Mapping the Innovation Universe in Europe
The data from the CIS1 pilot project demonstrate the importance of taking representative samples (see 'Teething Problems' ). This is being addressed in CIS2, the follow-up exercise.
Technological innovation is essential if companies - and countries - are to maintain their competitiveness in global markets.
But where and how does innovation actually occur? How big and how rapid are its impacts on turnover, export sales, and employment? What are its costs? What are its key mechanisms and constraints? And how widely do these factors vary between regions and industrial activities? (1)
The effectiveness of measures to stimulate or facilitate innovation depends to a very large extent on the ability of policy-makers to answer these kinds of question. Without a detailed understanding of innovation's complex processes, resources may not be directed where they will do most good. Until recently, however, both policy-makers and business and investment strategists were forced to rely on largely theoretical studies, supported by small-scale or at best national surveys of enterprises.
In the US, Japan and Southeast Asia, this remains the case. But in Europe the Commission, in collaboration with EU Member States, has embarked on a unique and ambitious project to gather data from enterprises in every region of the Union, in a common Community Innovation Survey (CIS). A pilot survey, CIS1, was conducted in 1992/93. A second, improved in the light of the pilot experience, is currently being carried out, and should be completed by mid-1998. Thereafter, the Commission intends to conduct the CIS on a regular basis, once every two or three years.
Policy-makers across Europe at all levels - EU, national, regional and local - will in future be able to base their deliberations on robust and comprehensive evidence.
In the Netherlands, a higher proportion of R&D is carried out in universities and institutes than in any other Member State. But its Ministry of Economic Affairs is keen to improve the efficiency of institutional-industrial technology transfer. Like Germany, it has for some time conducted its own national innovation survey, and has therefore had better data than most Member States. But before CIS1, inter-country comparisons were only possible at sector or macro levels.
"We wanted to know what type of innovation support structures and interactions had worked well elsewhere, in order to improve our own policy instruments," says the Ministry's Dr Theo Roelandt. "What makes the CIS database so valuable is that it enables us to compare the performance of our own structures and those of other countries, using firm-level data. That is something we have never been able to do before.
"Innovation theory told us that the interactions between a country's public knowledge infrastructure and its firms and knowledge suppliers is critical to its industrial performance. But this was an assumption rather than a fact. The actual data was scarce. The CIS confirmed the theory, and gave us a better understanding of the processes involved. The data cannot tell you exactly what to do, of course. But it stimulates constructive policy debate by highlighting areas of under-performance where policy intervention is called for, and by suggesting hypotheses which can then be researched in more detail on a case-study basis."
The larger the firm, the more likely it is to be involved in collaborative R&D projects.
The problems are of three kinds. First, which enterprises should be invited to participate? It is clearly important that firms of all sizes and from all sectors, both innovators and non-innovators, are represented in the sample. In CIS 1, for example, some countries focused on known innovative companies, which skewed their results upwards (see figure ).
Second, what data should be collected, and in particular what financial data? Georg Licht of the Mannheim-based Centre for European Economic Research (ZEW) is responsible for the collection of CIS data in Germany. He admits that questions about innovation costs are time-consuming to answer, because these costs are not shown directly in a firm's accounts. But he says that firms themselves are acutely aware that their answers to simple multiple choice questions about their innovation activities are almost meaningless unless these are accompanied by detailed financial information.
Professor Sergio Cesaratto of the University of Rome, who used CIS data to carry out a detailed study of the impacts of innovation on employment in Italy (EIMS publication 37), believes that some financial data could more easily be acquired from accounts submitted to national authorities. "The quality of the data would be improved if the CIS questionnaire omitted questions which firms have answered elsewhere," he says.
Georg Licht disagrees. "The survey does add to an already heavy administrative burden on business," he says. "But in most Member States data protection laws would prevent us from matching CIS data with that gathered for other purposes. It is essential that the CIS includes both financial and innovation data."
The third problem concerns the actual wording of the CIS questionnaire. "Take the word 'technology'", says Licht. "Not only does it have quite different meanings in German and in English. But different firms, particularly in the service sector, are likely to interpret it in different ways. For example, do a software house's products constitute 'technological innovation' or not? We should avoid a terminology which is oriented towards the manufacturing sector."
The Innovation Programme's European Innovation Monitoring System (EIMS), which with Eurostat co-ordinates the CIS, acknowledges the shortcomings of the first survey. But it is confident that a firm foundation has been created for a European resource of steadily increasing value.
The selection of indicators in consultation with the OECD, the development of the database itself, the use of micro-aggregation as a means of anonymising enterprise-level data to safeguard commercial confidentiality, and the establishment of working methods by the EIMS, Eurostat and national statistical offices - these are solid achievements, and the difficulties revealed by the first survey are largely being addressed in CIS2.
(1) The charts in this article illustrate the range and power of the CIS data, but are based on individual researchers' interpretations. Inter-country comparisons, in particular, should be viewed as indicative only.
Context: The CIS: Facts and Figures
What is the Community Innovation Survey?
Who operates and funds the CIS?
How is data collected?
Data on an enterprise's innovation activities could be commercially sensitive. How is confidentiality protected?
How is CIS data used?
Do participating enterprises themselves benefit?
Are the findings of studies based on the CIS data available to anyone?
CIS1 was a pilot. What lessons have been drawn from that experience?
How will these lessons be applied in the future?