher than public switches due to the fact that
there has not been as much competition in the market for mobile switches as it is in the
market for public switches (Lundqvist 1997). Switches are more a development of an
existing product for a new application area.
Radio Base Stations and Cellular Telephones
Radio base stations and cellular telephones both rely on radio communication.
Ericsson can be said to have diversified in these two product areas, which are based on
similar technologies. They moved into products based on new technologies and with
new markets. Their move into the products was through SRA. Most of the small
41Thus, Televerket continued to manufacture even after Ericsson and Televerket developed the AXE
switch in Ellemtel.
57
Swedish firms—namely SRA, Sonab, and Magnetic—moved from military to civilian
applications in radio communication, although Magnetic had worked on TV
transmitters before cellular networks. Although the basic knowledge area is the same,
cellular networks are new products, based on further developments of knowledge and
technology. There have also been two new innovators, namely Radiosystem and
Radiodesign, where both were started as spin-offs from other companies (by the same
entrepreneur).
Firms which had the initial competencies for radio communication had it because of
contacts with the Swedish military. In fact, even some of the initial SRA equipment
(RBS) for the first USA order were directly taken from military products (Lundquist
1997). The Swedish military has also had its own competencies in relevant technical
areas, but the importance of the military has declined over time. Today the military has
a very limited role when it comes to having an impact at the development of the mobile
telecommunication technology (Öhrvik 1997).
In addition to military uses, these small firms were also involved in the small scale
systems for civilian use, which included providing both RBS and phones. Moreover,
small firms could provide RBS or components of RBS for the initial Nordic standard
NMT because the technical specifications and interfaces were defined in such detailed.
The specified standard allowed them to concentrate on parts of the infrastructure rather
than providing the whole thing. To compete with later standards internationally, they
would have to cooperate more directly with other firms to provide a whole
infrastructure or else fight on the consumer market, phones.
Over time, the small Swedish companies in these areas have all been bought up by
Ericsson. SRA was from the beginning partly owned by Ericsson but independent, and
this independence, along with an explicit strategy to turn to civilian products, allowed
room to pursue mobile telecommunications at a time when few believed in its future
potentials. When Ericsson wanted to expand its in-house competencies, it first bought
the rest of SRA and then also bought the other Swedish firms. Moreover, Ericsson
expanded rapidly internationally, both in Sweden and abroad and developed
international collaboration with other firms as seen necessary.
Although Ericsson had previously made phones for fixed networks, the two types of
phones are based on different types of technologies, and users have different types of
demand. Over time, competition in markets for cellular phones has become more and
more over price and less over technical finesses. Or rather, in the ratio of price to
performance, it seems that when cellular phones become a mass market, many new
users chose price rather than performance. Sales of phones, and hence expansion of the
number of subscribers, are also influenced by operators of networks, particularly their
tariffs and whether/how much they subsidize purchase of the phones. Ericsson’s work
58
on phones seems to be mainly based on in-house Ericsson R&D, although with careful
monitoring and access to external technical developments as needed.
In summary, Ericsson’s move into the two products of RBS and phones as based on
radio communication is a case of diversification. It is a case where the firm moved into
this area, expecting it to be a small percentage of sales, not the dominant business
within a few years. The other small firms either diversified from military to civilian
products or else were started to take advantage of these new innovation opportunities.
System integration
Ericsson’s move into becoming an infrastructure system provider for cellular
networks can be classified as diversification because the technologies were new to them
(and often, to the world), markets differed, and internal organization change was
required. This is so even though the linchpin technology for the diversification has been
the AXE switch, and there, Ericsson adapted a new, yet existing, product to provide
switches for cellular network. Ericsson had to acquire external and develop internal
competencies in radio communication and in the integration and design of systems. As
explored in the next section, Ericsson both expanded rapidly in-house in Sweden and
bought Swedish firms but particularly expanded abroad, including R&D centers.
10.2 Envisioning and Acting upon Opportunities
Table 5 summarizes the results, in relation to key innovation opportunities for
mobile telecommunication as explained in this report. These are innovation
opportunities in the sense that these were key events, where the firm could choose to
act, or not, upon the market and technical potentials. The seven innovation opportunities
identified are:
1) Setting of the NMT standard,
2) Selling AXE for NMT,
3) Radio communication for mobile telecommunication,
4) Infrastructure system provider,
5) Ericsson Information Systems (EIS),
6) Setting of the GSM standard, and
7) The move from analogue to digital standards, with different standards
internationally.
59
Table 5: Relations in SI at key innovation opportunitiesIm
port
ant s
ales
Nor
dic
PT
Ts
Sa
ud
i Ara
bia
Th
e N
eth
erla
nd
s
Nor
dic
PT
Ts
Sa
ud
i Ara
bia
Th
e N
eth
erla
nd
s
Th
e N
eth
erla
nd
sO
per
ato
rs
Ger
ma
ny
Fra
nce
US
Jap
an
Etc
.
Rel
atio
ns to
othe
r fir
ms
No
dire
ctin
volv
emen
t of
firm
s
-Elle
mte
l's R
&D
wor
k-B
X m
odify
Pu
rch
ase
of f
irms:
Ma
gnet
ic (
19
83
)S
ona
b (
19
78
)R
ad
iosy
stem
(1
98
8)
Pu
rch
ase
of f
irms:
-Fa
cit
(19
80
)-D
ata
saa
b (
19
81
)
-Firm
s of
fer
alte
rna
tives
-Tel
ever
ket
spre
ad
info
rma
tion
Str
ate
gic
alli
an
ces
Gov
ernm
ent
acto
rs
Nor
dic
PT
Ts
Tel
ever
ket
an
dN
ord
ic P
TT
s
-Mili
tary
-Tel
ever
ket
(TV
tra
nsm
issi
on)
-Tel
ever
ket
-In
tern
atio
na
lst
an
da
rds
bod
ies
-PT
Ts
-Tel
ever
ket
-Nu
tek
-IT
4-R
AC
E-
Inte
rna
tion
al
sta
nd
ard
s b
odie
s
Uni
vers
ity
Non
e d
irect
imp
act
Non
e d
irect
imp
act
Non
e d
irect
imp
act
Non
e d
irect
imp
act
-Ch
alm
ers
-Lu
nd
-KT
H-L
inkö
pin
g
How
acq
uire
in-h
ouse
com
pete
nce
-Tes
ting
inT
elev
erke
t's' l
ab
s
-Tra
nsf
er fr
omE
llem
tel
-BX
Mod
ifyso
ftwa
re-S
RA
dev
elop
sp
roto
typ
e
-In
tern
al b
uild
-up
-Coo
per
atio
n w
ithB
X-
Acq
uire
firm
s
-BX
-BR
coo
per
atio
n-S
RA
ma
na
ges
-Tre
men
dou
s in
-h
ouse
exp
an
sion
-Acq
uire
firm
s
-R&
D-C
olla
bor
atio
n
-ID
EO
N-4
0 R
&D
cen
ters
inte
rna
tion
ally
-Hig
h R
&D
exp
end
iture
s in
-h
ouse
Vis
ion
-Wha
t-W
ho
-Com
mon
sta
nd
ard
an
d m
ark
et-N
ord
ic P
TT
s, e
sp.
Tel
ever
ket
-Sp
ecia
lized
nic
he
-SR
A/L
un
dq
uis
t
-RB
S a
nd
ph
ones
-SR
A/L
un
dq
uis
t
-Pro
vid
e a
com
ple
tein
fra
stru
ctu
re-
SR
A/L
un
dq
uis
t
-Pa
per
less
offi
ce,
inte
gra
tion
of
com
pu
ters
an
dte
leco
mm
un
ica
tion
-CE
O S
ved
ber
g
-Com
mon
Eu
rop
ean
dig
ital s
tan
da
rd-7
alte
rna
tives
from
firm
s-O
ne
alte
rna
tive
from
Tel
ever
ket
-Eri
csso
n s
ell a
llst
an
da
rds
-Dev
elop
all
sta
nd
ard
s
Dat
e
19
80
Fro
m 1
98
2
Late
70
s, e
arly
80
s
19
82
19
82
-19
88
-Pa
ris t
estin
g 1
98
6-M
ad
eira
dec
isio
n1
98
7
Late
80
s a
nd
on
Set
ting
of N
MT
stan
dard
AX
E s
witc
h fo
rN
MT
Rad
ioC
omm
unic
atio
nfo
r m
obile
tele
com
mun
icat
ion
Infr
astr
uctu
reS
yste
m P
rovi
der
EIS
Set
ting
of G
SM
stan
dard
Ana
logu
e to
Dig
ital
and
inte
rnat
iona
lst
anda
rds
60
Table 5 indicates two important points. Firstly, the Swedish / Nordic context was
initially very important for Ericsson, both as a market and for technical specifications
by government actors. Over time, its importance diminished, not only because Ericsson
became truly multinational but also because competing firms in the industry were the
ones with the knowledge and competencies to drive technical and market developments,
including standards. Secondly, Ericsson’s relationships to others, and hence the system
of innovation itself, are dynamic over time. As the nature of market and technology
changes, the types of relationships important for accessing vital economic and technical
knowledge also changes. Over time, contacts with universities and international firms
with complementary assets has been increasingly importance, parallel with an in-house
building-up of competencies.
A system of innovation must therefore be seen from a dynamic perspective. Instead
of assuming it exists and affects innovation and economic growth, this report
demonstrates the importance of asking under what circumstances different types of
relationships and institutions can, or have been, important.
10.3 Implications
The implications for firm management and for government policy will be drawn of
this study of Swedish firms developing a mobile telecommunication industry.42 There
are a number of points making our an interesting example of the dynamics of high tech
industry, specifically a high tech product area. As argued in Edquist, Hommen and
McKelvey (1997) and Edquist and Texier (1996), high tech industries are characterized
by a high percentage of sales spent on R&D, and especially those developing products
have important positive effects on employment and productivity growth. What our
report shows is that for the firm Ericsson, this business area has been the motor behind
firm growth in terms of employment and sales, and that more broadly within the
telecommunication industry, mobile telephony has grown rapidly and also gained
market share relative to fixed telecommunication. However, by the early 1990s,
Ericsson had no direct Swedish competitors, mainly because there were only a few
small ones to begin with and Ericsson had already bought them up to expand their
technical competencies and to ’clean up the market’.43 Thus, other Swedish firms have
42They should also be read in connection with the study of Nokia (Lemola and Palmberg 1997) and
of technological diversification in telecommunication internationally (Malerba et al 1997).43There have been, however, a number of subcontractors, but these are not examined here.
Ericsson’s subcontracting is international rather than national. The basic policy has been to purchaserelatively standard items on the market, although often from suppliers with a long term relationship, andto produce in-house critical components and assemble. Over time, the definition of what is criticalchanges.
61
not grown, although there are a few small niche firms as well as supplier firms. Ericsson
is, moreover, an international company, with the majority of its business and more and
more expansion coming internationally, rather than within Sweden.
Thus, one conclusion of this case is that small and/or new innovating firms have not
been important for technological change, industrial dynamics and economic growth.
Instead, what is shown to be important here is a large firm which diversified into related
products, but which were based on different technologies and markets than traditional
ones. Because of these differences in core technologies and types of markets (as well as
organizational change), moving from fixed to mobile telecommunication can be seen as
diversification. Re-orienting the firm has not been an easy process, and this report has
focused on the nature and types of challenges facing a large firm which is trying to
develop, access and use new competences in a situation of much market and technical
change.
The role of small firms has here been as a source of regeneration for the large firm
rather than as strong competitors for elements of the mobile telecommunication system
and for the phones. Early on, small firms were direct competitors and innovative, but
they had less possibilities to compete directly when the responsibility for designing and
forming the whole network moved from the purchasers (PTTs and other operators) to
the producing firm. Thus, there have been costs of developing large scale, network
technologies and taking responsibility for large projects requiring many complementary
assets which make it difficult for the small to compete. In fact, the large firm has
generally bought the few national small firms after awhile, to renew competences, etc.
(Note, however, that we have not investigated the relationship between Ericsson and
non-competitor firms like suppliers or software developers where the small firms may
play a significant role).
Having said that mobile telecommunication is an example of a successful high tech
product area, policy-makers might want to jump to the conclusion that if they want to
support high tech product industries, they should ask firms which technologies and
products they believe in, and then direct policy there. That conclusion is not supported
here. In fact, what previous sections tell us is that Ericsson top management—as well as
most in the indutry sector—believed in a different high tech future during the 1980s,
namely the ’paperless office’ or integration of telecommunication and information
technology. This was also when the basic technologies for civilian mobile networks
were being developed and the first orders sold. Similarly, there was a general consensus
among competing firms and analysts that mobile telecommunication would be a small
niche market. Twenty years into its history, the estimations of future expansion usually
still fall short of actual growth in demand. However, there were what we can call
internal firm entrepreneurs within Ericsson, SRA and small companies who believed in
the mobile telecommunication vision even then. They were willing to pursue their
62
vision, even when others laughed at their ideas and regardless of whether they were
losing or making money.
One implication, therefore, is that if firm management and/or government policy-
makers want to support new technologies with innovation potentials, they have to
accept that decision-making cannot be totally a rational, top-down process. An example
of this way of thinking would be to think that they can gather all relevant information,
and based on that, ’pick the winners’. Although laudable in the sense both of the
necessity of having information to make decisions and of trying to use resources
efficiently for a specific goal, this approach to policy-making will not work. Or, rather,
it might work but the overall policy is more likely to fail whereas some small projects in
it have unexpectedly important effects.
One of the reasons that this top-down decision-making does not work so well,
especially for high tech area, is the problem of information creation and coordination
under conditions of uncertainty about future technical and market development (See
Malerba 1997, Cohendet and Llerena 1997). Actors in a SI will always have access to
only limited technical and market information and even if an actor could gain all
information about today, they still could not know about future changes in markets,
technologies and innovation opportunities. For example, Ericsson top management not
only strongly supported the integration of telecommunication and computers in the
paperless office, they also invested massive resources and expanded employment in that
area. It still failed, partly because of difficulties to find and implement technical
solutions but also because the market expanded much slower, and with other demands,
than expected.
Thus, one implication of uncertainty and of different actors interpreting the same
information differently is that decision-makers must try to solve problems of
coordinating information flows but also of acquiring or developing bits of information
crucial to innovation opportunities. Based on that information they can find and
interpret, decision-makers still have to take risks to try to develop innovation
opportunities. They must also accept that failures occur and also be willing to redirect
strategies when circumstances change.
Thus, the other way to approach the problem of how to support high tech areas,
which is supported by the empirical material here, is to accept that innovation processes
involve uncertainties, failures as well as successes, as well as problems of gathering,
developing and interpreting information and knowledge.44 What Ericsson management
did right, partly because SRA was independent within the larger concern, was to allow
room for the initial ideas and technologies to develop. When market opportunities were
identified, as in the first orders in the early 1980s, then SRA managers could make their
44See (Smith 1997, Malerba 1997, McKelvey 1997b)
63
case and convince others to contribute to the cause, either by putting aside some man-
hours for development and modification work or later by purchasing firms with
necessary competencies.
Within the firm, this implies not only daring to take risks to pursue a vision, but
also being willing to support a variety of visions and approaches to take advantage of
innovation opportunities. There are always multiple ways to tackle a technical problem
and a variety of potential solutions, even when one general overall approach guides the
thinking (McKelvey 1996a). As Schumpeter argued, taking advantage of an innovation
may mean modifying an existing product to a new application / market, as in the case of
Ericsson switches, or it may be developing new areas of knowledge, as in radio
communication for radio base stations and cellular phones. Too much direct
management techniques requiring immediate estimation of returns and potentials can
easily stifle areas which have great potentials, that is, potential both for great losses and
great profits.
In this particular case, initial opportunities for innovation were made much more
immediate and apparent by the Swedish PTT Televerket. They were instrumental in
defining the first Nordic standard, along with the other Nordic PTTs, and the group not
only spent ten years defining the standard, Televerket also tested the equipment of
firms. They actively worked for a standard open to all (international) firms involved,
and the standard was also so very detailed that an operator with technical competence
could purchase equipment from different firms and create a system. Small firms could
thus also sell, without being directly dependent on a large one. Televerket could play
this role particularly because of its technical competencies and willingness to spend
resources over a long period of time. This standard was important for Ericsson and
other Nordic firms not only because they could identify a potential Nordic market but
also because some other countries used the standard. This gave firms potential
economies of scale, if they could sell products, and although the total population of the
Nordic countries involved was sparse, the system was to cover much of the countries
(partly in the interest of provision of a public good), and so the fact that the population
is spread out geographically meant the systems would still be large. Ericsson top
management felt that having a large functioning system(s) would help them win later
orders as well, and they were correct in this.
It is important to point out, however, that bidding for the Nordic PTTs’ orders was
international, implying that international firms also knew the standards, were in the
running and did win some of them. This is not a case of direct government industrial
policy for national firms, although Nordic firms did tend to win more of the orders.
Moreover, other countries were also developing competing standards, due to actions by
firms and/or government agencies. In fact, SRA was so concentrated on the American
market that it did not know so much about the technical solutions for the Nordic
64
standard. This created a disadvantage when they wanted to try to sell to their home
market. One implication is that although a government actor with strong technical
competences was important here, other standards and other ways of setting standards
could have been used to highlight this innovation opportunity.
For Ericsson and the other Swedish companies, relationships within the Swedish
system of innovation were important for the initial move into mobile
telecommunication in the late 1970s and early 1980s. These relationships were
important for providing and interpreting technical information as well as market
demand. Firms moving into any of the component technologies had their technical
competencies from longer term relationships and orders with other Swedish actors,
especially Televerket and the military. Firms’ development and use of technical
knowledge had relatively little contact with universities or institutes of technology.
Thus, for example within SRA, it was probably easier to create room to explore a new
vision of these in-house entrepreneurs when a long-standing Ericsson partner,
Televerket, was expressing a clearly defined demand. At the same time that Swedish
relationships were important, firms were oriented to international markets as well,
particularly for new orders (but less so for relationships to catalyze technical
developments). Internationally here really means international rather than European,
perhaps because initially telecommunication was heavily regulated and especially
American, European and later Japanese markets monopolized by national firms and
hence closed to non-national firms.
Over time, two things changed in the dynamic relationships between firms and
systems of innovation. For the first, the relative importance of the national Swedish
dimension versus the international one decreased in importance. This is particularly true
for Ericsson, which today has 40 R&D centers internationally; 97% of sales outside
Sweden; collaboration with other firms as it sees expedient; sells networks and phones
for all the international standards, etc. The only element of European integration that
has continued to be important is the European digital standard GSM. It has lead to
economies of scale for the firms because GSM is used not only for European markets
but also in many other countries.
The other aspect which changed is which types of actors and relationships have
become more important as others become less important. For the first, national PTTs
have decreased in importance for national firms. A PTT or even a group of PTTs have
neither the resources nor can work fast enough to determine new standards; standards
are instead set by international bodies, including representatives from many different
interested actors. Instead, it is that the large firms that have the in-house competence to
evaluate and ’bet’ on certain solutions or techniques, and Ericsson, for example, has
been very active in trying to influence the setting of basically all standards.
65
Secondly, relationships to other firms with competencies in other areas have
become increasingly important. Although Ericsson employees have stressed that they
do much themselves and that the firm expends enormous resources on R&D to do so,
one firm cannot master all necessary bodies of knowledge, which moreover are
changing rapidly. They therefore have various types of relationships like licensing,
collaboration, joint firms, etc. in order to gain access to relevant (technical) information.
What is interesting here is not just that Ericsson spends much on R&D to innovate but
that they do so based on explicit strategies to focus on core activities but also closely
monitor external events.
Thirdly, the shift from analogue to digital has made relationships with universities
increasingly important, as new aspects of knowledge have often had to be developed in
parallel with developing new ways of doing things, or implementing ideas into practice.
Advances in knowledge in a number of areas relevant for production and for product
components have led to phones with a lower price but higher capabilities, which has
attracted new types of buyers and hence expanded the market.
66
11. ReferencesAffärsvärlden (1981). ’LM byter strategi, Jättemarknad lockar’, 6/5 nr 1, pp. 25-29
Affärsvärlden (1982). ’Ericsson Information Byter Både Marknad och Kunskap’, 27/1nr 4.
Affärsvärlden (1985a). ’Ericsson inför stor Omställning - Bromsar och Satsar menTappar Tempo’, 6/2 nr 6, pp. 15-20
Affärsvärlden (1985b). ’Sverige och Norden erövrar världen’, 13/11, nr. 46, pp. 17, 19,23, 24.
Affärsvärlden (1987). ‘Ericsson mijlardsatsar för täten i tillväxtbranschen’, 4/3, nr. 10,pp. 17, 19, 23, 26.
Affärsvärlden (1988a) ’Ericsson InformationSystem in Ny Fas - Nollresult Uppnått . NuSäljs Datasidan?’ 20/1 No. 3, 56-58.
Affärsvärlden (1988c) 'Ericsson går in i Offensivare Skede' 27/1, No 4, pp. 4-5
Affärsvärlden (1988b) ’Ericsson/Radio System: Mycket dyrt köp’ 27/28 No.8
Ahlbom, Helen (1992). “Ericsson gör allt—än så länge”, Veckans Affärer nr. 43, 21Okt., pp. 52, 54.
America’s Network (1996). ’Wonders of the world phone: L.M. Ericsson’s planneddevelopment of the world phone’. v.100, Nr. 22, pp.46-7.
Andersson, Göte (1992). ’Alla räknar med att öka sina GSM-andelar’Elektroniktidningen. Nr. 12, pp. 23, 25.
Ansoff (1954) Strategies for Diversification, Harvard Business Review, 35 (5), 113-124.
Arnold, Erik, and Guy, Ken (1989). Evaluation of the IT4 Programme, Phase I. Booz-Allen and Hamilton.
Bevenius, Ingvar (1995). History of Ericsson Radio Systems AB. Mimeo. PR, EricssonRadio Systems.
Cohendet, Patrick and Llerena, Patrick (1997). ‘Learning, Technical Change, and PublicPolicy: How to Create and Exploit Diversity’ in Edquist (ed.).
Communications International (1992). ’Ericsson: Growing up and leaving home’, July,pp. 30-31.
Dahmén Erik (1988). 'Development Blocks' in Industrial Economies, in ScandinavianEconomic History Review, Vol 36, No1, 1988, pp. 3-14.
Dataquest (1997). ’Dataquest reports a major shakeup in the US Digital HandsetMarket’, www.dataquest.com, Spring 1997.
Datornytt (1985) Forsatta förluster PCn "lik i lasten", 9/85, p.37.
Dosi, G.; Freeman, C.; Nelson, R.; Silverberg, G.; and Soete, L. (eds.)(1988).Technical Change and Economic Theory. Pinter Publishers. London.
Edquist, C. (ed) (1997). Systems of Innovation—Technologies, Institutions , andOrganizations. Cassell, London.
67
Edquist, Charles, Hommen, Leif and McKelvey, Maureen (1997). Product vs. ProcessInnovation: Implications for Employment written for the Innovation Systems andEuropean Integration Program.
Edquist, Charles and McKelvey, Maureen (forthcoming 1997). ’High R&D Intensitywithout High Tech Products: A Swedish Paradox?” in the edited volume of theEAEPE 1994 Copenhagen Conference: in Nielsen, Klaus and Johnson, Björn (eds.)Evolution of Institutions, Organizations, and Technology(Aldershot: Edward Elgar).
Edquist, C., Texier, F. (1996), The Growth Pattern of Swedish Industry in InnovationSystems and Competitiveness, Osmo Kuusi (Ed.), Taloustieto Oy Publisher, Helsinki1996.
Ehrnberg, Elinor and Jacobsson, Staffan (1997) 'Technological Discontinuities andIncumbents' Performance: an Analytical Framework' in Edquist (ed.).
Ekwall, Johan (1991). ‘Ericsson får se upp med nya Motorola’, Veckans affärer nr. 33,14 Augusti, pp. 38-41.
Emmett, Arielle (1996). ’Wonders of the World Phone’, America’s Network, v. 100, no.22, p. 46-47.
LM Ericsson (LME) (1990). Annual Report
LM Ericsson (1995). Lars Magnus Ericsson 150 Years and the Company’s First 120Years. Ericsson.
LM Ericsson (1996). ’Ericsson Interim Report: Nine months ended Sept 30, 1996’.October 29. Business Wire.
LM Ericsson (1996b). Reference List: Wireless Systems. Mimeo. July 1996.
LM Ericsson (1996c). Annual Report
LM Ericsson (1997a). ’Mobile Telephony—Market Overview’. Information fromEricsson home page http://www.ericsson.se, spring 1997.
Ersman, Nina (1987). ’Vi måste bli mindre svenskt’, Veckans affärer Nr. 8, 19 feb, pp.39, 41.
Free, John (1983). ’Take-along Telephones—cellular radio unlocks new channels forsubscribers’, Popular Science 223, October, pp. 20, 24, 28.
Froste, Calle (1997). ’Philips mobiltelefoni kniper andelar’, Svenska Dagbladet Ons. 4Juni, p. 30.
Grandstrand, Ove (1993) the Evolution of Nordic Mobile Telephony, CIM WorkingPaper No: WP 1993: 06.
Henderson, R. (1994), ’The Evolution of Integrative Capability: Innovation inCardiovascular Drug Discovery’, Industrial and Corporate Change 3, 607-630.
Hultén, Staffan and Mölleryd, Bengt (1993). ’Cellular Telephony in Retrospect—Foundations of Success’, Tele 1, pp. 2-7.
International Management (1994). ’Sweden’s Mobile Leader’, May, pp. 24-25, 27-29.
IT4 (1990).’Projekt inom IT4-Programmet’, Januari 1990, IT-Delegation, Stockholm.
IT4 (1991).’Projekt inom IT4-Programmet’, Januari 1991, IT-Delegation, Stockholm.
68
IT4 (1991b). ’Provnät för GSM-Systemet’, IT4-Projekt 3102, IT-Delegation,Stockholm.
Jacobsson, Saffan and Rickne, Annika (1996) New Technology Based Firms in Sweden- a Study of their Impact on Industrial Renewal, working paper, department ofIndustrial Management and Economics, Chalmers University of Technology.
Jismalm, Greger and Lejdal, Jan-Olof (1990). ’Cell Planning—Products and Services’,Ericsson Review No. 2, pp. 84-91.
Jismalm, Greger and Rydbeck, Nils (1987). ’Ericssons telefonapparater förmobiltelefonsystem’, Ericsson Review 3, pp. 141-150.
Källen, Anna-Karin (1989). ’Icke Önskvärd’, Ledarskap 12/89, pp. 16-21.
Lemola, Tarmo and Palmberg, Christopher (1997). Case study of Nokia, ISE project.Draft presented at the Vienna Workshop, September.
Lilliesköld, Gunnar (1989). ’NMT, GSM and “sladdlös PABX” utvecklingsprojekt iLund’, Modern Elektronik 3/89, p. 35.
Lindmark, Sven (1995). The History of the Future: an Investigation of the Evolution ofMobile Telephony, Licentiate Thesis, department of Industrial Management andEconomics, Chalmers University of Technology.
McKelvey, Maureen (1996a). Evolutionary Innovations: The Business of Biotechnology.Oxford University Press, Oxford.
McKelvey, Maureen (1996b). ’Technological Discontinuities in Genetic Engineering inPharmaceuticals? Firm Jumps and Lock-in in Systems of Innovation’ TechnologyAnalysis & Strategic Management 8:2. pp. 107 - 116.
McKelvey, Maureen (1997a). ’Using Evolutionary Theory to Define Systems ofInnovation’ in Edquist (ed.).
McKelvey, Maureen (1997b). ’Mobile Phones: Success for Scandinavian Firms?’Report from a workshop with the same title, held at Deparment of Technology andSocial Change (Tema T), Linköping University, on January 16, 1997. Tema Tworking paper No. 179.
McKelvey, Maureen (1997c). ’’ (Economic Evolution: Small but also Radical Changes)in von Pierer, Heinrich, and von Oetinger, Bolko (eds.) (1997). Wie kommt das Neueauf die Welf? (Visions of the New—Reconceptualizing Innovation). München: CarlHanser Verlag.
McKelvey, Maureen (forthcoming) ’R&D as Pre-market Selection: ManagingUncertainty in Genetic Engineering and in Broadband’ in Green, C. and Eliasson, G.(eds.) Entry, Competition and Economic Growth Selected papers from the 6thInternational Joseph A. Schumpeter Society Conference, University of Michigan.
McKelvey, Maureen and Edquist, Charles (1997). ’Swedish Specialization in Researchand Development: Strength or Weakness?’ Tema T working paper No. . Originally apaper presented at the EAEPE 1995 conference Transforming Economics andSocieties: Towards an Institutional Theory of Economic Change’ Krakow, Poland,19-21 October.
McKelvey, Maureen and Texier, Francois (forthcoming). ’Large Firms SurvivingTechnological Discontinuities: Mobile Telecommunication Industry’ Paper acceptedfor the annual EAEPE (European Association for Evolutionary Political Economy)conference, Athens, Greece, November 6-9, 1997.
69
Malerba, Franco (1997). Public Policy in Industrial Dynamics: an EvolutionaryPerspective. Written for the Innovation Systems and European Integration Program.
Malerba, Franco, Breschi Stefano and Lissoni Francesco (1997) Patterns ofTechnological Entry and Exit: Evidence From France, Germany, Italy and the UK.written for the Innovation Systems and European Integration Program.
Mannberg, Lare-Erik (1984). ’En äkta svensk’, Teknik för alla 12, p. 64.
Meurling, John and Jeans, Richard (1994). The Mobile Phone Book, the Invention of theMobile Phone Industry, London: Communications Week International.
Meurling, John and Jeans, Richard (1995), A Switch in Time, AXE - Creating aFoundation for the Information Age, London: Communications Week International.
Månsson, Erik (1981). ’LM Ericsson: Telefonjätte på väg att bli datajätte’ Data 1981:4,pp. 31-33.
Månsson, Erik (1985). ’Ericsson: Kris eller succé?’, Nordisk DATAnytt, 2/85, pp. 24-25.
Mölleryd, Bengt (1996). Så Byggdes en Världsindustri -Entreprenörskappets Betydelseför Svensk Mobiltelefoni, Licenciate Dissertation, EFI Research Report, StockholmSchool of Economics.
Nelson, R. (1991). ’Why do Firms Differ, and How does it Matter?’, StrategicManagement Journal, Vol. 12, 61-74.
Nutek (1997). Telecommunication: Evaluation of the Research Program 1993-1996. RNutek 1997:5, Stockholm:Nutek.
Ny Teknik (1996b). ’Ericssons Exportraket: Från 0,7 till 10 miljarder på bara fem år’,1996:51-52, pp. 12-13.
Ny Teknik (1997a) ‘Han bestämmer Sveriges framtid: Ramqvist and Ericsson’, 20/97,pp. 24-27, 29-31.
Ny Teknik (1997b) ‘Sveriges FoU liga’, 21/97.
Nylander, Jan (1990). ’Att ligga i Lund’, Svensk Export 1990:2, pp. 30-31.
OECD (1993). Main Science and Technology Indicators
OECD 1997. Communications Outlook 1997, Vol. 1. Paris: OECD.
Oskarsson, Christer and Sjöberg, Niklas (1994= Technology Analysis and CompetitiveStrategy—The Case of Mobile Telephones’.
Palmgren, Anders (1992). ’Rebell bakom framgången’, Veckans affärer Nr. 41, 7 Oct,pp. 52-53, 55.
Perroux, F. (1969). L’Économie du XXe Siècle, 3rd ed., Paris: Pressses Universitaires deFrance.
Radiosystem 1985, Information Brochure Before Introduction On The Stock Market.
Skidé, Ulla (1994). ’Jan Uddenfeldt i spetsen för mobiltelefonin’, ElektroniktidningenNr. 1, p. 13-14.
Skole, Robert (1982). ’New Ericsson division aims at U.S.’, Electronics 55, Feb.pp.101-102.
Smith, Keith (1997). Systems Approaches to Innovation: Some Policy Issues. Writtenfor the Innovation Systems and European Integration Program.
70
Smith, Ray (1984). ’Ericsson bids on Bell COs’. Telephone Engineer & ManagementDecember 1, pp. 136-137.
Sundqvist, S.I. (1988) Ericsson misslyckade satsning på inormationssystem - tremiljarder kostade äventyret, Dagens Nyheter, 1988/01/24
Technopolis and SPRU (1992). Evaluation of the IT4 Programme, Final ReportStockholm, IT-Delegation.
Teece, D.; Rumelt, R.; Dosi, G.; and Winter, S. (1994). ’Understanding CorporateCoherence: Theory and Evidence’, Journal of Economic Behaviour andOrganization Vol. 23, 1-30.
Tushman, M. and Anderson, P. (1986). 'Technological Discontinuities andOrganizational Environments' Administrative Science Quarterly 31, pp. 439-465.
Veckans äffärer (1987). ’Nu måste Ericsson visa resultat också!’, Nr. 21/22, 21 maj, pp.88-91, 95.
Vedin, Bengt Arne (1992). Teknisk Revolt - det Svenska AXE-Systemets BrokigaFramgångshistoria, Atlantis.
Vedin, Bengt-Arne (1994). ’Så blev AXE ett sjumilakliv’, Daedalus Stockholm:Tekniska Museets Årsbok.
Wolmesjö, L-G (1992). ’I hans huvud finns kunskap för 30 miljarder’, Dagens IndustriTors. 13 Feb, p. 16-17.
Interviews:
Lundquist Åke, former CEO of Ericsson Radio System, interview April 10, 1997 at hisprivate residence, Södertälje.
Kågström Leif, Engineer from Magnetic who was in the starting team of Radiosystem,now Senior Development Engineer, Ericsson Radio Access, interview April 16,1997 a Ericsson Radio Access, Kista.
Meurling John, retired from Ericsson Public Telecommunication, interview April 14,1997 at his private residence.
Mäkitalo Östen, Director of Telia Research Laboratories, interview May 07, 1997 atTelia Research laboratories.
Persson Åke, Engineer from BX who adapted the AXE switch for BR, now VicePresident Marketing and Sales, Ericsson Radio System, interview April 10, 1997 atERA, Kista.
Öhrvik Sven-Olof, Engineer at Ericsson Radio System, now consultant for the nextgeneration of mobile communication system, interview May 13, 1997 at LinköpingUniversity.
Örnulf Erik, Engineer from BX who adapted the AXE switch for BR, now VicePresident, Ericsson Radio System. interview April 10, 1997 at ERA, Kista.
71
List of Abbreviation List of Abbreviation
AMPS Advanced Mobile Phone Service
AXE Digital Swich develped by Ellemtel
BR Business Radio at Ericsson
BSC Base Station Controler
BX Business Public Telecommunication at Ericsson
CDMA Code Division
D-AMPS Digital Advanced Mobile Phone Service
EIS Ericsson Information System
ERA Ericsson Radio System
FDMA Frequency Division Multiple Access
GSM Global System for Mobile Communication (previously Groupe
Spécial Mobile)
ISDN Integrated Services Digital Network
MTA Mobiltelfonsystem A
MTB Mobiltelfonsystem B
MTD Mobiltelfonsystem D
NEC Nippon Electronic Corporation
NMT Nordic Mobile Telephone System
PDC Personal Digital cellular
PTT National Telephone Operator
RBS Radio Base Station
SI System of Innovation
SRA Svenska Radio Aktibolag, also called BR
TACS Total Access Communication System
TDMA Time Division Multiple Access
72
Appendix A: Summary of developers and suppliers for differentstandards implemented in Sweden