The above presented different types of cluster studies are all primarily concerned with identifying industrial production clusters. The relations studied are first of all supply and demand oriented, and dynamic features of the relations are secondary to the analysis. Ongoing research carried out in relation to the Business Development Council-financed DISKO project (a study of the Danish innovation system in a comparative perspective) has its main focus on innovation-related interdependencies.
Ongoing work on identifying technological and innovative clusters at the industry level is applying a variety of methods, resulting in different types of clusters. A graph theoretical analysis of technology flows between industrial sectors, using technology indicator (R&D, patents and technical staff is used as technology indicators) weighted input-output relations in the identification of flows, is supplemented by a survey based analysis of innovation flows. The definition of a cluster in this study rests on the assumption put forward in DeBresson (1996;1996): innovative sectors cannot innovate alone, they need supplier industries for new components and user industries for new applications and requirements. The innovative clusters
Figure 3: Innovation flows in Denmark, 1990-1992.
List of sectors:
1. Food and beverages 11. Rubber and plastic 21. Raw materials/other manufacturing 2. Textile and clothing 12. Stone, clay- and glass industry 22. Public utilities (electricity, gas, heating,
water)
3. Leather 13. Iron and metal industry 23. Construction
4. Wood 14. Machines 24. Trade and repair
5. Furniture 15. Electronics 25. Hotels and restaurants
6. Paper 16. Electrical machinery and apparatus 26. Transport services etc.
7. Graphical industry 17. Office machinery and computers 27. Finance and insurance
8. Pharmaceuticals 18. Manufacturing of telecommunication 28. Public administration, defence and
equipment social insurance
9. Chemicals 19. Instruments 29. Teaching
10. Mineral oil 20. Transport (manufacturing) 30. Health and welfare institutions Sectors no. 15 and 22-30 are included as users only, sector no. 21 is only included as supplier.
are identified from “innovative activity matrices”, which are matrices inspired by input-output tables, but in stead of measuring flows of goods and services, the cells in the matrix expresses a flow of product innovations at the industry level. An analysis of this sort rests on the availability
The Danish C.I.S. survey includes only manufacturing, implying that it is only manufacturing that
10
appears as a supplier of innovations in the matrix.
of extensive survey data on innovative interdependencies. In the Danish case, data from the European Community Innovation Survey (C.I.S.), carried out in 1993 and covering innovative activity during the period 1990-1992, are used. The Danish C.I.S. questionnaire included supplementary questions on the supply of product innovations in the form of means of production, raw materials or intermediary goods to main user industries, as well as on the participation of firms in other industries in the development process. On the basis of this information a matrix of innovation flows can be constructed.10
The diagonal of figure 3 shows that intra-industrial innovation flows, measured as percentage of firms supplying innovations, are predominant. Manufacturing of electrical machinery and apparatus is a general supplier of innovations, supplying close to all other sectors in the system.
Machinery, iron- and metal industry, rubber and plastic as well as chemical industry (excl.
pharmaceuticals) are also general suppliers of innovations. The main characteristic of these industries is that the percentage of firms supplying innovation to a single industry is relatively low, which could indicate a high degree of specialisation between the firms, regarding the industries, that are main receivers of product innovations.
As a contrast to these general supplier industries we find a number of industries with a high intensity of innovation flows to few other industries. All firms in the mineral oil industry have supplied product innovations to firms in the food and beverages industry, public utilities, construction as well as transportation. All firms in manufacturing of office machinery and computers have supplied product innovations to firms in the graphical industry, while 40-60%
of the firms have supplied innovations to textiles and clothing, electronics, manufacturing of electrical machinery and apparatus, public utilities, transportation as well as intra-sectoral to other firms in manufacturing of office machinery and computers. More than 80% of the firms in paper and pulp have supplied innovations to food and beverages, while the graphical industry is a medium size receiver and pharmaceuticals, chemical industry, iron and metal industry, machinery and transportation are minor receivers. All firms in textiles and clothing are intra-sectoral suppliers of innovations. Manufacturing of telecommunication equipment has the electronic industry as a main receiver of product innovations.
The two main types of supplier industries can be used for identifying two types innovative clusters:
clusters consisting of a single industry which is a general supplier of product innovation to a broad range of receiving industries. This type of industry-cluster is important for the
Figure 4: Information flows (active participation in the innovative development process) in Denmark, 1990-1992.
diffusion of technology in the economic system;
clusters consisting of industries which are intensive suppliers to a single or few receiver(s) as well as the main receivers. In this second type of cluster, a close relation between the supplier and receiver industries is assumed in the innovative process.
Figure 4 shows the information links, measured as active participation in the development process, between industries. The “dimensions” of figure 3 are maintained, thus breaking some of the logic of a traditional input-output table: As the rows contains innovation suppliers in both figures, the flows have opposite directions. This can be illustrated by the relations between paper
List of sectors:
1. Food and beverages 11. Rubber and plastic 21. Raw materials/other manufacturing 2. Textile and clothing 12. Stone, clay- and glass industry 22. Public utilities (electricity, gas, heating,
water)
3. Leather 13. Iron and metal industry 23. Construction
4. Wood 14. Machines 24. Trade and repair
5. Furniture 15. Electronics 25. Hotels and restaurants
6. Paper 16. Electrical machinery and apparatus 26. Transport services etc.
7. Graphical industry 17. Office machinery and computers 27. Finance and insurance
8. Pharmaceuticals 18. Manufacturing of telecommunication 28. Public administration, defence and social
equipment insurance
9. Chemicals 19. Instruments 29. Teaching
10. Mineral oil 20. Transport (manufacturing) 30. Health and welfare institutions Sectors no. 15 and 22-30 are included as users only, sector no. 21 is only included as supplier.
and food & beverages. Food & beverages is an important receiver industry of product innovations from the paper industry, and between 40 and 60 percent of the firms in the paper industry indicate, that firms from the food & beverages industry have participated actively in the development process. This finding of close relations between paper and food & beverages is supported by the findings of Christensen et al.(1996) of developments of packaging playing an important role for the food and beverages industry.
The relation between paper and food & beverages is one example of the assumed relations between innovation suppliers and their most important user. The relation between manufacturing of telecommunications equipment and electronics is another example.
Identification of innovative industry clusters of the above kinds can be important supplements to the production cluster studies presented in section 2 and 3, since they can contribute with a dynamic angle which is largely missing in the production-defined clusters. In relation to the agro-industrial complex and the food and beverages resource area, figure 3 shows that it might well be that biotechnology is crucial for the development of this area, but food and beverages have a wide range of innovative sources and the industry plays a particularly important role as a user of innovations from the paper and mineral oil industries.
The identification of an innovative cluster around electronics is another important finding in relation to understanding the forces driving technology development and diffusion in the economic system.
Since the findings presented here are part of an ongoing research project, we have only scratched the surface in relations to the analytical possibilities of this kind of analysis. Thus the need for a development and improvement of both the methodology and data collection is obvious.