Wednesday, September 27, 2006

Concurrent workshops

• Session C-1 Understanding the impacts of innovation
• Session C-2 New indicators for science and technology policies
• Session C-3 Global knowledge flows – the spatial dimension

Session C-3 Global knowledge flows – the spatial dimension

• What are policy-relevant indicators to measure the performance of innovation clusters?
• Cross-regional and intra-sectoral analysis of clusters
• Of triple helixes, classification schemes and knowledge value chains
• Innovation at regional level: What we can learn from the CIS4 two-tiered survey in Italy

What are Policy-relevant Indicators to measure the performance of innovation clusters?

Biographical Note

Charles Davis, David Arthurs, David Wolfe, and Erin Cassidy have collaborated during the past 2 years to develop indicators and frameworks for the measurement of innovation clusters in Canada. They have measured seven of the innovation clusters that are part of the National Research Council's recent geographically focused R&D and innovation initiatives in Canada.

Charles Davis is a Senior Associate with Hickling Arthurs Low (HAL) Corporation and a Professor at Ryerson University. He received his Ph.D. from the Université de Montréal in science and technology policy and management. He has more than twenty years of work experience as a policy researcher, analyst, research program and project manager, and as a university teacher and researcher.

David Arthurs is the President of Hickling Arthurs Low (HAL) Corporation, a consultancy specializing in technology management, strategy, and economics for science-based public sector organizations. His education includes a B.A.Sc. in engineering from the University of Waterloo, an MBA from the University of Ottawa, and a Ph.D. from Queen's University where his studies focused on the economics of technological innovation.

David Wolfe is a Senior Associate with HAL and Professor of Political Science at the University of Toronto, where he is Co-Director of the Program on Globalization and Regional Innovation Systems. David is also the National Coordinator of the Innovation Systems Research Network, which is investigating the role of local and regional clusters in Canada. David holds a B.A. and an M.A. in Political Science from Carleton University and a Ph.D. from the University of Toronto.

Erin Cassidy is an evaluation officer in the Planning and Performance Management group with the National Research Council. Her education includes a B.Mus. from the University of Ottawa. Her career in the Government of Canada has spanned several departments where her work as a senior analyst has focused on evaluation and policies for industrial development, science and technology, commercialization, regional economics and the Canadian music industry.

Abstract

In the past two decades enormous scientific, practitioner, and policy literatures have appeared on innovation clusters, which we broadly define as geographically bounded concentrations of interlinked firms and related institutions.  Innovation clusters now feature centrally in contemporary thinking about meso-level social organization of innovation.  Because of the importance of innovation clusters in the economic performance of places, widespread interest has developed in understanding the genesis and dynamics of clusters and in measuring their performance.  Typically the organizational ecology of innovation clusters encompasses a range of actors, in addition to a core group of firms.  However, responsibilities for innovation cluster monitoring and development strategy often devolve to government.  This is particularly true in the case of clusters that are in an incipient state of formation around publicly funded scientific or educational institutions such as R&D labs or universities.  It is usually hoped that these organizations will serve as catalysts or magnets to induce the emergence of an innovation cluster. 

In this paper we develop a framework for innovation cluster indicators from the perspective of policy relevance. This requires clarification of three sets of issues. First, we review the theoretical grounds for policy intervention in innovation, which lie more in correction of structural or systems failure than in market failure, and we discuss systems failure in the context of cluster genesis and management. Second, we address the question of what are the underlying features of clusters that need to be measured, showing that indicators that link structure to performance are key. Third, we identify groups of policy instruments that address these failures, and we consider which innovation cluster actors have control over them.

Innovation cluster policy necessarily requires coordination of policy interventions among public actors concurrently and sequentially. Use of appropriate innovation cluster indicators can significantly improve policy coordination, which in turn can help to improve cluster performance. Our work is based on research on innovation cluster measurement and mapping that is currently being undertaken under the auspices of the National Research Council of Canada.

Cross-regional and intra-sectoral analysis of clusters

Top of Page

Biographical Note

Anne Plunket is an Associate Professor University Paris Sud 11. She is currently involved in the 6FP framework Innova-omninet studying opto-micro-nano innovative networks and clusters. She has just completed an empirical analysis of the biotechnology cluster of the region Ile de France. Currently testing the geographical, technological and scientific sources of local externalities using firm level data (see forthcoming contribution at the Fifth Proximity Congress).

Abstract

Understanding the factors of geographical agglomeration and its impact on economic performance is one of the very dynamic fields in the economics of science, technology and innovation. It is also a key issue for regional innovation policies that aim at fostering their innovation system by supporting local activities, industries/sectors, research centres and their relations. Despite major progress in this field, concepts such as (Regional innovation systems) RIS remain difficult to operationalize and in particular, it is difficult to evaluate the impact of regional innovation systems and policies on the structure and evolution of industries and their clusters.

The literature is somewhat unclear about the impact of regional innovation systems on local industries:

On the one hand, RIS literature explains the differences in innovation patterns and economic performance across regions through the interplay of local actors and institutions (Cook, 1997, Edquist, 1997, 2004). Yet, a limitation of this literature is that the economic impact of these institutional settings is difficult to be measured.

On the other hand, the literature on sectoral systems of innovation and production shows that major differences exist across sectors, but, for the same sector, these patterns are rather similar across countries (Malerba, 2002, Breschi, 2000). This similarity is explained by common features of technological regimes, knowledge base and products, that are somewhat invariant across countries. Despite the emphasis on major similarities, cross country differences appear, nevertheless, in these studies (Malerba, Orsenigo, 1996).

The question, then, is "how much the features and dynamics of the same sectoral system is similar and how much is different across countries or regions?" (Malerba, 2002, p.260).

In order to contribute to this discussion on the impact of regional innovation systems on the structure and evolution of industries, I compare the same industry (optical-electronics) across six European regions (Paris region, Baden Wurttemberg, Thüringen, Hessen, Helsinki region and Scotland).

The comparison will rely on the following methodology and data: 1) determine the specific value chain in each cluster for the opto-electronics industry. Each sector of the value chain is assessed by the number of firms and their activities (using NACE codes and regional eurostat data), 2) the technological specialization is assessed using patent application counts at the regional level (using eurostat data at the NUTS 2 level) for each sector by applying OECD concordance tables. These data are available for more than a decade which enables to compare the sectoral patterns of innovation and production. 3) the RIS's data are gathered through interviews of regional authorities and industrial cluster agencies in the six clusters (as part of the European Commission 6 FP programme INNOVA-OMNINET in opto-micro-nano innovative networks) and through eurostat data on regional education, R&D labour and research centres.

The aim is 1) to test and discuss the determinants (institutional, technological and sectoral) of significant differences (if any) in the features and evolution of each industry across regions, and 2) to infer implications for regional innovation policy.

Of triples helixes, classification schemes and knowledge value chains

Top of Page

Biographical Note

Brian Wixted is a Visiting Fellow with CPROST at Simon Fraser University in Vancouver. His doctorate examined the international linkages between industrial clusters. He has worked for the Australian Public Service on science and technology indicators (1989-1995) and agricultural science and innovation policy (1995-2000) and an Australian university (2000-2004).

Susan E. Cozzens is Professor of Public Policy and Director of the Technology Policy and Assessment Center at the Georgia Institute of Technology. Dr. Cozzens' research interests are in science, technology, and inequalities. She is actively internationally in developing methods for research assessment and science and technology indicators.

Abstract

One of the most prominent indicators of research development used in national S&T indicator scoreboard reporting is that of manufacturing industry R&D intensity. Countries that collect data by socio-economic objective can collect data on expenditure in these industry categories by research groups in universities and government, although such data is typically reported as separate statistical silos in indicator reports. Thus, despite the growing policy interest in the role of universities, business and government laboratories (the triple helix) together play in fostering the development of new products and services – current statistics can offer little in the way of information on this process. Often it seems that government and university research creates a knowledge support infrastructure that contributes to the success of industries. The food industry benefits from agricultural research usually performed in universities and government labs. Or another example, the motor vehicle industry is supported by research into fuels, infrastructure and highway management systems. However, we know little about the way such contexts develop and evolve. This paper suggests that a major challenge for the future is to make progress on frameworks that can incorporate the concept of knowledge value chains, not to compete with the existing system that avoids double counting but to complement it. Currently we do not have either the conceptual tools or the data to begin to analyse these knowledge value chains. This paper aims at clarifying the nature of the problem outlined here and suggesting a number approaches that could move us towards improving the reporting of science, technology, and research and development data.

Innovation at regional level: What we can learn from the CIS4 two-tiered survey in Italy

Top of Page

Biographical Note

Giorgio Sirilli is Research director at the National Research Council of Italy and NESTI delegate.

Giulio Perani is Senior Researcher and Head of the Unit for Innovation and R&D Statistics at the Italian National Statistical Institute; NESTI delegate.

Valeria Mastrostefano is Researcher in the Unit for Innovation and R&D Statistics at the Italian National Statistical Institute.

Abstract

At the European Union the development of a comprehensive strategy to foster innovation policies at regional level has strengthened the need for a regional breakdown of the innovation indicators produced by the Community Innovation Survey (CIS). The demand for regional innovation indicators is rising both from European regional authorities (who are increasingly taking over the responsibility for innovation policies in their regions) and from the European Commission which need stronger statistical evidence for its benchmarking exercises.

The production of CIS regional indicators has been, so far, hampered by the lack of methodological guidelines on the regionalisation of innovation surveys in the 1997 Oslo Manual.

The main problem was the definition of the statistical unit, since the adoption of the "enterprise" as the statistical unit can introduce a bias in the measurement of the technological potential of those regions hosting a large number of production units belonging to enterprises having their headquarters and R&D facilities located elsewhere. This problem was addressed in the 2005 version of the Oslo Manual. In the current Manual the adoption of a "two-tiered approach" in data collection for innovation surveys is explicitly recommended when the production of regional innovation indicators is planned.

An additional problem is the identification of the indicators which can provide a meaningful measurement of the innovation phenomena taking place at local unit level. When surveying the local units a special effort has to be made to identify the innovation activities and practices which are actually managed at local unit level in order to collect relevant data for regional analysis. Other issues dealing with the overall innovation strategies of an enterprise will have to be necessarily collected at enterprise level.

In order to pilot the production of reliable regionalised CIS statistics for Italy, ISTAT is currently (April – May 2006) carrying out a statistical survey of around 2,000 Italian multi-plant innovating enterprises in order to collect information about how they are managing their innovation projects on a multi-regional basis. This survey is connected to the Italian CIS4 as a supplementary data collection activity. Thus, the Italian CIS4 is based on a two-tiered approach which consists of treating separately:

• primary statistical units in the framework of the EU harmonised CIS4, and
• secondary statistical units (that is, local units including all the activities within a region at NUTS2 level) to be surveyed through computer-assisted personal interviews.

The paper will present the results of this two-tiered Italian CIS4 focusing on:

• methodological issues emerged in the pilot survey on multi-plant/multi-region innovators; recommendations will be made in view of future surveys;
• analysis of standard CIS indicators and of new indicators of the regional dimension of multi-plant innovating enterprises stemming from the survey; recommendations for additional indicators to be collected will be made, relying on the qualitative information provided by the survey respondents about their innovation projects;
• use of indicators for policy analysis.