Innovation System Research and Policy: Where it came from and where it should go Globelics Academy Lisbon May 4, 2007 Bengt-Åke Lundvall Aalborg University and Tsinghua University
Conceptual issues: Innovation and innovation systems • Innovation is a new idea brought to the market or into the production process. • Technical innovation vs. organisational change. • Innovation as interactive process • System of innovation as constituted by interdependent organisations – focus on relationships. • National systems of innovation differ in terms of economic structure and institutional set-up. They are learning systems (including R&D-efforts).
Structure of the lecture • A brief history of innovation system research • Theoretical challenges for innovation system research • Comparing the Danish innovation system with other European systems
The message • The drift away from the original concept of NSI made it narrow and focused on science as sole source of innovation. • Strong science with weak innovation is not a paradox – reflects NSI-weakness in organisational learning and lack of market orientation • A need for research to develop theories that makes it possible for us to understand national systems as learning economies
The Original NSI-concept • Friedrich List (1840) as the Grandfather and Freeman (1982) as the Great Father. • My 1992-version was based upon collaboration in Aalborg with Andersen, Johnson, Dalum, Villumsen and Fagerberg among others. • List, Freeman and Aalborg versions were broad and linked innovation to the production system and to the organisation of firms.
Context for original NSI • List as catch-up analyst and as involved in policy advice - Freeman as modern List • The Ad Hoc OECD-group on Science technology and international competitiveness 1981-83. • In common for Freeman and myself was a criticism of competitiveness strategy based on low wages – ’structural competitiveness’ • The OECD TEP-program 1989-92 diffused the NSI-idea.
Brief history of innovation research – antecedents to the NSI-concept • Adam Smith on the role of both experience-based and science based learning as the basis for innovation. And on the importance of vertical division of labour for wealth of nations. • Friedrich List on the need for govenrmental infrastructural investment to build national innovation systems • Schumpeter and Schmookler and debate on Supply- or Demand-driven innovation. Modern view of innovation as an interactive learning process.
Christopher Freeman: The father of modern innovation theory • Economist from Cambridge – went to Keynes’ lectures, read Marx and Schumpeter. • Among Freeman’s favourite themes beginning of the 80’s were: • The need to overcome the split between innovation as driven by supply factors versus innovation as driven by demand factors. • The importance of understanding interaction between agents in the innovation process • 1982 Freeman introduced the concept ’national system of innovation’ in unpublished paper for OECD-group.
STI-mode and DUI-mode of learning – getting the NSI-concept back on track • STI=Science-Technology-Innovation mode is characterised by science-approach – formalisation, explicitation and codification • DUI=Learning by Doing, Using and Interacting mode refers to experience-based, implicit, embedded and embodied knowledge. • Jensen, Johnson, Lorenz and Lundvall, ’Forms of Knowledge and Modes of Innovation’, Research Policy, 2007
The paradox and the built in STI-bias The Paradox: ’Systems with a lot of good domestic science but less successful in innovation’ • Reflects the limited perspective with too much focus on Science based learning (STI) to the neglect of Experience based learning (DUI). • Reasons for bias: • STI-learning can be measured and manipulated more easily than DUI-learning – cf. the Meadow-project in Europe. • The policies involved are less controversial – cf. The LOK-program in Denmark.
The double change in context • ICT and access to elements from the science base becomes increasingly important for firms in all sectors – calls for a strengthening of STI-mode of learning • But these changes and globalisation contribute to accelerating change and requires learning organisations – calls for a strengthening of DUI-mode of learning
Illustrating empirically how DUI and STI-learning promote innovation • Year 2001, DISKO survey on technical and organisational change addressed to Danish firms in the private sector,. • Survey and register data from 692 firms included in the following analysis. • Jensen, Johnson, Lorenz and Lundvall in Research Policy 2007.
DUI-learning - seven indicators reflecting ’learning organisation’ and ’user focus’ • The firm makes use of some of the following practises: • Interdisciplinary workgroups • Quality circles/groups • Systems for collecting employee proposals from employees • Autonomous groups • Integration of functions • Demarcations between groups of employees have become less sharp 1998-2000. • The firm has established closer relationships with customers 1998-2000.
STI-learning – three indicators reflecting R&D-effort and networking with science infrastructure • The firm has positive expenditure on R&D. • The firm has personnel with academic degree in natural science or engineering. • The firm interacts with researchers attached to universities or other science institutes.
Odds ratio estimates (with control variables for sector, size & ownership)
On the need to combine science- with experience-based learning • Firms combining science-based (STI-mode) with experience-based (DUI-mode) learning are more innovative than firms biased toward one mode. • Calls for analytical efforts that establish the connection between knowledge creation through research and knowledge creation through organisational learning and inbteraction with users. • Implies broad definitions of innovation systems, innovation policy and knowledge management.
Two kinds of bias in industrial policy • Promoting the science base of high-tech firms assuming DUI takes care of itself • SME policies sometimes neglects the importance of linkages to sources of codified knowledge The big challenge lies in stimulating firms to combine the DUI- and the STI-mode.
Similar tension in the knowledge management literature reflects different emphasis on different types of knowledge • Implicit/tacit………..Explicit Knowledge • Know how………….Know why • Skills………………..Information • Intuition……………..Analytical skill • Experience based……..Science based • Local……………………Global • Most economically relevant knowledge mix elements placed differently on these scales
The national context affects what is good practise knowledge management • Education and labour markets differ (Lam –Lundvall paper) • The mode of learning in firms differs across countries (Lorenz – Lundvall-Valeyre paper for conference) – affects what is going on inside firms • Social capital and networking opportunities differ- networks and alliances show different patterns.
Cluster analysis of how people learn in different parts of Europe • Based on household survey in 15 European countries • Restricted to 8000 people who work in the private sector in firms with more than 20 employees. • Emphasis on the degree of independent problem-solving and learning at the workplace. • Links DUI-learning to innovation. • Work in progress: But see Lorenz and Valeyre in Lorenz and Lundvall (eds), How Europe’s economies learn, 2006
The four clusters • Discretionary learning • A lot of learning, complex tasks and delegation of responsibility for quality • Lean production • Job rotation, team work and quality control but little discretion • Taylorism • No problem solving, no autonomy • Simple production • Little learning but some discretion and problem-solving
Results: International diffusion – after correcting for sector and job function • Discretionary learning and lean production in Nordic countries and Netherlands • Little DL and a lot of Lean production in UK, Ireland and Spain • Taylorism and simple production in Portugal, Greece and Italy. • Germany and France in between 1 and 2 above.
Conclusion • In order to explain how new ideas are brought to the market and transformed into economic performance it is necessary to take into account how learning takes place in working life. • National systems of work organisation and learning are dramatically different. • NSI is a useful perspective also for microstudies of specific firms. In spite of globalisation the management challenge is nation specific.