TECHNOPOLIS. An International review of Competence Centre Programmes. Erik Arnold Jasper Deuten Jan-Frens van Giessel

Transcription

1 TECHNOPOLIS An Internationa review of Competence Centre Programmes Erik Arnod Jasper Deuten Jan-Frens van Giesse Apri 2004

2 An Internationa Review of Competence Centre Programmes Erik Arnod Jasper Deuten Jan-Frens van Giesse Technopois Group Apri Introduction 1 2 What are Competence Centres? 2 3 Competence Centres in Theory 4 4 Some Competence Centre Programmes in Practice Engineering Research Centres, USA Co-operative Research Centres (CRCs), Austraia Kpus Centres, Austria 10 5 How Competence Centre Programmes Operate Objectives Operations Assessment criteria 14 6 Success Factors 16 7 Effects of Competence Centres: What Does the Taxpayer Get for the Money? 19 8 Concusions 23

3 1 Introduction Competence centres are a comparativey new form of university-industry research aiance that do both fairy fundamenta but aso more appied, probem-oriented research. Their ong-term nature and the comparativey high rates of subsidy invoved aow them to have a structuring effect on sub-systems of innovation, educating and generating communities of research practice between researchperforming institutions and industry and generating common, use-oriented research agendas with potentia to have significant positive socio-economic effects. RCN is considering how to create centres of innovation exceence that can expoit strong academic and institute research capabiities to buid custers of innovation. The Counci therefore asked us to make a survey of competence centre programmes internationay. In agreement with RCN, we seected eight competence centre programmes to study. For each, we buit a description using secondary sources, which we sent to the responsibe programme manager for comment. Obtaining the managers comments aso aowed us to interview them about some of the more soft factors invoved in defining and running such programmes. This paper summarises essons from the revised programme summaries. There are growing numbers of internationa exampes 1 of competence centre programmes (Exhibit 1). A share an intention to change research structures permanenty, by providing a ong-term but nonetheess time-imited financia and organisationa impuse. Exhibit 1 Competence Centre Programmes Country Start Date Agency Competence Centre Programme USA 1985 Nationa Science Foundation Engineering Research Centres Ireand 1988 EOLAS/Forfás Programmes in Advanced Technoogy Canada 1989 NSERC, CHIR, SSHRC Networks of Centres of Exceence Austraia 1990 Ministry of Industry Co-operative Research Centres Sweden 1994 NUTEK/STEM/VINNOVA Competence Centres Netherands 1997 Ministries OCW and EZ Top Technoogica Institutes Austria 1999 BMVIT/TiG Kpus Austria 1999 BMWA/FFF Kind, Knet Hungary 2000 Ministry of Education KKK Co-operative Research Centres Estonia 2004 Ministry of Industry Competence Centres This paper is based on a survey of a of these except the ones in Ireand and Estonia. We omitted these because the Irish case is rather od, very compex and invoves many eements that appear specific to Ireand s state of deveopment in the eary 1990s. The Estonian centres were set up at the same time as we began work on this study, so there is so far no rea experience from which to earn. Other important inputs to this paper incude work for VINNOVA and STEM over the past year, 1 Note that the we-known German Kompetenzzentren concept is a different kind of initiative, invoving regiona custers that incude a research eement together with active network management, rather than setting up new institutiona entities. 1

4 aiming to improve our understanding of the impacts of the Swedish competence centres, the work of the MAP networks on compex R&D programme initiatives, and participation in the mid-term evauation of the Austrian Kpus centres over the past three years, as we as a reading of reevant evauations. 2 What are Competence Centres? Competence centres use a combination of academic exceence with industria needs and probems to focus joint academic industry R&D on areas of high innovation potentia. As poicy instruments, they typicay focus on comparativey highcapabiity industria and academic participants: typicay medium-arge companies and high-capabiity SMEs (Exhibit 2, which is of course a simpification) though there is scope for a fair amount of variation. Exhibit 2 Focus of Competence Centres Minimum capabiity required to participate Linked and Biatera Support Competence Centres Pre-competitive Coaborative Increasingy fundamenta research Increasing Neson-Arrow market faiure As one moves North-East in the Exhibit, R&D tackes increasingy fundamenta questions and the traditiona or Neson-Arrow market faiure 2 (that companies under-invest in research) comes more and more into pay. Participants need increasing capabiities or absorptive capacity in order to make use of the resuts of R&D. As one moves South-West in the Exhibit, market faiure decines, as does the need for interna technoogica capabiities. This is typicay refected in the eves of subsidy provided in state programmes. Linked and biatera support (such as brukerstyrt forskning) tends to get 25-40%; pre-competitive coaborative work (such as the Framework Programmes) 40 50%; whie competence centres tend to be in the range 50 70%. Each of these instruments serves a somewhat different purpose, not east because it tackes a different segment of need. The idea of fundamenta research in the context of industria innovation can sound paradoxica, because the basic science obby ikes to define basic as bue skies, 2 Ken Arrow, Economic Wefare and the Aocation of Resources for Invention, in Richard Neson (Ed.) The Rate and Direction of Inventive Activity, Princeton University Press, 1962; see aso Richard Neson, The simpe economics of basic scientific research, Journa of Poitica Economy, 1959, vo 67, pp

5 curiosity-driven or research-driven research. But, as Donad Stokes 3 points out in his 1997 book Pasteur s Quadrant, much fundamenta science is actuay done with considerations of use in mind (Exhibit 3). His exampes of the curiosity-driven Bohr, of Pasteur (doing fundamenta research in order to understand and contro disease) and Edison (with his ruthess empiricism, not much interested in underying mechanisms) give a better sense of how research actuay operates. The instruments in Exhibit 2 operate mosty in Edison s quadrant. They can have adventures in Pasteur s quadrant, but competence centres systematicay devote more of their resources Pasteur s Quadrant it than do the other instruments. Exhibit 3 Types of Research, According to Stokes Yes Pure basic research (Bohr) Use inspired basic research (Pasteur) Quest for fundamenta understanding No Pure appied research (Edison) No Considerations of use Yes Competence centres have some recognisaby specia features reating to their roe, especiay They are normay funded by three partners: industry, university and a state agency. They are intended to have an effect on university resource aocation and strategy, in addition to reinforcing university-industry inks They invove ong term contractua arrangements, requiring a much bigger commitment than traditiona project by project funding of coaborative R&D They create new on-campus structures, and therefore make new organisationa and structura demands on the universities They are interdiscipinary and generay probem-focused in the research they do, demanding horizonta networking across traditiona university structures Their ong-term presence on campus and their engagement with postgraduate education draws them into coser contact and co-operation with universities core business of education and research than is often the case with inkage actions, which tend to focus more purey on research By drawing industry personne onto campus to join in research, they aso extend academics networks into the industria research community 3 Donad Stokes, Pasteur s Quadrant: Basic Science and Technoogica Innovation, Washington DC: The Brookings Institution,

6 The prototype was the US Nationa Science Foundation s Engineering Research Centre programme, aunched in Whie competence centres are normay categorised as R&D funding instruments that aim to improve academic-industry inkages, they generay aso have a ess ceary stated ambition to ater the research cuture of the universities: moving towards greater interdiscipinarity; and making cose co-operation with industry more acceptabe. The extent to which competence centres are needed, as change agents within the university system, naturay varies, not east because there are many different kinds of university. Nonetheess, this aim to change university-industry interaction by changing the way the universities work is something that most ceary distinguishes competence centres from other inkage initiatives, which tend to take university norms and cuture as given. 3 Competence Centres in Theory Competence centres operate in a context of significant change, not ony in the way the knowedge and innovation system works, but especiay in our understanding of it. They conform with our essentiay systemic understanding of research and innovation, often discussed under the sogan of Nationa Innovation Systems. In part this seems to represent an improvement in theory: a better way to describe what aready exists. But there is aso some evidence that research and innovation operate in ways that are increasingy systemic: that there is aso a quaitative change in reaity in progress. There is evidence for this, for exampe, through the increasing citation of academic research resuts in industria patents. A particuary interesting expression of the systemic nature of innovation systems is the emergence of so-caed Knowedge Vaue Coectives (KVCs) 4 essentiay the set of actors which defines, creates and uses reated knowedge. Competence centres represent contractua, institutiona reations among some of the actors within individua KVCs. The systems view of innovation 5 is based on ideas in evoutionary economics that have been emerging over the past few decades. These see the firm as constanty chaenged to innovate by externa and interna changes. There is no economic equiibrium, but a constant search for ways to create advantage. Competence centres respond to this need constanty to search for knowedge that can be used in change. The baance of knowedge production is shifting towards what Gibbons 6 et a. Caed Mode 2 : namey, interdiscipinary, probem-focused and heterogenous in its performance, as opposed to the discipinary focus of traditiona universities. Such interdiscipinary, probem-oriented research is precisey the sphere of the centres Barry Bozeman and Juan Rogers, A churn mode of scientific knowedge: Internet researchers as a knowedge vaue coective, Research Poicy, Vo 31, 2002, pp See Christopher Freeman, Technoogy Poicy and Economic Performance: Lessons from Japan, London: Frances Pinter, 1987; Bengt-Åke Lundva, Nationa Systems of Innovation: Towards a Theory of Innovation and Interactive Learning, London: Pinter, 1992; RR Neson, Nationa Innovation Systems, New York: Oxford University Press, 1993 Michae Gibbons, Camia Limoges, Hega Nowotny, Schwartzman, S., Scott P. and Trow, M., The New Production of Knowedge, London: Sage,

7 One crude indicator of the growth of Mode 2 is the growing proportion of the nationa R&D effort financed and conducted by the private sector. The emergence of mass education since the mid-20 th century means not ony that there are arge numbers of research-capabe peope working outside the knowedge infrastructure but aso that companies have many more options about where to do R&D and that they increasingy have the capabiity to do research in partnership with universities. The economicay driving technoogies are increasingy to be found at the boundaries between discipines (sometimes 7 referred to as the hyphen technoogies ), not in what the universities woud traditionay see as the discipinary mainstream of knowedge production. Whie there are changes in the mode of knowedge production, peope continue to pay a centra roe not ony in the generation but aso in the movement and expoitation of knowedge. Competence centres tend to have an important roe in post-graduate training. In what ooks initiay ike a paradox, the major parts of industry that depend on, and compete in, technoogy-intensive products and services, there is a movement away from conducting fundamenta research in-house 8. The reasons appear to be party the traditiona probem that it is hard to appropriate fundamenta research resuts, so the economic incentives to doing it are ow. Party, aso, the breadth of the fundamenta knowedge needed by industry seems to be growing, so that companies need new strategies for accessing word knowedge. Competence centres provide contexts in which companies can infuence the direction and content of research that, in the medium term, they need, but that is more fundamenta than they can afford to support. Gobaisation of industry has been accompanied by a wiingness to conduct R&D in mutipe ocations primariy within the Triad of the USA, Japan and Europe. As a resut of the growing interaction between arge, mutinationa companies and the research and higher education sector, the shape and quaity of the knowedge infrastructure becomes one of the factors infuencing industria R&D ocation. Competence centres can pay important roes in ensuring not ony that the research performed in the knowedge infrastructure has, over time, user reevance, but aso that it is of high quaity. These are necessary, but aone not sufficient, aspects of attracting and retaining industria R&D activity. In the context of a changing socia contract between science and society, many research funders are giving priority to funding reevant work. Socia goas for both research and education are bringing universities increasingy under pressure to change. At the same time, many of them have weak strategic capabiities and resources, and are strongy ocked in to existing activities and trajectories by the way they are governed 7 8 Foowing Frieder Meyer-Krahmer Frieder Meyer-Krahmer and Guido Reger, New perspectives on the innovation strategies of mutinationa enterprises: essons for technoogy poicy in Europe, Research Poicy, Vo. 28, No. 7,

8 4 Some Competence Centre Programmes in Practice In this section, we sketch three exampes of competence centre programmes, incuding the NSF one that has functioned as a prototype for the competence centres movement and the Kpus programme in Austria that appies the mode in an economy that is in important respects not very different from that of Norway. 4.1 Engineering Research Centres, USA The US Engineering Research Centres programme that started the competence centres movement predates the emergence of the Nationa Innovation Systems iterature, but this new mode of creating a common network of companies and academics fits very we with the new way of ooking at innovation. With a budget of some $50m per year, it is sti the second-argest competence centres programme, of which we are aware. The ERC programme was set up in 1985 to deveop a government-industryuniversity partnership to strengthen the competitive position of U.S. firms in word trade. More specificay, the ERC programme grew out of concern expressed in the eary 1980s by the Nationa Academy of Engineering (NAE) and NSF that: (1) rapid technoogica advances were occurring at the intersection of engineering and other discipines, requiring a cross-discipinary approach to engineering that had not been incorporated into engineering research or practice; and (2) a mismatch had deveoped between the way engineering was carried out in industry and the way students were being trained. The ERC mission has three main eements 9 Cross-discipinary and Systems-oriented Research Education and Outreach Industria Coaboration and Technoogy Transfer Individua ERCs are required to have the foowing features A strategic vision guiding both the production of advances in a compex, nextgeneration engineered system and the creation of a new generation of engineers needed to strengthen the competitive position of U.S. industry in a goba economy A dynamic, evoutionary strategic research pan to focus the ERC on achieving its vision A cross-discipinary research programme promoting the synthesis of engineering, science and other discipines that spans the continuum from discovery to proofof-concept in test beds and invoves undergraduate and graduate students in research teams An active, ong-term partnership with industry and practitioners in panning, research, and education to achieve a more effective fow of knowedge into innovation and the education of a new breed of engineers 9 6

9 An education program for undergraduate and graduate students that produces an integrative, systems-oriented inteectua environment and corresponding curricuum innovations Outreach to other institutions to enhance the capacity of the ERC to achieve its goas and broaden the impact of the ERC cuture in academe and society NSF's support across a ERCs ranges from $1.0 M to $3.6 M per year and per centre; the average ERC award was $2.5 M per year in The actua funding eve in any given year wi depend upon a detaied anaysis of proposed work, progress to date, financia need, and the avaiabiity of funds. In 2002, NSF was supporting 19 ERCs, pursuing research foci in: bioengineering; design and manufacturing; earthquake engineering; and microeectronic systems and information technoogy. As of January 2002, 13 ERCs were said to be sefsustaining after the concusion of NSF support. Currenty there are some 20 centres. Each invoves severa tens of facuty and an industria consortium. Generay, the companies are big and have high technoogica capabiities. Exhibit 4 gives two exampes of industria participations in ERC centres. Exhibit 4 Exampes of ERC Industria Consortia Centre for Neuromorphic Systems Engineering at CaTech Boeing Digita Persona 7 Centre for Subsurface Sensing and Imaging Systems, NorthEastern University Raytheon Air Force Office of Scientific Research (AFOSR) CardioMag Imaging Hewett-Packard Company Lockheed Martin Mercury Computer Systems Genera Motors Honeywe Inte IRIS (Internationa Remote Imaging Systems, Inc.) Rockwe Int. The MathWorks MicroBrightFied Jamie Whitten Nationa Centre for Physica Acoustics Nationa Geospatia-Inteigence Agency. Textron Systems TransTech Systems, Inc Zomega Technoogy Corporation Source: Centre web sites The industria consortia comprise a mixture of very arge companies and sma to medium sized technoogy-based firms. In European terms, many of these sma firms are rather big, refecting the fact that in the USA it is possibe to grow a new technoogy-based company very quicky indeed, compared with our norma European experience. A competition for new ERCs is now hed periodicay, usuay every two years. ERCs are estabished by NSF as a resut of peer-reviewed competitions preceded by programme announcements Each new centre receives a five-year cooperative agreement stipuating that a

10 forma review (incuding by externa site visitors) be conducted in the third year for renewa of support. Centres passing the third-year review receive a new five-year cooperative agreement requiring a forma review in each centre's sixth year. The fu term of an ERC is 11 years. As a centre begins its seventh year, if not before, it shoud begin to generate even higher eves of non-nsf funding to ay down a pathway to sef-sufficiency after year 11. Funding is phased down in an ERC's ninth and tenth years. The centre is expected to preserve the ERC cuture as it moves into sef-sufficiency. Sub-groups within an ERC may choose to regroup with others to compete to form a new ERC in the fina two years of the ERC's award. 4.2 Co-operative Research Centres (CRCs), Austraia The Austraian Cooperative Research centre scheme is the argest competence centre programme considered here in terms both of overa budget (about $120m per year) and number of centres (71). The individua centres can therefore be comparativey sma. The programme was set up in 1990 in response to perceived weaknesses in the nationa Innovation System 10 Austraia s combined S&T resources were dispersed both geographicay and institutionay. This separation made it difficut to buid strong research teams and ed to unnecessary dupication of faciities, and difficuty in ensuring that they were word cass. There was a ack of critica mass due to this dispersion In addition, there were chaenges of effective internationa inks for a country isoated from the internationa centres of research and innovation Existing funding arrangements were not suited to buid arge integrated mutidiscipinary research. (Most research funding in Austraia was from institutiona sources and was distributed through administrative channes to operationa units and individua researchers) Links between research organisations and users were weak. In fact, there were disincentives to coaboration among research providers and Austraian businesses The absorptive capacity of Austraian industry was imited There was a ack of mobiity of personne between government research, academia and industry Graduate programmes in Austraian Universities did not prepare students we for jobs outside the academic word as we as denying students access to the skis and experience of many of Austraia s best researchers, and researchers the stimuus of interaction with students The (officia) objectives (in 2002) for the CRC Programme are To enhance the contribution of ong-term scientific and technoogica research and innovation to Austraia s sustainabe economic and socia deveopment To enhance the transfer of research outputs into commercia or other outcomes of economic, environmenta or socia benefit to Austraia To enhance the vaue to Austraia of graduate researchers 10 Howard Partners, Evauation of the Co-operative Research Centres Programme, Canberra: Department of Education, Science and Training,

11 To enhance coaboration among researchers, between researchers and industry or other users, and to improve efficiency in the use of inteectua and other research resources In addition to the officia objectives, the CRC Programme aso addressed the unofficia goa of changing Austraia s research and innovation cuture. Unofficiay, CRC aso stands for changing research cuture. Indeed, the emergence of pubic-private research partnerships refects a fundamenta change in the way in which knowedge is generated and appied as we as changes in approaches to the management of industria R&D. Research commerciaisation has come into prominence. The CRC Programme has been an important contributor to buiding the required capacity to carry out partnership-based research and innovation, business deveopment based on research commerciaisation, and for scientists to engage in pubic programme design and deivery. As Exhibit 5 iustrates, there is a wide diversity of types of company invoved, from arge and highy capabe companies through sma, technoogy-based firms to farmers co-operatives, which traditionay have ony modest technoogica capabiities. Ceary, the threshod for entry into this scheme is much ower than in the US case. Exhibit 5 Exampes of CRC Industria Membership CRC for Advanced Composite Structures Core Members: Hawker de Haviand Aerospace P/L Supporting Members: GKN Aerospace Engineering Services Pty Ltd MSC Software Austraia Pty Ltd Pacific Engineering Systems Internationa Pty Ltd Associate Members: Austraian Urethane & Styrene Pty. Ltd Pacific Composites Pty Ltd Structura Monitoring Systems Ltd. Internationa Coaborators: Office of Nava Research, USA Airbus, Deutschand GmbH Airbus UK Ltd Source: Centre web sites CRC for Innovative Dairy Products Industry members: Austraian Dairy Farmers Ltd Dairy Austraia Commercia members: Dairy Farmers Cooperative Genetics Austraia Probio Inc Tatura Mik Industries Ltd The CRC programme operates with a series of cas for proposas atogether 8 Cas have been issued in the period 1990 to Centre contracts are normay for 7 years, though an eighth may be granted in order to wind up a centre. There is a performance review after 3 years, on which the baance of the funding depends. At the atest during the 6 th year of operations, each centre must have deveoped a pan wind up its activities. Participants may appy again in a new CRC, but the od one wi not be extended. 9

12 4.3 Kpus Centres, Austria Austria has many important simiarities with Norway, incuding an industry structure that is not especiay R&D-intensive, a weak tradition of university-industry cooperation and a rather strong appied research institute sector. In order to reach the Lisbon/Barceona target (3% of GDP aocated to RTD) increasing the investment in R&D aone is not sufficient. Additiona pubic and private investments in R&D need to generate sufficient high returns. This requires enhancing the overa efficiency of the Austrian Nationa Innovation System by Encouraging existing firms to engage in more radica types of innovation. Promoting technoogy based start-ups Increasing the roe of the higher education sector in providing a research base which can be utiised co-operativey with industry The main reason to aunch the Kpus programme was the ow eve of scienceindustry co-operation in Austria. Kpus aims at bridging the gap between fundamenta research carried out by universities and industria R&D. Other reasons to aunch the Kpus programme were the foowing deficiencies in the Austrian innovation system Short term R&D panning in industry Dominance of SMEs in R&D in Austria Lack of critica mass within the knowedge infrastructure, especiay the universities Low internationa visibiity of many Austrian R&D capabiities The main goa of the programme is to perform research that is highy reevant for both the academic word and industry and to deveop human capita in areas that are either muti-discipinary or which are reevant for a number of sectors/companies in Austria. Other goas are to Improve ong-term co-operation between science and industry Improve transfer of know-how Define new areas of research through bottom-up approaches Reaching focus and critica mass in research. Use pubic funding to trigger additiona private/industria expenditures Ensure internationay competitive quaity of K pus-centres through a strict seection process and periodic evauation. Create exampes of good practice in research management As in Austraia, an important hidden agenda of the programme was to change the research cuture and attitudes of both the industria and the scientific actors. The Kpus programme was designed on the basis of extensive research into possibe foreign exempars There are currenty 18 Kpus centres in Austria, which have been aunched in groups of 6 foowing three cas for proposas starting in The annua centra 10

13 government spend on the programme is about Euro 10m. Some Austrian Länder cofinance centres As Exhibit 6 shows, in the Austrian context there is a range of industria participation, from mutinationas through major Austrian companies to mediumsized forms and some start-ups in some cases as sma as one-person operations. A minority of the centres is muti-site. Exhibit 6 Exampes of Kpus Centre Industria Membership Light Metas Research Centre, Ranshofen AuLight Eckhart Granues Fronius Foseco Hütte-Kein-Reichenbach Linde Magna Steyr Rauch Stofig SAG UBE TCG Unitech Source: Kpus centres Competence Centre for Wood Composites and Wood Chemistry, Linz & St Veit Agroinz Meamin Internationa Dynea Austria Fritz Egger GmbH Funder Industrie Kooperationsabommen Forst Patte Papier Lenzig AG Kpus centres have a panned ife of 7 years. Many were evauated by a sma peer committee after 2 years, to ensure that conditions aid down at the time of origina funding were being fufied. In their 4 th year, the centres are evauated again by a visiting pane, to decide whether they shoud be funded for the second haf of their panned ives. One of the 12 to be evauated so far at mid-term, one has been wound down (over a period of time, to protect the post-graduates). The major uncertainty is the viabiity of the centres in year 8 and onwards, at which point they need to find aternatives to the Kpus core funding. In severa cases, the Land has stepped in to take over part of this responsibiity. There is growing consensus that the centres wi not survive in their current roe of doing a mixture of appied and more fundamenta work without continued pubic subsidy. This contrasts with the abiity of some of the US ERCs to survive beyond the end of their NSF grants (though there is presumaby an eement of pubic funding invoved here, too). Generay, the size of firms invoved in the Austrian centres and their more technoogica capabiities compared with those in the ERCs are more imited. Whie Kpus centre participants may reappy to the programme in new configurations, the poicy is that the Federa government wi not finance existing centres beyond 7 years. An important concusion from these descriptions is that the competence centre idea can function at a number of different ambition eves, not east in reation to the eves of technoogica capabiity of participating firms but aso in terms of their scae. 11

14 5 How Competence Centre Programmes Operate In this section, we sketch the characteristics of what we can think of as a composite competence centres programme, based on the eight exampes reviewed. Where there is significant bandwidth, we try to give an impression of the variation. 5.1 Objectives Competence centre programmes aim to tacke both ends of the academic-industry ink. They encourage firms to undertake more radica kinds of innovation based on more fundamenta understanding of the technoogies with which they work. They aim to re-focus some of the activities in the knowedge infrastructure (universities pus research institutes, though more often the first than the second) towards interdiscipinary probem areas of importance to industry. They work primariy with estabished firms that have some absorptive capabiity. Often, they pay a roe in making the knowedge infrastructure attractive and supportive for mutinationa companies with R&D faciities in the country. Sometimes even companies ocated outside the country may participate, but it seems that the physica distance is an important obstace to being deepy engaged. In many cases, they aso contain a proportion of new technoogy-based firms (NTBFs), which may incude spin-offs. They do not work with ow-capabiity SMEs. Competitors are not often present within the same centre. Where this happens, they tend to tacke different topics within the same centre, or to hande ong-term questions of common interest, such as how to tacke environmenta requirements. Centres normay incude a significant proportion of PhD education, producing PhDs who are more used to and interested in working with industria probems than many, and who are more quicky and easiy absorbed into industria companies. Most countries operate one or more separate, academicay-oriented centres of exceence programmes, but there is itte contact with these, which tend to be seen as compementary rather than competing. Competence centres work with a custer of industria partners, whose activities share a common knowedge base. They buid networks and communities of knowedge with vaue that goes beyond R&D. In some cases, they have a regiona focus (Canada, some Austrian and Austraian centres); in others they are nationa. In many cases, they have a sma internationa participation. Most competence centres programmes have been set up in the beief that industry wi take over the funding roe of the state at the end of the subsidy period. Whie thirteen of the NSF Engineering Research Centres have been abe to carry on without core funding after their NSF grants ran out, there is itte evidence to suggest that industry is more generay abe to ignore Neson-Arrow market faiure and step into the roe of funding fundamenta but industry-oriented R&D. This is creating an interesting puzze in some funding systems: whose job is it to fund Pasteur s Quadrant? 5.2 Operations Stakehoders have rarey been invoved in the design of competence centres programmes. However, their active participation in the panning and execution for individua centres is aways effectivey a precondition for funding. 12

15 Competence centre programmes are run by ministries, innovation agencies, research councis or a dedicated organisation Austraia) a modes are practised. Budgets are often arge (Euro 10 60m per year), because the centres themseves tend to be rather big and because programmes commit to fund centres for ong periods. The duration of centre funding varies from 4 years (Kind in Austria, which is unusuay short) through 7 years (Kpus, Austria) to 11 years (US ERCs) and years (Canadian NCEs). The centres are funded for ong periods in order to buid reputations and capabiities, and to provide a context in which more ong-term and fundamenta work can be done. Industry normay commits itsef for muti-year periods, too. Centres are evauated periodicay and poory performing ones wi be phased out (usuay over a coupe of years, in order to protect the postgraduates). Competence centres are rather intensivey evauated, generay through modified peer review, though as programmes age there is rising interest in impact studies. Active evauation is often used in programme management notaby in the eary stages, to test whether individua centres are on track or need to change course; and at ater stages in order to support decisions about whether individua centres shoud continue to receive funding. Typicay, a centra agency pus the universities invoved together contributes over haf the cost. In some cases, regiona governments aso make a contribution. Most programmes work with a rather forma two-step ca for proposa. Outine pans are assessed and a short ist of appicants is invited to submit a fu proposa. Appication / acceptance ratios vary dramaticay, from 10:1 down to amost 1:1. In most cases, scientific quaity is a key assessment criterion, and is secured using internationa peers. Appications are therefore generay written in Engish, since this is the scientific ingua franca(!) The majority of programmes have an additiona and separate appraisa process focusing on reevance, though some take the position that industria funding is itsef a guarantee of reevance. Technoogies supported tend to emerge bottom up, though sometimes peope are encouraged to appy if there are evident gaps in the programme. Since the centres ast for a ong time, it is best to aunch cas for proposas in waves. Otherwise, the programme becomes a snapshot of needs at a particuar point in time that may become many years out of date before the programme can be renewed. Since centres span academic and industria needs and need to have scientific critica mass, they tend to be fairy arge: 20 to 100 or more peope. At the minimum, they need a director abe to tacke the scientific and commercia issues invoved. Often these aspects are spit between two peope. In some cases, training and handbooks are provided for centre managers and others invoved with the programme. Very arge centres (the US ERCs) have more peope invoved in management. Normay, each centre has a board comprising a mixture of industria and academic interests, and often a separate scientific advisory group. Centres usuay have a great dea of autonomy to set and vary their work programmes, and to recruit additiona industria partners (especiay as some partners aways drop out over time). Centres vary as to whether they have a cear, singe ocation or as to whether they are virtua centres functioning across campuses or universities. Our participation in 13

16 centre evauations suggests there are cear and obvious advantaged to a physica centraisation mode. Lega forms vary, adopting whatever form is convenient under nationa aw. Some programmes even eave the ega form of centres up to the participants to decide. Many programmes use a standardised or mode contract for the centres and their reations with their partners. Most programmes set Inteectua Property Rights (IPR) rues, but there are often variations within the individua programme. IPR regimes aso differ among countries, based on prevaiing aw and practice. A specificay Norwegian mode wi be needed here. Centres normay pubish through academic and industria channes. A key aspect of the centres is that industria participants shoud work activey in projects preferaby on more fundamenta projects as we as industriay focused projects. Agreements specify governance structures in some detai, because it is important to maintain a baance of power between industria and academic participants. Normay, there is some kind of Board where the stakehoders sit, and the centre director reports to this Board. Experience is (as one woud expect) that excessive industria infuence tends to make the research agenda rather short term, so that in effect the companies are abe to use the high eve of subsidy as economic rent. If the academics are too powerfu, the companies become mystified or ose interest. An externa scientific committee can be a usefu way to ensure that centres stay in the research mainstream and understand where the technoogica frontiers ie. Competence centres are now od enough that serious ex post evauation is beginning. This immediatey runs into a severe probem, because of the ong ives of the centres: namey, that most of the data needed about the earier ives of the centres have been ost or forgotten. NSF operates a secure web site for recording key data, incuding a record of who has been invoved. Since peope are key to the operation of the centres, keeping such records is vita to understanding them. 5.3 Assessment criteria Each programme has deveoped its own assessment criteria. Many aso track a number of performance indicators, such as numbers of pubications, patents, and so on, but none seems to know how or whether these actuay measure their success, since programme goas are set at a much higher eve. Assessment criteria normay fa into two categories: technica; and economic / reevance. Often, these are assessed by different experts. Technica criteria tend to incude Scientific quaity and novety of the proposed work Likey effects on postgraduate training and production of PhDs Probem orientation and interdiscipinarity Track record of the researchers and participating industry Technoogica capabiities of industry invoved Economic criteria tend to incude 14

17 Conformity with the rues of the programme The apparent viabiity and ikey ongevity of the partnership Likey effectiveness of technoogy transfer mechanisms Likey economic impacts Likey externaities through networking, pubication, raising participants capabiities Active invovement of industry In the work programme Viabe management and governance These criteria can be specified at various eves of detai. The Kpus programme ists about 50 criteria. Others tacke them in a more aggregated way. The NSF criteria provide a good mode, as foows. What is the inteectua merit of the proposed activity? How important is the proposed activity to advancing knowedge and understanding within its own fied or across different fieds? How we quaified is the proposer (individua or team) to conduct the project? (If appropriate, the reviewer wi comment on the quaity of the prior work.) To what extent does the proposed activity suggest and expore creative and origina concepts? How we conceived and organized is the proposed activity? Is there sufficient access to resources? What are the broader impacts of the proposed activity? How we does the activity advance discovery and understanding whie promoting teaching, training, and earning? How we does the proposed activity broaden the participation of underrepresented groups (e.g., gender, ethnicity, disabiity, geographic, etc.)? To what extent wi it enhance the infrastructure for research and education, such as faciities, instrumentation, networks, and partnerships? Wi the resuts be disseminated broady to enhance scientific and technoogica understanding? What may be the benefits of the proposed activity to society? Additiona Review Criteria Proposa defines an emerging engineered system with strong potentia to spawn new industries, transform our current industria base, service deivery system or infrastructure, and have a broad societa impact; Research pan targets critica systems goas, identifies chaenging scientific and technica barriers to be overcome and proposes research projects and proof ofconcept test-beds to address these barriers; Proposa demonstrates a cear knowedge of the state-of-knowedge and the state of-the-art and presents a persuasive strategy for advancing them; Education pan integrates the ERC's research activities and resuts into curricua at a eves, achieves a team-based, cross-discipinary cuture for undergraduate and graduate students, and incorporates effective pans for impementation, assessment and dissemination of curricuar materias; Outreach wi expose a broad spectrum of facuty, teachers and students to the ERC's research cuture, impact pre-coege curricua and motivate students to study engineering; 15

18 Proposa provides a convincing rationae for the seection of industria/use partners and engages these partners in panning, research, education, and technoogy transfer. Institutiona configuration is appropriate to the goas of the ERC and, for mutiuniversity ERCs, coaboration is integrated across the participating universities; ERC has expertise in a discipines required to attain its goas, a capabe eadership team, and eadership, facuty and student teams diverse in gender, race, and ethnicity; Organizationa structure and management pan effectivey organize and integrate the resources of the ERC to achieve its goas and incude strong advisory and project seection/evauation systems. In a muti-university proposa, the resources of a institutions must be effectivey integrated; Experimenta, computationa, and other required equipment, faciities, and aboratory space are in pace or proposed to support the research of the Center; The participating institutions have committed to encourage, support and faciitate the dissemination of the interdiscipinary research, educationa and diversity programs of the ERC. For fu proposas ony Headquarters space proposed for the Center wi effectivey encourage and faciitate interdiscipinary coaboration and house the management functions of the ERC. Commitments from firms to be fee-paying members of the ERC, if an award is made. Proposed terms of the industria membership agreement wi structured a centerwide program of industria coaboration to support overa ERC goas, as opposed to a coection of individua sponsored projects; proposed terms of the inteectua property poicy wi faciitate technoogy transfer. 6 Success Factors Competence centres are compex creatures. There is a ot to get right, and there is not compete agreement among a programme or centre managers about what the success criteria are. Success factors identified in the Networks of Centres of Exceence evauation 11 in Canada incuded Word cass scientific eadership Strong administrative support, incuding having a strong network manager and board of directors A strong and active roe for partner organisations throughout the network panning and research process (not just a roe in name ony ) True coaboration among researchers (not coaborations of convenience ), who represent the nest peope in the fied 11 Dennis Rank, Evauation of the Networks of Centres of Exceence: Fina Report, Ottawa: KPMG,

19 An integrated research programme, in which the themes are mutuay sefsupporting A mutidiscipinary approach, in which periphera discipines are we integrated into the network strategy This is at east a partia ist of success factors, emerging from our overview of centre programmes. Centre managers are very important. They have to be seen as egitimate by both the research community and the industria participants. Often, this is taken to mean that a centre manager has to have a track record as a researcher. In some cases, the roe is spit between a scientific director and a commercia manager, since the job is a demanding one. This overcomes the probem experienced in the Swedish competences centres programme by some centre managers, who found that they ost momentum in their research careers. Managers obviousy need to be perceived as good eaders and managers, and this has consequences for the kinds of personaities that are appropriate for the job. This means in part that they have to be supported by a governance structure that baances academic and industria power in the centre, and which aso deegates authority adequatey to the manager. Research co-operation ony works if researchers actuay co-operate. This assertion is trivia but aso important. Academic and industria participants need to be physicay together and working on the same or reated probems for some of their time, otherwise there is no rea subject for co-operation and there is itte earning. For this reason, and in order to create an esprit de corps, it is important for a centre to have a physica existence, and preferaby to be ceary abeed. In some cases, competence centres are seen by academics as simpy another source of money and by industry as a source of subsidy or cheap (subsidised) R&D abour. Centres where these attitudes are refected in practice never become viabe. Centres and commitments need to be ong term. It takes time to buid up trust among the participants in a competence centre. One of the most successfu of the Swedish centre managers argues that It takes 5 years to become famous, and if there is to be a period of harvesting as we as of growing, this tends to argue for centre ives of the order of 10 years. A consequence of this for the state funder is that there needs to be a way to deiver muti-annua funding, and not to have to get a new funding decision for the programme or for each centre every year. Another consequence is that assessment and seection procedures need to be transparent as we as fair. The aocation of arge, ong-term funding quicky causes envy within the academic system. Inteectua property issues are a source of major conficts. Party inspired by the US Bay-Doe Act, universities are increasingy trying to contro and expoit inteectua property. At the same time, participating companies wi try to contro as much as they can. A compromise seems to be that, where more fundamenta work is done ad where it is not part of a reationship between the centre and a singe company, the IPR beongs to the centre. In cases where a company is paying a high proportion of the cost of the work, it tends aso to get the IPR. Probems of background IPR 17

20 feeding through successive projects can arise, and need management. Academics often aso need training in good aboratory notebook practice and discosure routines. In order to keep the academics interested and to quaity-assure the operations of the centre, it is important that it participate in norma conference and pubication activities. Some programmes have created handbooks and training courses for centre managers. These appear to be widey appreciated. Funding a centre after the programme money runs out is a big and difficut issue. Centres tend to think about this too ate, and then to ask funders for more core money. This kind of onger-range panning needs to be buit into the centre ife cyce as a requirement for funding. Amost no-one invoved with the NCE programme beieved the centres coud carry on unchanged, once the period of network funding (up to 14 years) was over. The eight networks that had been terminated eary significanty decreased tota funding, and their activity emphasis changed significanty. Ony one was abe to find onger-term research funding. An experience with the Swedish programme, which aunched 28 centres at the same time, then had no money to do another ca for neary a decade, is that the resut is a snapshot of nationa needs at one point in time. It is better to aunch smaer, more frequent cas, so that the programme continuay adjusts to needs. In principe, there is a design choice between having a ife cyce at the eve of the programme or of the individua centre. In a number of cases, individua centres have been probem oriented, and have moved from an eary phase of strong focus on rather fundamenta research to a greater emphasis on technoogy transfer and commerciaisation ater in their ives, as important parts of the probems they were set up to address have been soved. The impication is that these centres may move into a sti more commercia mode after centre funding expires or that they need to reinvent themseves so as to find and tacke new probems in a more fundamenta way. Bunching the Swedish investment, as was done, opens the door to having a ife cyce at the eve of the programme as a whoe. It is ess obvious that this is sensibe. After a, is it ikey that a a country s important industriay reevant probem areas for academic industria research can be identified and soved in phase? Yet this is what the atest evauation 12 of the Austraian CRCs effectivey proposes. To date, as far as we are abe to determine, there has been ony one systematic attempt 13 to compare competence centres programmes: between the Austrian Kpus of the BMVIT ministry and the Kind/net programmes of the BMWA ministry. It concudes that the much more stringent and forma approach of Kpus has paid dividends in terms of having greater infuence on changing research and industria Howard Partners, Evauation of the Cooperative Research Centres Programme, Manuka, ACT: 2003 Jakob Eder, Susanne Bührer, Vivien Lo, Caudia Rainfurth and Stefan Kuhmann, Assessment Zukunft der Kompetenzzentrenprogramme Kpus, und Kind/net und Zukunft der Kompetenzzentren, Karsruhe: FhG-ISI,

21 cutures. Given the high eve of pubic support provided, it is especiay important to ensure that the funding is used to support onger-term research rather than to support the type of shorter-term activity that often characterises university-industry co-operation. Another key distinguishing feature of the Kpus programme is the compete separation of the ministry responsibe from the operationa impementation of the programme. 7 Effects of Competence Centres: What Does the Taxpayer Get for her Money? Some of the competence centre programmes discussed have been evauated, with a view to understanding their effects. This section is based upon the avaiabe evauation materia. Competence centres occupy a distinctive position in the research and innovation system. Our own impact study of the Swedish competence centres 14 indicates that they provide A ong-term inkage between university and academic research, which tackes more fundamenta questions than are handed in norma biatera research reationships or than are avaiabe from conventiona network or custer programmes Longer term research than is typicay provided by research institutes, focusing especiay on Pasteur s Quadrant of use-oriented fundamenta research A mechanism to buid (permanent or temporary) critica mass in subjects directy reevant to industry but within the university research system A arge suppy of research-trained peope, who are aready used to working for industry and who are highy sought-after by industry Enhanced networks or coectives among peope working with distinct bodies of industriay reevant knowedge, eading to increased co-operation and personne mobiity within the reevant custers or sub-systems of innovation A suppy of innovations and company spin-offs, with considerabe economic vaue. There are big methodoogica probems in making sensibe estimates about this kind of vaue, but it is cear at east that the vaue of the economic activity generated by the programme in the fairy short term aready exceeds its cost A mechanism to increase the attractiveness of the nationa knowedge infrastructure to existing companies, new start-ups and foreign investors. In Sweden, for exampe, competence centres have payed a significant roe in retaining in Sweden parts of the R&D capabiity of major firms The 2002 Canadian NCE evauation found that the programme has transformed the way research is conducted. Most of those invoved beieved the processes invoved were the same, better or much better than in norma granting agency support. Advantages of working through the NCEs incuded coaboration, mutidiscipinarity, cross-discipinary student training, partnerships with users, knowedge and technoogy transfer, inteectua property protection and deveopment of critica mass. 88% of postgraduates emerging from NCE centres in found empoyment, over haf of them in industry. 56 patents were granted from 14 Erik Arnod, John Cark and Sophie Bussiet, Impacts of the Swedish Competence Centres, report to STEM and VINNOVA, Brighton: Technopois,