|Cities and the ‘new economy’.
The clustering of the software industry in Oslo
By Arne Isaksen,
Agder University College, Grimstad, Norway, and
the STEP group, Oslo, Norway
Paper prepared for the
IXth National Meeting in ERA,
Lisbon 27-29 June 2002
Draft version. No ‘languish wash’
This paper examines the reasons for the large concentration of software firms to Oslo. The paper analyses how software firms perform individual activities, and how they interact with other players in performing the activities. The concentration of software companies to Oslo rests first of all on the need of dense interaction among software consulting companies themselves, and between consulting companies and important customers and platform suppliers. The fact is that consulting activity is project based and involves lots of constellation building and face-to-face contact, which is facilitated when players co-locate. Keeping these characteristics in mind, a large scale decentralising away from Oslo will not take place in the near future. The paper discusses four distinct theoretical approaches that may inform the empirical analyses. The approach that emphasises a strong clustering effect of ‘new economy’ sectors in large cities is a fruitful analytical tool in interpreting the concentration of software firms in Oslo, with one exception. The approach postulates that firms in new industries tend to cluster near universities and research institutes, however, few software firms in Oslo uphold any dense contacts with the regional knowledge infrastructure.
The period since the mid 1970s seems to represent a transition to a new phase of capitalist development that is distinctly different in its organisation and geography from the preceding Fordist period (Amin 1994). It is argued that the transition is accompanied by a different geography of production. In the 1980s and in the first part of the 1990s, the changes were characterised as a resurgence of regional economies, as a number of local areas were seen to be experiencing new growth (Piore and Sabel 1984, Storper 1997).
The model postulates a reorganisation of production away from the former centres of capitalist accumulation, i.e. the large city regions. The new way of organising production is said to result in a number of ‘self-propelling’ regions, or clusters, as it has come to be called (Porter 1998). Much studied clusters are industrial districts in Italy (Sengenberger and Pyke 1992) and ‘new industrial spaces’ of high technology firms (Scott 1988). However, the postulation of a large-scale geographical decentralisation of industry and jobs away from central areas does not agree with the reality in most western countries. For example, a convergence in GDP per capita took place between regions in the EU in the period 1960-1980. However, regional convergence ceased after 1980 and up until 1995 (Cappelen, Fagerberg, Verspagen 1999), meaning that the well-off regions defended their lead against the poorer ones. The largest cities continued to grow at a fast pace.
The growth of the large cities is also explained by referring to a new phase in the development of capitalism, although of a somewhat different kind from the one postulated in the 1980s. The late 1990 version of the capitalist transition focused on the coming of a new economy, a more knowledge based economy, a learning economy or the digital economy. Terms like the new economy are often fuzzy concepts (Markussen 1999), i.e. they may have two or more alternative meanings, theoretical propositions may lack empirical validation, or they are often based on very selective evidence, as a few case studies. However, the new economy has come to mean in particular the growth of new economy sectors like ICT (information and communication technology) and biotechnology, as well as an ever-increasing use of new technology in industry1.
New economy sectors are seen to form regional cluster to a large degree, which are frequently located in large cities. Thus, ‘the new economy sectors are highly skewed in geographical terms. They are overwhelmingly found in or near large, well-diversified services and knowledge-based cities or specialist research university campus cities or towns (Cooke 2002: 130-131). In contrast, new economy clusters are not seen to exist in rural areas or in areas containing old and declining industries.
This paper examines the cluster building mechanisms and the centralising and decentralising forces in the software industry in the Oslo area. The industry is narrowly defined as NACE 722, ‘Software consultancy and supply’. The Oslo area2 holds as much as 66% of the 18.700 jobs in this industry in Norway in 1999, while the area has 26% of all jobs. The software industry may be seen as the archetype of a new economy sector. The sector produces standard and tailor-made software to be implemented in companies in other sectors.
The paper builds on personal interviews with leaders (mostly managing directors and sales officers) in 14 software firms in the Oslo area, including a good number of the largest firms3. The interviews above all aimed to ‘unpack’ the sector, i.e. to determine what firms actually do ‘on the ground’. Thus, the interviews intended to find out what types of activities the firms perform, to examine how innovation and production goes on, to analyse the relations between the software firms and their clients etc. The aim of the interviews was to learn more about how the software industry works in order to be able to construct a good questionnaire for a follow-up telephone survey to amongst others a larger population of software firms and potential clients of these firms.
The detailed case study of the software sector in the Oslo area, which is based on the personal interviews, addresses one major problem with existing research on the geography of new economy sectors. There seems to be a tendency for this research to rely on secondary data, often derived from official classifications, which are insufficiently specified and calibrated to address key research questions. In studies of industrial adjustment, there is a crucial need for researchers go beyond statistics and to ‘get inside’ firms and industries in order to understand the origins, developments, composition, core competencies, location requirements etc. of industries.
The paper explores three main questions, two empirical questions and one theoretical:
Which activities constitute the software industry in the Oslo area?
To which extent comprises (part of) the software industry an innovative cluster in the Oslo area? What are the clustering mechanisms? Which mechanism may lead to decentralisation of activities and jobs from the Oslo area?
Which of four specific theoretical approaches are most relevant in explaining the development and clustering effect of the software industry in Oslo?
The rest of the paper is divided into four parts. The first part reviews and discusses four main theoretical approaches that may be relevant guides in exploring cluster building and decentralising forces in the Oslo software industry. The second and third parts consist of empirical analyses of the software industry in the Oslo region. The second part examines important aspects of the main activities in the software industry, while the third part analyses the cluster building process in the software industry in Oslo. Finally, the forth part discusses wider methodological, empirical and theoretical lessons from the case study.
The importance of the 'new economy' on regional clustering4
Information and communication technologies (ICTs), and the new economy with which they are associated, may have great implications for regional industrial development. ICTs may ‘”change the balance” between centralising and decentralising dynamics in the space-economy’ (Gillespie, Richardson and Cornford 2001: 109). A number of quite different spatial expressions of the rapid introduction of ICTs can be postulated, some which may benefit the industrial development of large cities and others that may disadvantage large cities.
The paper reveals four possible, but not mutually excluding, impacts of ICTs with regard to regional industrial development. The first two refer to the locational impacts of ICT as a new economic sector in itself. The third and fourth refer to the way ICTs are used in the economy and the possible impact of ICTs on the industrial location pattern more generally. The four spatial expressions are (cf. Table 1):
The clustering of new economy sectors in large cities;
The growth of new industrial spaces (or clusters) outside the large cities containing ensembles of firms in new economy sectors;
The decentralisation of (some) industrial activities away from large cities rendered possible by the distance-transcending capabilities of new technologies;
The inclusion of firms in evolving global networks orchestrated by transnational corporations using ICTs to co-ordinate global activities.
i The clustering of new economy industrial sectors in large cities
One possible outcome of the new economy ‘emphasises the strong clustering effect that can be witnessed in those regions most associated with the emergence of the new economy, such as Silicon Valley in California, or in the concentration of "dot.com" start-ups in major cities such as London' (Gillespie, Richardson and Cornford 2001: 1). While the cluster building in areas like Silicon Valley and London is based on many of the same mechanisms, the history of the areas may diverge in important aspects. The result is also starkly different as regards implications for regional development. The development of new economy clusters in major cities like London has a centralising effect on the space-economy. Silicon Valley, on the other hand, is the archetypical new industrial space, in which industrial growth often occurs at the expense of the traditional industrial centres in and around the large cities.
The first step in the growth of a new economy cluster is the birth of the pioneer firms in a new industry. The birth of these firms is often based on some radical innovations: the innovations so to say create the new industry. The innovations and new firms can often be traced to historical circumstances, such as the availability of raw materials, specific knowledge in R&D organisations, or the specific or sophisticated needs of a certain group of (geographically concentrated) customers or firms. The first approach claims that pioneer firms in a new industry often start in large cities. Large cities have the most developed knowledge and technological infrastructure, and thus are the areas where new, scientific knowledge typically is created and commercialised.
The important point behind this argument is that new companies and industries do not develop in a vacuum; they rise from existing firms and organisations, and based on existing knowledge. The clustering effect of new economy sectors relates then to the fact that important types of knowledge are generated and transmitted more efficiently as a result of local proximity. The interaction between knowledge creators, as universities and research institutes, and firms, often takes place within short geographic distances, as this kind of knowledge diffusion generally requires face-to-face contacts between persons (Tödtling and Kaufmann 1999). Spin-off firms from knowledge organisations often establish themselves near their ‘mother organisations’.
However, close distance to important knowledge organisations is not a sufficient condition to develop a new economy cluster. ’For there are many research universities, and even many that have generated lots of knowledge in semiconductors, but there is a much smaller number of Silicon Valleys and Route 128s’ (Storper 1997: 16). The relevant knowledge has to be commercialised. New commercialised innovations develop usually in collaboration between several players. In new economy sectors important players are often researchers, venture capitalists and business angles (Cooke 2002), typically found in large cities. Thus, new economy clusters often originate in metropolitan areas that contain industrial and knowledge environments able to breed firms in entirely new industries.
Table 1: Possible impacts of ICTs on regional industrial development
New scientific knowledge develops in organisations in large cities and is generated and transmitted more efficiently through local proximity
Clustering of ICT and other new economy sectors i large cities
New industrial spaces
Pioneer firms locate in ‘new’ areas in which new institutional infrastructure can be built, and new industries then cluster around themselves in some of the ‘new’ areas
Clustering of ICT and other new economy sectors mainly outside large cities
Strengthened spatial division of labour
Activities that depend on tacit knowledge, face-to-face interaction and trustful relations need to remain agglomerated
Information, knowledge and some production and services are distributed immediately and cheaply over electronic networks, reducing the ‘friction of distance’
Decentralisation of some routine production and service activities, while ‘front-end’ activities (as innovation and complex production) remain in large cities
Global division of labour orchestrated by TNCs
TNCs link up to knowledge intensive and dynamic clusters, often found in large cities, by locating some firms in cities
TNCs are using ICTs to coordinate global production chains, and are transferring some activities to low cost countries in the face of increased competition
The strengthening of some knowledge intensive clusters in large cities, whereas many activities are becoming more diffused on a global scale
ii The growth of new industrial spaces outside of large cities
The second approach postulates that new economy clusters arise outside of the traditional centres of industrial activity. Although there may be many specific explanations behind the formation of each new industrial space, the assumption is that the same common causal mechanisms have created each area. In short, it is a matter of uncertain, unstable markets together with increased competition causing vertical disintegration of production chains in a new era of 'flexible accumulation' (Scott 1988). This provides market opportunities for a number of new firms, and if there is rapid growth in the market, completely new industries can arise.
The pioneer firms in new industries may, according to Storper and Walker (1989), locate themselves in many places. This locational freedom distinguishes the second approach from the first one, in which pioneer firms establish near the sources of knowledge normally found in large cities. The locational freedom relates to the fact that the pioneer firms for a period of time may achieve high profits as they have developed or put to use radical innovations. Achieving a high profit, the pioneer firms can attract the necessary production factors at many places. The firms are able to compete for recruiting the best workers, and they are able to create an attractive local market for suppliers. In new industries the necessary know-how by firms, workers and suppliers is often created through trial and error and learning by doing. Thus, the approach postulates that some places are not have particular advantageous as regards the development of new firms in growing industries.
The locational freedom for the pioneer firms results in new industries often developing outside the traditional centres of capitalist accumulation, although not in the most peripheral parts of a country. However, the firms are often established in areas where they will not meet the same degree of resistance from workers and labour unions towards new, and more flexible working conditions than in the old industrial centres. The new industries are seen as part of a new techno-economic paradigm, which demands a new, adapted and supporting institutional infrastructure (e.g. Freemand and Loucã 2000). The institutions are seen to be easier to build in ‘new’ areas not dominated by powerful defenders of the old ways of doing things.
When the pioneer firms and a cluster start growing, the further cluster development follows the same logic in the first and second approach. The first step often involves new firm spin-offs leading to a geographical concentration of firms in nearly the same production stage. Once an agglomeration of firms becomes established, progressively more external economies are created, forming a cumulative process. The first external economies often include (i) the creation of a set of specialised suppliers and service firms, frequently originating from vertical disintegration of firms, and where firms agglomerate in order to reduce inter-firm transaction costs, and (ii) the creation of a specialised labour market (Storper and Walker 1989). New organisations that serve several firms in the growing cluster may be established, e.g. knowledge organisations, specialised education establishments and business associations. In this way new industries create localised externalities. The agglomerations themselves are considered as sources of industrial dynamics, and the regions are in particular seen as the locus of what Storper (1997) denotes as 'untraded interdependencies'; which are conventions, informal rules and habits that coordinates economic actors under the conditions of uncertainty or complexity.
iii The decentralising of some industrial activity away from large cities
The third and fourth approaches refer to the way ICTs are used in the economy and how this affects centralising and decentralising forces in the space-economy. The development of ICTs, and in particular the use of Internet, has led some commentators to proclaim the 'death of distance' (Gillespie, Richardson and Cornford 2001). At least, many kinds of economic activity are seen to be less dependent upon their location in a large city. The reduced ‘friction of distance’ relates to the possibilities of distributing information immediately and almost without costs over electronic networks. Firms may then more easily find collaborators in other places, and not least keep in current contact with these. Firms making active use of ICTs may also, it is claimed, locate in areas having lower costs, less pollution, better living conditions etc.
Some products and services may also be distributed worldwide through Internet. In many R&D-intensive companies, ‘dotcoms’ and producer service firms, in particular, the transport costs are small as Internett may be an important distribution channel. The idea is then that firms do not need to be close to customers, and also that some industrial activities do not need the good access to information, knowledge and collaborators found in large cities.
However, these arguments seem to mix up information and codified knowledge that is mass-produced and widely distributed with the more scarce resource of non-codified knowledge. This kind of knowledge includes know-how and skills that are more or less tacit, ‘sticky,’ and highly embedded in individual experience, human relations, communication channels and organisational routines. New knowledge in particular is often ‘sticky’ and includes combinations of tacit and codified knowledge. ICT may facilitate dispersal of activities that can be accomplished away from large cities, while activities that depend on tacit knowledge, face-to-face interaction and trustful relations tend to remain in cities, or in knowledge intensive clusters. The Internet allows long distance ‘conversations’ (interactive long-distance exchanges of visual and oral information) but not ‘handshakes’ (information exchanges requiring persons to be in the same physical space) (Leamer and Storper 2001).
Innovation activity is often based on combinations of codified and tacit knowledge and is in need of coordination of several, often complex, activities. The development and supply of some specialised services also need close and long-term contact between user and producer. Co-location facilitates these kind of coordination. Thus, ICTs seem to promote decentralisation of some activities (such as routine and mobile production and service activity) away from large cities, strengthening older development trends. In contrast, 'home base' activities, associated with innovation and complex production, are still prospering in large cities. This is the ‘double-edged geography of the Internet age, with its tendencies towards specialisation and agglomeration, on the one hand, and spreading out on the other’ (op. cit.: 20).
iv The inclusion of firms in evolving global networks
The fourth approach examines the way in which large corporations use ICTs in particular in co-ordinating their supply chains, which increasingly includes activities taking place in many parts of the world. The world economy is seen to develop in the direction of increasingly supranational, functional integration run by multinational companies (Dicken, Peck and Tickell 1997). Lower tariffs and non-tariff barriers, and standardisation of some products and services, have facilitated the movement of products across borders, and have increased the competition for many companies. Thus, ICTs have enabled new global divisions of labour to emerge, resembling the global corporate structures of activities, and ICTs have lessened the need for plants to be near either their corporate headquarters or their markets. It is also argued that former sheltered local and national markets are about to disappear in many industries. Firms have then a stronger pressure to reduce costs (by, for example, automatisation and the transfer of some production to low cost labour areas), and to increase their innovation activity.
Economic globalisation may, on the one hand, lead to reduced significance of regional networks and clusters compared to the linking up of firms in global value chains. On the other hand, clusters may constitute what Amin and Thrift (1992: 577) denoted some years ago as ’neo-Marshallian nodes in global networks (where they) act as ”centres of excellence” in a given industry’. Corporations link up in different ways to specialised firms and knowledge organisations in dynamic and innovative clusters, which are often to be found in large cities.
By locating some activities in large cities, or knowledge intensive clusters found elsewhere, units of TNCs may have access to a competent work force and specialised local competence, proximity to knowledge and training organisations. New ideas and economically useful knowledge can then come about through contact and co-operation between e.g. skilled workers, engineers and researchers. Corporations may tap the knowledge base of such a cluster, i.e. the knowledge intensive cluster becomes a ‘listening post’ for relying back product development and marketing information to the TNC. Such a role may be especially important in new industries and in unstable external environments as ‘when the content of knowledge is changing rapidly only those who take part in its creation can get access to it’ (Lundall and Borrás 1997: 34)5.
The theoretical overview above provides the point of departure from which to analyse potential clustering effects in the software industry in the Oslo area. All the approaches point to the continued importance of geographical proximity, localised learning, dense human relations etc. in stimulating some kind of production and innovation activity. These factors are seen to have clustering effects on new economy sectors, if not always promoting clusters in the large cities. ICTs may, however, also have decentralising effects on some activities.
‘Unpacking’ software consultancy in the Oslo region
The Norwegian software industry (NACE 722) is strongly centralised to the Oslo region. The region had 66% of all jobs in this sector in Norway in 1999, as compared to 53% of the jobs in the entire ICT sector (as defined by OECD) and 26% of all jobs6. In examining the comparatively large number of jobs in the software industry in Oslo, we first ‘dissect’ the sector. That is, we analyse the main activities that characterize the sector and how innovation, production and customer relations actually take place ‘on the ground’.
The paper distinguishes six main activities in the Oslo based software industry. Individual firms normally perform several of these activities. The activities follow a rough value chain of software production and distribution (Figure 1). The value chain model consists of activities to transform raw materials and other factors of production into refined goods and to market and distribute these to clients. The model is seen to describe very well the manufacturing of physical products, but to be less appropriate in analysing producer service activities (as software consultancy). Producer service firms often contribute to solving a unique problem for their client, and these firms are described as value shops by Stabell og Fjeldstad (1998). The activities in value shops typically consists of problem diagnosis, proposals for solution, selection of one solution, accomplishing this solution, and control of results. While the value shop describes important activities for consultancy projects, the entire software industry is well characterised by the rough value chain model in Figure 1.
Platform suppliers deliver generic technology and tool that are the basis for developing software solutions. Oracle Norge AS has, for example, three main platforms including application server, database and development tools. The platform suppliers are mainly large, global and US based corporations with daughter companies or branch offices in the Oslo area. Other parts of the software industry, which use the platform technology to develop their own products and solutions, are important customers.
The platform suppliers build their competitiveness mainly through large research and development efforts, which primarily take place in the ‘home base’ of the companies. The technology development and sale of licences (which dominates in the Norwegian branch offices of the US companies) ‘drive’ other activities. It forms the basis of training, support and consulting services.
The platform suppliers may also sell applications (software solutions) developed by use of the companies’ own technology. Thus, Oracle puts two main solutions on the market, one economy system (ERP, Enterprise Resource Planning) and one sales system (CRM, Customer Relation Management).
Figure 1: Main activities in the software industry in the Oslo area, following a rough value chain model
Running of systems
Software production consists of constructing standard solutions, within economy and account, customer relation management, logistics, case handling, portals, web applications etc., to a large number of customers, which are companies or public organisations. Some companies distribute standard software with none or little modification for individual customers. The software products of other companies require adaptation for each customer. Even if there is no or minor modifications, consulting services, as installation, integration of new solutions in the customers’ organisation, training, converting of existing data to a new software system etc., come with the deliveries. The new solutions normally require changes in ways of working, and they require organisational development. How much the consulting part amounts to vary between companies. Some companies consider that their customers use the same amount of money on installation, modification, training etc. as on purchasing of the software product itself. In other firms, the consulting services amount to twice the costs of software products.
The software products for the most part undergo a continuous development, which is based on demands and wishes from clients. The firms allocate considerable resources to development of their existing products. In several firms about 10 per cent of the turnover is used on development work.
The standard products are put up for sale through a net of branch offices and authorised distributors, which service local markets in Norway and in other countries. The distributors handle the contact with the clients as long as there are no large problems in implementing and using the software. The software producers often come in if deliveries require considerable modification of the standard solutions, or if the clients have problems that the distributors are unable to deal with.
The producers build and maintain their competitiveness by redistributing some of their turnover to product development, which most often consists of upgrading existing products. The product development takes place by use of internal competence, and often in dedicated R&D departments. In for example Webmaster Unique ASA 40 of 160 persons work in product development and product improvement, Agresso AS has a central R&D division in Oslo employing 200 persons7, while in Visma Software about 100 of 600 persons are developing and improving products. Although product development takes place by use of internal competence, signals from clients are important. In some cases, firms have dense contact with demanding, Norwegian customers, often found in industries where Norwegian firms or organisations are large and/or early users of software solutions. Development of new products is also based on interpretations of technologic and market trends. Some firms obtain knowledge about trends from large, international consulting companies, and they cooperate with platform suppliers among other things to achieve knowledge about technology development early on. Some large companies also build competitiveness through acquisitions of other firms with high-quality products.
Consulting projects require tailor-made solutions. However, the solutions are often based on generic tools and/or familiar components and knowledge, which have been developed in previous projects. The projects include development and implementation of a new software system at a client. In these cases the client cannot use a standard product, but needs a specially tailored solution. Based on successful consulting projects, firms may develop standard programmes or solutions, for example tools, routines, methods, and advisory programmes for selecting and implementing new IT solutions. Development of standard methods and solutions means converting experience based (often tacit) knowledge into explicit knowledge (Nonaka, and Reinmöller 1998). This knowledge is then used in new consulting projects, giving way to still more tacit knowledge.
Some consulting companies perform long-lasting projects for clients and act more or less as the IT division of the clients. In addition to the development of software systems, the projects may include advise in the purchasing and implementation of software products. These activities also involve analyses of labour processes, organisation of activities, reorganisation and competence needs, and preparing IT-strategies. Consulting companies and clients cooperate closely in projects, partly to tailor-make software solutions to the needs of clients, partly so that skills in running of the new system remain by the clients after implementation. Some firms work increasingly at the clients and together with their expertise. EDB Business Partner AS, for example, works as a system integrator in linking old and new software systems at clients.
The consulting firms often specialise in undertaking projects in specific industries. The most important users of software solutions are found in the public sector, bank and finance, the energy sector (oil and gas and power plants), and telecommunication. The largest software consulting company in the Nordic countries, TietoEnator AS, aims to service industries in which Nordic firms are world leaders in making use of software solutions. The company provides solutions that are specific to individual industries and customers, and which build on reusable components and good knowledge about the Nordic industries that the company has identified as particular IT-intensive.
Consulting companies gain competitive strength through their continuing building of competence and methods, through the effort of individual workers to keep their knowledge up to date, and through good routines in diffusing information and knowledge inside the companies. Contact with clients in large projects is important for learning and for the building of competence. To some extent, firms put together project teams consisting of experienced and less experienced employees. Some firms seem to have a sophisticated system for reusing knowledge and components from one project to another, in which developing and maintaining the firms’ intellectual capital are important. Firms have, for example, internal groups responsible for developing dedicated subjects, and for diffusing knowledge inside the firms through meetings and web sites. Knowledge is also obtained from external specialists, however, firms seek to develop internally what they see as their core competence and competence that are used in several projects.
Firms often have activities in several of the first three categories in Figure 1. Consulting companies, like Computas AS, have generic tools that are used by other software firms. Other firms both produce standard products and perform consulting activities. The consultants make use of the firms’ standard products, but also products from other software producers. Firms have to some extent focused on specific activities in different time periods. Some firms were founded on one innovative product or family of products. In general, these products are early in their life cycle, and firms have been unable to obtain a considerable market share when the products are becoming more mature. Larger competitors oust these small and young firms, and the firms lack resources and traditions in producing standard products for a large market. On the other hand, some consulting companies develop standard solutions and components using their consulting experience.
After sale services
Most firms perform the three last activities in Figure 1. These activities include the last part of the value chain, which consists of the distribution of products and the running of software systems. Thus, firms instruct clients in the use of both tailored solutions and many standard software programmes. Software companies also offer standard courses when new software or technology is introduced. The companies often have course and training departments that arrange coursed and workshops.
Companies also offer their clients assistance as concerns questions and problems with their software solutions, and this support has to be offered day and night. Lastly, some firms operate and manage hardware and software systems for large clients, while others run, for example, clients’ account and wage systems. A firm like EDB Business Partner AS operate the old software systems of clients for a definite time period, while the client build competence in running new systems.
The after sales services are mostly standard services requiring specific routines, as when firms run clients’ systems. However, these services also give some important information of clients’ need to be feed back into product refinement activities. Some firms (as Mogul Technology AS) also see the arrangement of external courses as important in the building of internal competence and in diffusing competence internally and externally. To arrange courses, Mogul (which has web technology as an important product) has to be in the forefront as regards new technology. Courses in new technology may most easily obtain participants in Oslo, which have much of the Norwegian IT industry, have several potential users, and have good transport communications to the rest of Norway.
The software industry in the Oslo area constitutes a regional cluster due to (i) a concentration of Norwegian software firms to Oslo, and (ii) considerable interaction between the firms. The consulting activity in particular fuels the cluster building process, as this activity is project based and involves lots of constellation building.
The clustering process rests first of all on the need of dense interaction between software consulting companies as regards large tenders and projects. Companies often have short deadlines and small budgets for preparing tenders. They actively explore possibilities of making alliances. Face-to-face meetings with potential suppliers of component and with other consulting companies are often required. Hence, being located in the same area as several potential collaborators and suppliers is advantageous in preparing tenders. The same applies to the completion of projects. To be capable of carrying out large projects, consulting companies often join together in projects and may help each other out if deadlines are short. In addition, companies may obtain special competence from one-man firms or other consulting companies. Thus, the software consulting industry is relation intensive, and geographical proximity lessens the cost of face-to-face interaction.
Another important contributor in the clustering process is the need for consulting companies to visit and influence the decision makers at large and important customers in order to gain projects. Most firms find about two thirds of their clients in the Oslo area. Companies often have lasting projects at some clients, in which several employees work on the clients’ systems and machines and at their premises. To some extent consultants ‘slide’ in at clients by arranging courses and performing small projects. Though such work consultants may get a good reputation at the client, which may lead to larger projects. In such case it is advantageous to be located in an area containing many large, potential clients. Most large consulting companies have branch offices in one or several places in Norway, which are of advantage in gaining clients in different parts of the country. At the same time, many companies merge or gain a foothold in different countries in other ways. The reason for this strategy is the fact that international companies want to meet the same consulting company in several countries.
Consulting companies cooperate with platform suppliers in some projects. The suppliers also arrange courses that are necessary to pass for consultants who want to use the suppliers’ technology in projects. The majority of the suppliers are international corporations, which generally find Oslo as the natural place to locate their Norwegian branch office. Oslo has, as said, by far the largest market for software systems.
Another cluster building factor is a large, specialised labour market in the Oslo area. Consulting companies recruit from other software firms, from the university and university colleges and from other industries, in particular from industries in which they have, or try to get, important clients. The large labour market is also making it easier to hire and fire employees. Some consulting companies cooperate with universities and research institutes in completing projects and in developing core competence. However, the knowledge infrastructure in Oslo does not seem to be an important collaborator in firms’ development activity. Some firms, however, recruit lots of candidates from the local knowledge organisations, in particular firms where the entrepreneurs comes from such an organisation.
The concentration of software companies to Oslo results in an information rich area as regards the software industry. Companies obtain much competence about technological and market trends by talking with large suppliers in Oslo, and by participating in branch forums, meetings and seminars in this area. To some extent information can be obtained through social interaction. Employees in different companies meet privately and exchange experience. A network of managers in software companies in Oslo is also said to exist. This kind of social milieus is seen to arise, and to be more fertile, in areas holding a concentration of software firms. It is advantageous to locate where action takes place, where a lot of large projects are carried out, information and knowledge are flowing, where experienced employees live etc.
Development, production and distribution of standard software also stimulate the clustering process, however, less than the consulting activity. Development of software products goes trough several phases, including collecting of demands and wishes from clients, prioritising between these demands and wishes, specifying the software solution, designing the software, programming, testing functionality and system, and then release and marketing. The mere programming can in principle take place anywhere. Most of the other phases require face-to-face contact with salesmen, consultants, distributors and pilot customers. These phases are easier to carry out in Oslo where, for example, numerous large customers are found.
Agresso AS has, for example, located an R&D division with 200 employees in Oslo. The division is not in direct contact with clients, however, Agresso has methods to register and prioritise demands and wishes from customers. The location of the R&D division in Oslo mainly reflects the fact the company was established there, that the employees live in Oslo, and that the area has a large labour market.
Software firms do not really need to be located near clients as they sell standard products or licences that they may distribute through CDs or Internet. However, software firms may use pilot customers in development projects and may have some very large customers, which they often meet face-to-face. Distribution of standard products also leads to training, consulting and support activities, which amount to a large share of the clients’ costs. Thus, firms may serve customers more economically if they are close to many of them.
Some activities in the software industry can more easily be placed outside Oslo, Visma Software AS, has for example, gathered all support activities in Fredrikstad, which is, however, within daily commuting distance from Oslo. Bringing together these activities in one place facilitates more effective routines and diffusion of competence between employees. Programming may also be easily decentralised from Oslo, provided that enough experienced employees are available in other places. Agresso AS has, for example, outsourced some standardised work to maintain existing solutions to South Africa, while Visma Software AS used to have 30 programmers in St. Petersburg, which got very specific projects.
This paper examines the reasons for the large concentration of the Norwegian software industry to the Oslo area. The empirical analyses are based on information from 14 personal interviews with firm leaders. Through these interviews we aimed to understand how the software industry works, as a necessary groundwork for a follow-up telephone survey to a larger population of firms. The interviews, however, shows the strength of the intensive research design (Sayer 1992) when it comes to realizing development processes. The paper analyses identifiable firms in their wider production system. By focusing on how firms perform individual activities, and how they interact with other players in performing the activities, the reasons for the strong concentration of software firms to the Oslo region seems to become clear. The work of ‘unpacking’ the sector ‘was like “switching the light on”’ (Sayer 1992: 284).
The concentration of software companies to Oslo rests first of all on the need of dense interaction among software consulting companies themselves, between consulting companies and decision makers and IT-personnel at important customers, and between consulting companies and platform suppliers. The fact is that consulting activity is project based and involves lots of constellation building and face-to-face contact, which is facilitated when players co-locate. Concentration is also stimulated by the large, specialised labour market in Oslo and by spill-over of knowledge in formal and informal settings.
The paper distinguishes four different approaches as regards the impacts of ICTs on regional industrial development. The first approach emphasises a strong clustering effect of the ICT sector and other ‘new economy’ sectors in large cities. This approach is a good analytical tool in interpreting the concentration of software firms in Oslo with one exception. The approach postulates that firms in new industries tend to cluster near universities and research institutes where new and important knowledge develops (Garnsey 1998). A few software firms in Oslo are established by academics, and these firms seem to uphold some contacts with academic milieus. The software industry in general seems to make little use of the knowledge infrastructure in Oslo, besides when recruiting new candidates.
The second approach postulates that new economy clusters arise outside of the large cities as pioneer firms in new sectors have a large degree of locational freedom. Such a freedom is not found in the software industry in general. This study stresses the importance of most activities in the software industry of being close to decision-makers in large public organisations and in private companies in Oslo, and for firms to base their activities upon pre-existing ensembles of software firms and knowledge. More generally, the opinion that the pioneer firms can locate almost anywhere may be greatly influenced by the ‘high mobility’ society that the inventors of the theory of new industrial spaces8 have analysed, which is the US and in particular California. It seems to be the case that specific conditions in this area are presented as a more universal theory, which means that one ‘freeze, and then present as universal, relationships which are contingent and historically specific’ (Sayer 1985: 17). Such a theorising of what may be specific developments in the US is not relevant in interpreting the locational pattern in the Norwegian software industry. The more general model of cluster development in the second approach, however, points to important mechanisms in the Oslo software industry. An important observation is how new industries create localised externalities, which causes new activity to cluster around it, and creates substantial first-mover advantages.
From a regional policy point of view, much interest has centred around the potential for ICTs to bring about the ‘death of distance’, and, hence, have a decentralising effect on the space-economy (Gillespie, Richardson and Cornford 2001). From somewhat different angels, the third and fourth approaches postulate that, in particular, routine and mobile activities may be decentralised away from large cities, while innovation activities and complex production remain. The paper has not examined the impact of ICTs on the location pattern of diverse industries, but has examined decentralising forces in the software industry, which seem to be few. Some activities that do not require face-to-face interaction with clients or external actors, such as programming and support, may be decentralised. Decentralising also occurs when Oslo based software companies establish branch offices in other parts of Norway, mainly to serve regional markets. However, keeping in mind the relation intensive character of most activities in the software industry, a large scale decentralising away from Oslo will not take place in the near future.
Amin, A. (1994), Post-Fordism: Models, Fantasies and Phantoms of Transition, in Amin, A. (Ed.), Post-Fordism. Blackwell, Oxford UK & Cambridge USA (1-39).
Amin, A. og N. Thrift (1992), Neo-Marshallian Nodes in Global Networks. International Journal of Urban and Regional Research, 16: 571-587.
Cappelen, Aa, J. Fagerberg and B. Verspagen (1999), Lack of regional convergence, in Fagerberg, J., P. Guerrieri and B. Verspagen (Eds.), The Economic Challenge for Europe. Adapting to Innovation Based Growth. Edward Elgar, Cheltenham and Massachusetts.
Cooke, P. (2002), Knowledge Economies. Clusters, learning and cooperative advantage. Routledge, London and New York.
Dicken, P., j. Peck and A. Tickell (1997), Unpacking the global, in Lee, R. and J. Wills (Eds.), Geographies of economies. Arnold, London (147-157).
European Commission (2002), Regional Clusters in Europe. Report in the framework of the Seventh Observatory of European SMEs, submitted to the Enterprise Directorate General by KPMG Special Services, EIM Business & Policy Research, and ENSR; Brussels, 2002.
Foss, O. and T. Selstad (1997), Bosettinger og næringer. In Foss, O. and T. Selstad (Eds.), Regional arbeidsdeling. Tano Aschehoug, Oslo (11-36).
Freeman, C og F. Loucã (2001), As Time Goes By. From the Industrial Revolutions to the Information Revolution. Oxford University Press.
Garnsey, E. (1998), The Genesis of the High Technology Milieu. Paper. University of Cambridge.
Gillespie, A., R. Richardson and J. Cornford (2001), Regional Development and the New Economy. EIB Papers, 6, 1: 109-131. European Investment Bank, Luxembourg.
Leamer, E. E. og Storper, M. (2001), The Economic Geography of the Internet Age. NBER Working Paper No. 8450. National Bureau of Economic Research, Cambridge, Ma.
Lundquist, K.-J. (1996), Företag, regioner och internationell konkurrens. Om regionala resursers betydelse. Meddelanden från Lunds universitets Geografiska Institutioner avhandlinger nr. 129. Lund University Press, Lund.
Lundvall, B-Å. and S. Borrás (1997), The globalising learning economy: Implications for innovation policy. Report from DG XII, Commission of the European Union.
Lundvall, B-Å og B. Johnson (1994), The Learning Economy. Journal of Industry Studies, 1: 23-42.
Markusen, A. (1999), Fuzzy concepts, scanty evidence, policy distance: The case for rigour and policy relevance in critical regional studies. Regional Studies, 33, 9: 869-884.
Nonanka, I. and P. Reinmöller (1998), The Legacy of Learning. WJB Jahrbuch 1998. Wichenshaft Zentrum Berlin.
Norton, R. and J. Rees (1979), The Product Cycle and the Spatial Decentralisation of American Manufacturing. Regional Studies, 13: 141-151.
Piore, M. J. and C. F. Sabel (1984), The Second Industrial Divide. Possibilities for prosperity. Basic Books, New York.
Porter, M. (1998), Clusters and the new economics of competition. Harvard Business Review, November - December: 77-90.
Sayer, A. (1985), Industry and space: a sympathetic critic of radical research. Environment and Planning D: Society and Space 3, 3: 3-29.
Sayer, A. (1992), Method in Social Science. A realist approach. Routledge, London and New York.
Scott, A. (1988), New Industrial Spaces. Pion, London.
Sengenberger, W. and F. Pyke (1992), Industrial districts and local economic regenration: Research and policy issues, in Pyke, F. and W. Sengenberger (Eds.), Industrial districts and local economic regeneration. International Institute for Labour Studies, Geneva (3-29).
Stabell, C. B. og Ø. Fjeldstad (1998), Configuring value for competitive advantage: On chains, shops, and networks. Strategic Management Journal, 19: 413 – 437.
Storper, M. (1997), The Regional World: Territorial Development in a Global Economy. Guilford Press, New York.
Storper, M. og R. Walker (1989), The Capitalist Imperative. Territory, Technology, and Industrial Growth. Basil Blackwell, New York.
Tödtling, F. og A. Kaufmann (1999), Innovation Systems in Regions of Europe – A Comparative Perspective. European Planning Systems, 7, 6: 699-717.