Int. J. Entrepreneurship and Innovation Management, Vol. 7, Nos.

2/3/4/5, 2007 117

Regional innovation, entrepreneurship and

talent systems

Philip Cooke
Centre for Advanced Studies and
Centre for Economic and Social Analysis of Genomics,
Cardiff University,
Cardiff, Wales CF10 3BB, UK
E-mail: cookepn@[Link]

Abstract: Growth in general interest, specific research projects and papers and
policy experimentation focused upon regional innovation systems has been
remarkable since the concept entered scholarly discourse in the early 1990s.
Sufficient time has elapsed for longitudinal studies of some of the earliest
exemplars to have been conducted. These are enormously useful intellectually
and in policy assessment terms, since they show how aspirant or actual
regional systems evolve in imponderable ways. That is, it is now clear how
global macroeconomic shifts may destabilise regional system equilibria,
considerably more than, say, national system effects. But it is unclear whether
such shocks assist system trajectories on to a higher or lower innovation
curve in a predictable way. Two elements that have hitherto been understated
in this innovation-led discourse are entrepreneurship and talent-formation.
In this paper, an attempt is made to categorise regional innovation system
evolution according to the robustness of the entrepreneurship and talent
variables.

Keywords: regions; innovation systems; knowledge outsourcing; implicit,

complicit and explicit knowledge; platforms.

Reference to this paper should be made as follows: Cooke, P. (2007) ‘Regional

innovation, entrepreneurship and talent systems’, Int. J. Entrepreneurship and
Innovation Management, Vol. 7, Nos. 2/3/4/5, pp.117–139.

Biographical notes: Philip Cooke is University Research Professor in

Regional Development and the founding Director (1993) of the Centre for
Advanced Studies, University of Wales, Cardiff. His research interests lie in
studies of Economics of Biotechnology (partner in ESRC CESAGen Research
Centre), Regional Innovation Systems, Knowledge Economies and Policy
Actions for Business Clusters and Networks. He co-edited a book entitled
‘Regional Innovation Systems’ in 1998, a fully revised 2nd edition of which
was published in 2004. He authored ‘Knowledge Economies: Clusters,
Learning and Cooperative Advantage’ published by Routledge in 2002.
In 2006 he co-edited ‘Clusters and Regional Development’ and ‘Regional
Development in the Knowledge Economy’ both with Routledge.

Copyright © 2007 Inderscience Enterprises Ltd.

118 P. Cooke

1 Introduction

The rise in interest in the scholarly investigation and empirical application of

Regional Innovation Systems has been shown (Carlsson, 2006) to have outstripped
that for all other systems of innovation research and empirical applications. Now
national governments in Europe, such as Sweden Norway, and South Korea, have
established policies to promote this concept in practice. In both continents, in countries
where regions as governance entities did not previously exist, the attractiveness of
embedding innovation policy at regional level was seen to offer the opportunity for
regional experimentation and policy variation in line with the obvious fact that
regional variation in industry specialisation is a pronounced feature of most if not all
economies. A ‘one-size-fits all’ policy decreed from the central apparatus of the state
was seen to be no longer fit for purpose and indeed may even have become
innovation-inhibiting.
However, policy intervention by means of lighter central ‘steering’ allowing
regional variation in policy to suit geographical variation in socio-economic
characteristics is by no means a panacea in innovation capability building. This is
particularly true in regional contexts in which a high degree of ‘catch-up’ is deemed
desirable. Innovation may be the talisman that regional stakeholders seek, not least for
the widespread belief, if not actuality, that innovation is the fountainhead of economic
growth and competitiveness. But what kind of innovation, placing special policy
responsibilities upon what kinds of institutional actors, with what systemic
relationships to which other key institutions? This is the conundrum for any stakeholder
implicitly or explicitly conscious of and curious about the problem of ‘asymmetric
information’ (Akerlof, 1970) or what may be more contemporaneously called
‘asymmetric knowledge.’1 Recent research suggests that three external environmental
factors can be seen to exert statistically significant effects upon regional firm
performance. The first of these is the positive effect upon firm innovation management
derived from a regional milieu characterised by cooperative industrial relations
(Doeringer et al., 2005). The second is the privileged conditions knowledge-intensive
regions supply for firms locating facilities devoted to search, planning and project
organisation rather than more routine activities within them (Sapsed et al., 2005).
Centrally important here are the capabilities of gaining trust, earning respect,
demonstrating credibility, generating commitment to action and generating social
capital among diverse project stakeholders. Sharing of identity, coordinating
responsibilities and ensuring involvement by participants are the key practices
referred to by Orlikowski (2002) as distinctive ‘senses of knowing’ deriving from
proximity. Like psychotherapeutic interaction, it is ‘proscribed intimacy’ that such
‘knowing’ relies upon. As Sapsed et al. interpret it, this is, analytically, what face-to-face
contact facilitates.
Innovation, being a creative act set within a socio-cultural process stimulated by a
quest for at least some degree of originality, is rooted in the notion of critique or at least
some level of disagreement or dissatisfaction with the status quo. The third dimension of
great significance to innovation seen in this way is the long-term historical dimension of
dissent versus quiescence that may be displayed in various regions. A tradition of liberal
and liberating questioning of received wisdoms – notably from the Catholic Church – is
shown by Iammarino (2005) to continue to underpin the creative and innovative impulses
found in certain (northern) regions of Italy that have key characteristics of being
 Regional innovation, entrepreneurship and talent systems 119

operational regional innovation systems, from Galileo’s heliocentric astronomy to

Tuscan monk Pacioli’s double-entry book-keeping methodology.
This paper has three main sections that seek to take elements of this tripartite
analysis forward by equilibriating these three key characteristics that support regional
innovation. They operate in an externalised, environmental way in juxtaposition to the
three key innovation conditions listed in the title: innovation, entrepreneurship and talent.
Section 1 conducts certain theoretical and conceptual work by way of a three-by-three
matrix that seeks to categorise regions into a typology of innovation systems. This builds
from a complementary typology juxtaposing business and governance characteristics that
has assisted theoretical development of regional innovation systems analysis from the
outset (Cooke, 1992; Cooke and Morgan, 1994).
Section 2 discusses, with empirical illustration, recent findings on the dynamics of
regional innovation systems based on the only extant study of a number of candidate
regional innovation systems studied at more than a single point in time (Cooke et al.,
2004). This overcomes a point of critique of much systems of innovation research to the
effect that it is overwhelmingly static (Iammarino, 2005). Finally, some textured case
analysis of recent organisational policy change that seeks to rebalance regional
innovation system capabilities by attending to entrepreneurship and talent-related issues
are presented. Conclusions are then drawn.

2 Typologising regional innovation, entrepreneurship and

talent systems

The typologising of regional innovation systems has evolved throughout 15 years since
the concept was introduced. It may be said to have undergone two static and
one dynamic iterations, the latter occurring in Cooke et al. (2004). In 1992, the first,
fairly simple, effort – albeit starting from almost a theoretical and empirical terra
incognita – was constructed as follows. At the time it was being researched (circa
1989–1990) the drivers of this work were the concepts of innovation replacing a much
narrower discourse of ‘technology’ and networks emerging to challenge ‘market’ and
‘state’ as the key coordination modes of complex institutional action and industrial order.
An influence upon highlighting these concerns was the then recently published book of
Best (1990) on new forms of competition. Notable were accounts of the superiority of
certain networked forms of innovation governance over markets.2 A notable instance of
this were the distinctive trajectories of the Italian design-intensive and the UK
scale-intensive furniture industries. The former succeeded while the latter atrophied.
Noticeably, such classificatory effort conducted at that time made no reference to the
main channel of national innovation systems literature, not least because no such effort
had been made (and still has not) by that school (Table 1).
The ‘industrial district’ RIS model was thus the genesis for the stylised category
‘grassroots’ since it was clearly a market-driven model but one in which neither ‘scale’
nor ‘science’ were particularly pronounced. Nevertheless, local, ‘weak ties’ coupling or
coordination were clearly present, as were diffuse – we would now say social capital
based – funding mechanisms and a regionally diversified grouping of distinctive
‘industrial districts.’ It could be said that communicative linkage in the ‘grassroots’
model was locally intense but mediated nationally and globally by looser entrepreneurial
intermediation.
120 P. Cooke

Table 1 Typology of regional innovation systems and key action impulses

RIS Characteristics Grassroots Integrated Dirigiste

Initiation Local Multilevel Central
Funding Diffused Guided Determined
Research Applied Mixed Basic
Coordination Low High High
Specialisation Weak Flexible Strong

Source: Cooke (1992).

This was contrasted first, with a more consciously guided mode of innovation
governance in which small- and medium-sized enterprises interacted regarding
innovation in significant ways regionally, nationally and beyond, with large corporations.
A scale above the ‘grassroots’ model but still by no means ‘governed’ or ‘coordinated’
either fully by market relations arising from corporate power or state planning modes of
intervention this was basically a ‘partnership’ model of networking in which power
relations clearly existed but so did elements of symmetry in the innovation interactions
among stakeholders. The ‘network paradigm’ enabling the model to be envisaged arising
from varieties of engineering excellence, from automotive to tooling and printing
machinery, was Baden-Württemberg’s industrial model. This was an exaggerated,
socially pervasive form of the Germanic type of industrial order in which constructed
advantage (Foray and Freeman, 1993) had arisen historically because the region was
once so poor and suffering from major outmigration.
Baden-Württemberg was the Ireland of a previous era, a Swabian rather than Celtic,
Tiger. Under the enlightened Kingdom of Württemberg, Ferdinand von Steinbeis in 1868
established the first industrial training facility in the locomotive works at Esslingen, near
Stuttgart. This is thus an early case of endogenous constructed regional advantage,
subsequently evolving and spreading its benign influence throughout a geographically
integrated administrative region, such that today there are over 500 Steinbeis knowledge
transfer centres within Baden-Württemberg and neighbouring regions and States.3
However, they are structurally embedded in a specific cultural and institutional context.
For example, they were designed as ‘fiefdoms’ mainly of Fachhochschule (polytechnic)
professors in a region with some 40 such institutions set in a national higher education
system that privileges chair positions much more than elsewhere. Hence, we can say with
some confidence that regional innovation systems may not only be categorised
distinctively on the basis of ‘epistemic community’ problems like this, but also that they
can only be ‘learned’ under highly specific conditions.4
Finally, the third category of regional innovation system explored at the dawning
of interest in the subject was the ‘top-down’ or dirigiste model. As the terminology
implies the inspiration for this was Francophone. During the 1970s and 1980s, France
(and Quebec) had been at the forefront of innovation policy. In Cooke (1985), it was
shown how the French idea of technopoles evolved from the post-war state planning
model inspired by Francis Perroux and his successful promotion of the ‘growth poles’
idea. This had lineaments to Schumpeter’s idea of ‘swarming’ by imitator firms of the
innovating entrepreneur that leads to the formation of what today we would term a
‘cluster’. However, it was rather the prosaic case of Germany’s Ruhr Valley that was
Perroux’ inspiration. Hence, large scale industry, such as the great steel complexes at
 Regional innovation, entrepreneurship and talent systems 121

Dunkirk and Fos-sur-Mer were designed as economic anchors for downstream

processing of such inputs. The technopole idea was its transmogrification into more
analytical, science-based and advanced technology activity. Science Parks and even
Science Cities, as at Lille were constructed, usually by decentralising government
research laboratories from Paris. This happened at the first such technopole with the
Meylan-ZIRST science park at Grenoble, opened in the 1950s.
By the 1980s, Grenoble and the wider Rhône–Alpes region were leading innovation
in energy, high performance engineering and materials. Populating the regional system
by the early 1990s were Regional Centres for Innovation & Technology Transfer
(CRITTs) in composite materials, agro-food, production, electronics and mechanics,
biotechnology and biomedical devices, chemistry and energy. There were also nine
DRRTs, decentralised offices of the French Ministry of Research and Technology and
numerous investments by ANVAR, the national risk-capital fund, in the region.5
Remarkably, despite the emergence of regional government in France some years
previously, all this infrastructure was designed and funded from Paris. The region’s
comparative advantage in hydro-electric power generation had morphed into a globally
significant expertise in nuclear energy production, hence also specialisation in materials
science, electronics and mechanics. Thus, the region had become a technological
powerhouse, with the pole itself including colleges, private and public research, large
businesses (e.g. Rhône-Poulenc, now Sanofi-Aventis, Caterpillar, Philips), SMEs and
financing vehicles, but the whole was, as it were steered or ‘directed’ from Paris with
relatively minimal regional or local innovation input.
Hence, it clearly differed from both ‘Third Italy’ and Baden–Württemberg’s more
associative models, but as Colletis (1993) concluded at the time, it seemed to work as a
system in which ‘talent’ was trained, labour market positions existed in innovative
business or research organisations and large and SME businesses coexisted with
ANVAR bankrolling the latter. This was the top-down system par excellence.
To conclude this round of the ‘thick description’ of distinctive systems (Geertz,
1973) let us anatomise three dimensions of the categorisation in terms of the way – at
that time – these regional types had been constituted as forms not of inherited but rather
constructed advantage. First, although each type of region occupied a position in a
distinctively national innovation system, and to a considerable extent, as we have seen,
these national institutional characteristics could be seen to be non-transferable, by no
means all regions in those national systems were alike. Thus, the asymmetric knowledge
issue comes forcefully into play. Each of these regional types had once been poor, with
major rural–urban and international out-migration, a few natural resources and skills of a
traditional craft kind, wood-working mainly.6 In all cases, innovative capabilities came
from an entrepreneurial impulse, probably with a regulatory flourish from noblesse
oblige. This enabled a specialist ‘tinkering’ survival strategy to construct advantage
(of course, monopoly always helps) that produced benign path dependence. In Germany,
the paradigmatic regulatory environment was the Württemberg royal house and
vocational training innovator von Steinbeis, forerunner of land Minister President Lothar
Späth, famously quoted in his first book7 as saying, to paraphrase, ‘…in this region we
have only ever had our wits to rely upon…’ (Späth, 1985). Finally in Italy, notably in
Emilia–Romagna and Tuscany:

“…the progressive traditions of the dukedoms of Parma and

Piacenza…Modena and Reggio…exerted a powerful influence on the historical
evolution of regional knowledge linkages. The industrial districts of Carpi,
122 P. Cooke

Sassuolo and Faenza are among the oldest in Italy… The economic prosperity
of Toscana started in the eleventh century based on the unbeatable cultural
traditions of its communes (Lucca, Pisa, Pistoia, Siena, Arezzo) ….Florence
prevailed and absorbed the most positive elements of Toscana’s typical pattern
of development.” (Iammarino, 2005, p.510).

The second mark of distinction requiring unpacking is entrepreneurship, so clearly a

pronounced feature in each category. In an important manner, ‘being entrepreneurial’
arises from Orlikowski’s ‘sense of knowing’ about how, in this case, to ‘do business’.
This comes from a different dimension of Iammarino’s historical analysis of
regional innovation systems in the longue durée. Although she writes of Italy, something
of this must also have imbued the competitive but communal entrepreneurship denoted
in the following: “…communes, all fierce competitors for market share, product
variety, trade intensity, cultural creativity and social dynamism…” (Iammarino, 2005,
p.510). Notice that, importantly, we are observing traditions of collective
entrepreneurship, familial, communal, social but also competitive, confident that the
reliability of significant others to provide finance, to assist in planning and production, to
speak well of the key artisan seeking a contract in a ‘system of production’ often, in
those early days perhaps for some kind of luxury consumption – a painting, a clock, an
(ornate) comb.
We have seen how constructed advantage took distinctive forms within a framework
set by comparable but by no means identical cultural and institutional conditions.
This approach, as Geertz (1973) describes in the methodology of ‘thick description,’
enables discovery of distinctiveness within a common broad framework. In his exemplar
study, this enabled him to unpick the varieties of cultural practice within Islam stretching
from Indonesia to Morocco and root them partly in preislamic cultural inheritances,
from Hinduism in the former case to Berber in the latter. However, in RIS set-ups
distinctiveness evolves as part of a more open, innovative, modestly liberal governance
regime in which entrepreneurship of a collective kind and innovation in meeting
competitive market demand are able to flourish. Iammarino compares the more
innovative Italian regions in the late middle ages with those, like Marche, Umbria and
Lazio of the Papal States, stifled under institutional dogmatism, changing to an almost
geological chronology. Hence, innovation and entrepreneurship require ‘talent’ and
crucially the means for producing talent. Moreover, in more adventurous settings
institutions for the formation of talent thrived, whether at Esslingen or in the creative
apprentice workshops of Florence. Even at Oyonnax, for the eyewear business the real
take-off came with the foundation of the Société Confraternelle des Ouvriers Lunetiers in
Paris in 1849. This confederation or ‘house of lens-makers’ was the ancestor of today’s
powerful Essilor group. Taking its inspiration from the collectivist ideas of Saint-Simon,
the association was founded by three small industrialists: Duez, Duriez and
Muneaux. But notice the dependence on Paris for this. Nevertheless, the Lycée
Technique Victor Bérard was France’s first optics school, founded in 1933 at nearby
eyewear district Morez.
It is thus clear that the historical embeddedness and path dependence of regional
innovation are rooted in and evolve from certain initial conditions that assist the
construction of specific kinds of regional advantage. But are such initial regulatory
conditions favourable to asymmetric knowledge evolution, a high degree of intra-system
openness and collaboration, but collective competitiveness that amounts to certain
exclusive, quasi-monopolistic practice towards outsiders, and institutions, such as
 Regional innovation, entrepreneurship and talent systems 123

business associations, to reproduce these enough? As suggested, two key conditions that
further enhance constructed advantage are those that nourish entrepreneurship and
talent-formation. These are fundamental features of the full-fledged innovation system
that in times of relative stability in economic conditions may be reproduced
‘organically’, that is without strong evidence of institutional or social reflexivity about
how the system ‘works’. But in times of disadvantage, instability and economic threat,
asymmetric knowledge and the associated knowledge capabilities must be ‘retheorised’
and the system put together in ways that facilitate its survival (Cooke, 1995). Otherwise,
as Brenner (2005) shows with theoretical and econometric finesse, regionally constructed
advantage, along with the competitive edge it gives rise to, are blunted and the economic
means of existence are lost as the system atrophies arising from loss of markets.

2.1 The dynamics of change in regional innovation systems

We see governance and business scope and scale (including innovation,
entrepreneurship and talent) as key analytical intersections in the further elaboration of
innovation systems typologies. The next stage of typologising regional innovation
systems took this insight seriously by evolving a matrix in which these two vectors
formed the coordinates (Cooke, 1998). More recently, with the publication of a revised
edition of the book in which this appeared, it was possible to dynamise the
categorisation since data on the position of regions studied were available for two time
points ten years apart. Figure 1 summarises the observed positions. The detailed
accounts of these regional trajectories may be consulted in Cooke (2004) but in brief,
three things may be said about both the general classification and the dynamic element
portrayed.
The first of these is that ‘regional’ is somewhat extended to encompass at least
three small nation-states; Denmark, Singapore and Slovenia. It is slightly ironic that
questions are sometimes raised by non-spatial economists about the validity of a
regional focus given that some are big and others small. The same critics seem to have
little difficulty in encompassing Singapore and China in economic analyses. Anyway, it
is important to draw attention to the fact that we have three states in the sample, albeit
small ones. It occurred because of analyst receptivity to being included in a globally
diverse sample of relevance to their interpretative capabilities. In one case (Denmark),
the author preferred to conduct the country than a regional analysis. The key point
is that such states clearly have in principle more autonomy to implement innovation
support initiatives than most regions even though in a globalising economy they may not
have as much leverage as they would consider ideal. Nevertheless, they pursue sovereign
policies that also autonomously favour firms in their jurisdiction on matters like
monetary policy over which regions usually have no control (though they often do on
some fiscal and to some extent budgetary matters). What is of interest despite
these different autonomies (also present among the regions to a degree) is that the
directionality of innovation system trajectories is not obviously out of line for these
regions or nation-states. Thus, for example, the tendency for business and governance to
have moved closer to the centre of the diagram, meaning demonstrating a higher
networking propensity is shared. A few regions go against this trend by remaining
relatively unchanged in their business innovation character and governance of
innovation support. In one case, although the society became more democratic as
124 P. Cooke

devolution to the Welsh Assembly occurred as part of the reforms in administration

in the 1990s, the new government has centralised all innovation administration
unto itself.

Figure 1 Regional innovation systems: typology and evolution

Source: Cooke (2004).

The second observation that interpretation of these trajectories suggests, is that the
concept of utilising innovation for business competitiveness and regional constructed
advantage has grown over the ten years encompassed in the regions researched. Even
Baden–Württemberg, already a highly accomplished innovation setting, as we have seen,
experienced new associative policy input from its latest Commission of Inquiry
published in 2000. Now policies to stimulate diversification of innovation across a range
of industries like biotechnology, photonics and new media are a pronounced feature of
the policy landscape. This is consequent upon recognition that the regional economy had
become undesirably ‘monocultural’ around automotives, to echo the terminology of an
earlier Commission of Inquiry’s findings. This desire by regional and smaller state
innovation policy-makers for a biotechnology presence is a marked evolution of
 Regional innovation, entrepreneurship and talent systems 125

innovation focus. It ranges from Singapore’s Biopolis to Denmark’s Medicon Valley now
linking Copenhagen to the Swedish side of the Øresund bridge.
Catalonia, Tampere region and Wales are among regions extending or building
Bioparks and Gene Parks. With these developments goes a heightened sense of the
importance of talent and entrepreneurship within innovation systems. This is because it is
not multinational companies that carry this development impulse but spinout firms from
university research ‘centres of excellence’. Attracting, recruiting and retaining research
talent, on the one hand and assisting in the formation of entrepreneurs to transform
exploration knowledge into exploited commercial innovations, on the other, have
become greater imperatives in the past ten years.
The final general point and corollary of the previous one is that, more or less
successfully, regionally constructed advantage is being gambled on industries at the heart
of the new ‘knowledge economy’ because the constructed advantage of only a decade
ago is swiftly vanishing. Consider Tuscany, Brabant and Wales as paradigmatic cases.
Tuscany has numerous, distinctive clusters, more traditionally referred to as ‘industrial
districts’ in its regional setting, ranging from textiles to leather and furniture. Each of
these traditional sectors is assailed by cheaper goods from, particularly, China. In Prato,
the heart of Tuscany’s woollen fabric industry the cluster is evolving into a clothing
manufacturing centre. The skills needed to, for example, stitch, sew-on buttons and make
clothing accessories were not traditional to Prato and, in any case, Italian wages would
render products prohibitively expensive. Remarkably, during the ten years covered by
these regional innovation system case studies, at least 23,000 Chinese workers and
entrepreneurs moved to Prato to supply these missing elements in the local division of
labour. They have been joined by numerous other Asian and North African ethnic
minority workforces. In the Netherlands, one of that country’s greatest firms, Philips,
was formerly headquartered in Eindhoven in the region of Brabant.8 Philips owes its
existence to a decision by the Dutch government in the last century to suspend
international patenting protocols as a means of constructing advantage for what was
perceived to be a lagging economy at the time. The Edison light bulb was simply
‘pirated’ as a consequence. Philips grew to be an almost classic M-form corporation.
Nowadays however, recognising inefficiencies in that ‘stand-alone’ corporate structure it
has adopted ‘open innovation,’ outsourcing R&D to entrepreneurial firms and research
9
institutes within and beyond its region. In this way, ‘knowledge entrepreneurship’ has
been externalised with all the network management skills that are thereby implied being
evolved. Finally, Wales’ innovation system was relatively over-governed but Japanese
and other Foreign Direct Investment (FDI) engineering firms stimulated local
entrepreneurship and talent formation as they elaborated regional supply chains. By 2005
even one of the first, iconic FDI arrivals Sony, had almost completely shut, following
other plant closures by Hitachi, Panasonic and Korea’s LG. A totally new
entrepreneurship and talent-based innovation approach is now painfully and slowly being
attempted (Cooke et al. 2004).
So, big changes in a relatively short time-period, each with massive implications for
the slope and direction of regional development trajectories. Clearly, regional
innovation systems are not islands, they are more like icebergs, swiftly affected by their
global environment, immediate external conditions and internal dynamics. Accordingly,
they are fascinating objects of study but devilishly difficult to ‘design’, especially in
times of economic development instabilities, notably the rise of major global
competitors. More and more, where policy consciousness and reflexivity involve
126 P. Cooke

recognition of the importance of innovation to economic growth and the connected

importance of entrepreneurship and talent to innovation, we see regional firms and
policy governance engaging with new challenges in the ‘knowledge economy’. This
stretches from the widespread embrace of new, thriving but complex sectors like
biotechnology to the equivalent organisational complexities of managing ‘open
innovation’ through R&D outsourcing. The knowledge economy demands sophistication
in knowledge exploration, knowledge management and knowledge entrepreneurship.
In Section 3, we examine some glimpses of the future for such ‘regional knowledge
capabilities’ in the construction of regional advantage by better integrating innovation,
talent formation and entrepreneurship through the redesign or the novel design of
regional innovation systems.10

3 Reinventing regional innovation systems

We could say that the old regional innovation systems, where they existed, assumed the
characteristics represented mainly on the left side of Table 2 while the new type seeks to
promote characteristics closer to those on the right. That is, there are moves from
knowledge scale to scope, from closed to open innovation, and from reliance on
corporate hierarchies to entrepreneurship. For convenience, this former may be termed an
Institutional Regional Innovation System (IRIS) and the latter an Entrepreneurial
Regional Innovation System (ERIS) as discussed in Cooke (2004).

Table 2 Aspects of regional and regional innovation systems knowledge

Institutional RIS (IRIS) Entrepreneurial RIS (ERIS)

Research and development driven Venture capital driven
User-producer relations Serial start-ups11
Technology-focused Market-focused
Incremental innovation Incremental and disruptive
Bank borrowing Initial public offerings
a
External supply-chain networks Internal EcoNets
Science Park Incubators
a
Venture capital cross-holdings in portfolio firms among whom outsourcing is promoted.
Source: Cooke (2001).

The key point regarding the (somewhat artificially opposed) sides of Table 2 is that ERIS
is rooted in entrepreneurship and talent formation that is largely outside the corporate
hierarchies presumed to be the drivers of innovation in IRIS set-ups. Such large
institutions showed during the 1990s that they experienced difficulties in many sectors in
bringing forth innovations. Already the global energy industry displayed low Business
Expenditure on R&D as a percentage of GDP. OECD (2003) estimated it at less than 1%
of GDP in the USA and Europe. Essentially, these large corporations have outsourced
R&D almost completely to ‘problem-solving’ companies like Halliburton and
Schlumberger who, in turn, operate their own knowledge supply chains. It is inside such
supply chains among entrepreneurial firms that R&D is principally conducted.
 Regional innovation, entrepreneurship and talent systems 127

In other industries, notably pharmaceuticals, which has a globally high expenditure

on R&D at some 18% of revenue, much of this (up to 50%) is outsourced to
knowledgeable biotechnology and pharmaceuticals knowledge suppliers, universities and
other research institutes. Chesbrough (2003) shows this to be common also in the ICT
and electronics industries, as we noted concerning Philips.12 Hence, this large scale
externalisation of knowledge exploration, examination and exploitation to knowledge
entrepreneurs of a public or private kind, including in universities, creates a new
knowledge and innovation landscape for regional governances to grapple with. In Table
3, the firm-level shifts towards management of open innovation compared to the tradition
of more ‘closed innovation’ are summarised. The latter was conducted in-house by
dedicated R&D departments of the kind firms like General Electric, Dupont and AT&T
pioneered, but which the last two at least have largely divested, decentralised or
externalised. It is evident that, to the extent such R&D, along with other outsourcing, has
progressed to the point where the likes of Indian and Chinese knowledge and services
suppliers are well-integrated in the knowledge value chains of Western companies,
globalisation itself is in transition to a new economic geography. In this, the ‘constructed
advantage’ of possessing regional talent pools of global significance now dictates
locational decision-making to multinationals in ways that were not so determinate when
largely ubiquitous cheap manufacturing labour was available in many global locations, as
it still is.13 We term this a transition from Globalisation (1) and the quest for cheap
routine labour, to Globalisation (2) and the magnet of regional talent pools of global
significance (Cooke, 2005b).

Table 3 Closed and open innovation

Closed innovation Open innovation

Strategic R&D – core business strategy Firms select desired technologies
Firms perform R&D in-house Technology acquisition
Firms put technologies in products Firms market products
Product revenues fund additional R&D R&D Outsourcing
Globalisation (1) adjusts products to markets Globalisation (2) taps global talent pools

Source: Chesbrough (2003).

What remains for Western regional innovation systems is less and less routine
manufacturing and private services (e.g. branch plants and call centres) and more
relatively untraded services (healthcare, education, administration), local traded services
(transport, retail, cuisine, entertainment) and knowledge intensive exploration (research),
examination (testing and trialling) and exploitation (innovation) activities.
As an indication of the way R&D outsourcing has evolved in the USA, the data in
Table 4 summarises the position for the most recent data year available. It is clear that
in the decade of the 1990s a significant shift towards R&D outsourcing and ‘open
innovation’ occurred on the part of large US corporations. Discussion with experts
suggests possibly half is accounted for by medical and bioscientific and a further quarter
by ICT and electronics. At the 1999–2001 rate of decline, markedly slower than that
from 1989 to 1999, it will have halved from its 1981 share of nearly 71% of total
US industrial R&D by 2006.
128 P. Cooke

Table 4 Percentage of US industrial R&D by size of enterprise

Company size 1981 1989 1999 2000 2001

<1000 employees 4.4 9.2 22.5 22.1 24.7
1000–4999 6.1 7.6 13.6 15.2 13.6
5000–9999 5.8 5.5 9.0 8.3 8.9
10,000–24,999 13.1 10.0 13.6 14.0 13.0
25,000 + 70.7 67.7 41.3 39.5 39.0

Source: NSF (2003–2005) Research and development in industry (2001).

To identify the extent to which much of this outsourced R&D has entered the purview of
the smallest SMEs in the USA, Table 5 presents the appropriate results. These show huge
percentage gains for SMEs in the 1997–2001 period, albeit slowing somewhat after
the end of the technology stock market boom in 2000. It is not known what share of these
quantities is internally-generated SME own funding but it is likely to be relatively small
since outsourcing is precisely how most R&D conducting SMEs survive, namely by
taking R&D subcontracts. Thus, in five years, the amount of R&D conducted by SMEs
was not too far short of doubling. Given what is known about the geographical
distribution of knowledge-intensive SMEs in the USA, it is highly likely a substantial
number are located in the high tech clusters identified by the likes of Norton (2000) and
Florida (2002) in Boston, San Francisco, San Jose, San Diego and Austin, Texas.

Table 5 R&D expenditures by small US companies (Millions of constant $1992/ *$1996)

Size of firm 1997 1998 1999 2000* 2001* Change (%)

<25 2536 3804 5579 6176 4346 71
25 to 49 2455 2525 3824 4507 3375 37
50 to 99 3415 5155 5779 6533 7382 116
100 to 249 5907 6622 5707 8118 11,634 114
250 to 499 5229 5522 6463 6731 7832 50
Total 19,542 23,627 27,352 32,065 34,569 77

Source: NSF (2003–2005) Research and development in industry (2001).

How to capture such clustered, talented, knowledge ERIS growth elsewhere? This is the
new challenge being faced by regional innovation stakeholders worldwide. In the
following few pages, evidence will be presented of two case exemplars for which such
ERIS knowledge-intensive development is being pursued in new ways. They are in
Scotland, UK and in Sweden.

3.1 Complicit knowledge: research transforming implicit to

explicit knowledge
It is by now clear that the academic debate concerning the value of geographic proximity
for implicit to explicit knowledge transformation as an explanation for clustering
practices by firms is based on something of a fantasy. There is a large literature in
organisation science reviewed, for example, by Sapsed et al. (2005) that shows
 Regional innovation, entrepreneurship and talent systems 129

permanent face-to-face contact is unnecessary. It is far more likely that such clustering
activity occurs to control knowledge access,14 involving protection of valuable
knowledge by researchers and/or their clients from competitors and communication of
such knowledge formally or informally only with the trusted few who earn that position
by virtue of possessing complementary and valuable knowledge of their own. In other
words, it is as valid to hypothesise clustering in proximity of this kind to be
quasi-monopolistic practice,15 as it is to argue it is primarily an open market practice.16
It has to be asked why firms would willingly pay the extra costs in land-rent (estimated
as double the norm in Cooke, 2005c), wages, congestion diseconomies and so on denoted
by spatial concentration, if knowledge freely flowed outwards in ubiquitous open market
fashion? They could operate through distant networks or arm’s length exchange.
Of course for some firms and some sectors that is precisely what happens. It depends on
the kind of knowledge in question. Thus, biotechnology firms often cluster near the
exploration knowledge base because, as Zucker et al. (1998) show, academic researchers
wish to control the knowledge advantage they have and tend to outsource examination
and exploitation to trusted spinout firms (in which they may have a financial interest)
who must locate in proximity to conduct joint, for example, patent-related or
commercialisation work. But this also involves complicit actors with specialised
knowledge of patient trialling, patent law, commercial due diligence and venture
capital for an innovation to have a chance of ensuing. However, such conditions
of perfect implicit-complicit-explicit knowledge interaction occur globally in relatively
few places.17
Contrastingly, evidence suggests that ICT firms are likely to cluster elsewhere despite
18
also being high technology businesses. They find location in proximity to customers
and suppliers far more important than proximity to exploration knowledge. Unlike
biotechnology, the expertise and equipment sophistication are insufficient to their
requirements as ‘synthetic’ knowledge specialists themselves outsourcing or being
outsourced to meet customised client requirements. Being located in proximity to such
contractual value chain opportunities enables them to exert control over the company’s
destiny since anticipatory knowledge concerning contracts is frequently ‘in the air’ as
Marshall famously put it.19 Finally oil and gas industry clustering, for example, occurs far
away from the theoretical knowledge base and equally so regarding the ultimate client
base, the multinational oil and gas corporations. Rather, in locales like Aberdeen,
Houston and Stavanger the great variety of offshore and onshore engineering and
logistics specialist firms gain from proximity to contractors that, again, frequently require
20
customised, innovative solutions to problems of extraction and transmission.
These three sectors are perceived to be of key importance to the economy of
Scotland, to which we now turn. The Scottish innovation system is interesting since it is
relatively strong in regard to exploration knowledge, mainly university research, there
are numerous spinout companies in the three sectors although in the energy sector these
are corporate not academic spinouts. Academic spinouts mostly remain small and often
benefit/survive through enterprise support platforms from the Scottish government or
Scottish Enterprise, its economic development agency. The leading research universities
also have strong industrial research collaborations with leading global corporations.
Where the Scottish innovation system is least ‘joined-up’ is with its indigenous SME
population comprised of firms in many sectors, including advanced technologies, that
were not academic spinouts. However, the highly exacting demands of the Scottish
Executive (government) are prioritised on pursuing the policy of a ‘Smart, Successful
130 P. Cooke

Scotland’ fully engaged with the global knowledge economy. Thus, improving the
exploitation of Scottish research excellence is seen as the priority, with integration of
SMEs a more complex challenge to be faced later. To that end, the key problem has been
identified as the inability of Scottish talent and entrepreneurship to innovate commercial
products and services. In other words, it is a problem of too little complicit knowledge in
the market, a problem shared in many European regions compared to the leading US
‘megacentres’ like Boston and San Francisco.
The innovation system instrument introduced in Scotland to better integrate the
knowledge value chain to construct advantage is the idea of the Intermediary Technology
Institute (ITI). Three of these, in Life Sciences, ICT and Energy were announced in
2003.21 They compensate for perceived market failure in knowledge commercialisation.
The manner in which this is achieved differs somewhat according to the focus of each
ITI. With respect to the ‘TechMedia ITI’, the colloquialism for ITI, ICT and Digital
Media, there are three innovative aspects to organisational strategy. Firstly, evolving in
interesting ways from the traditional sectoral perspective often associated with economic
development agency thinking, it adopts a Thematic approach to its technology promotion
activities. This means that the focus is upon ‘platform technologies’ having qualities of
‘related variety’ and ‘ubiquity’. Thus, they are likely to be pervasive in their impact
rather than being confined to a sector. Hence, Broadband Wireless Technologies,
Content Creation Tools, Human–Computer Interfaces and Networked Sensor
Technologies have wide potential application. Networked Sensor Technologies, for
example, allow remote monitoring of animals, wind farms or even personal healthcare.
This indicates some positive signs of escape from the commonly emulated ‘key sectors’
view of public innovation support targets, which is widely perceived in today’s
innovation discourse to entail possibly debilitating policy rigidities.22 For ITI Energy, the
fact that the world’s great oil and gas firms have outsourced R&D expenditure for all but
0.4% (0.2% in UK) of revenues, points to the central importance of knowledge
entrepreneurship opportunities in energy. Numerous oil and gas employees have shown
entrepreneurial skills (‘cashing-out’ during upswings and downswings) by spinning
businesses out of large firms to operate in specific knowledge value chains.
A second innovative aspect in ITI approach is to ‘walk the talk’ regarding
cluster-building, a policy approach of considerable provenance since Scottish
Enterprise’s pioneering adoption of this economic development methodology from the
early 1990s. The cluster idea involves taking policy actions that enable businesses to gain
from complementarities, collaborations and knowledge spillovers especially where
‘related variety’ firms operate in geographical proximity. From some viewpoints, this
may smack of exclusivity or even construction of quasi-monopoly conditions where a
group of firms enjoy ‘club’ benefits.23 But building the cluster may, in fact, mean
disrupting such equilibrium conditions by, for example, introducing a competitor. This
occurred in Life Sciences and caused anxieties from incumbents, from which new
knowledge of the problems of cluster-building arose. Nevertheless, this was justified in
terms of ‘building the cluster’ rather than protecting established interests. In other
respects, the cluster perspective, while being perceived as a worthwhile business
development model, is also seen as a policy retro-model, meaning it is based upon ‘a
compartmentalised 1980s picture’ rather than a future-shaping one. Thus, from the ICT
perspective, software is seen as a pervasive platform technology. But policy reality
was, from a firm perspective, inadequately expressed in software having no cluster
policy-identity. Moreover, the policy action-line that subsumed software under
 Regional innovation, entrepreneurship and talent systems 131

‘e-learning’ was found ludicrous in the industry, given ‘e-learning’ is something that
from a market perspective has zero profile. This is a clear indicator of the ‘epistemic
community’ problem where, for policymakers, e-learning is a social imperative while for
firms it is a market loss-maker. Here is a good instance of the new ‘knowledge economy’
policy requirement for greater isomorphism between innovation and policy platforms in
constructing regional advantage. Thus, platform technologies like software, displaying
pervasive characteristics, may not fit the cluster idea as practised in regional economic
development as well as an innovation system or network metaphor. This lends greater
credibility to the diverse interactions involved in complex, multiuse or ‘general purpose’
innovation systems than more conventional sectoral-spatial notions such as clusters
(Helpman, 1998). ITI Energy’s oil and gas sector is highly networked vertically and
horizontally, but renewable energy is more characterised by smaller firms, perhaps still
mainly competitive more than collaborative (lobbying practices, associations, industry
organisations, etc., are widespread in the former).
Finally, an ITI innovation is to seek to overcome barriers arising from the
existence of ‘communities of practice.’24 Here, for example, SMEs without university
traditions either in respect of management training or innovation frequently fail to
interact significantly with such knowledge centres for innovation purposes. Similarly
‘Triple-Helix’ models that privilege senior manager ‘communities of practice’ of
universities, industry and governance bodies may exclude the academics actually
conducting research. Accordingly, decoding of tacit research knowledge and its
relevance to large, small or spinout businesses, may be inadequate. Equally, interaction
with academics to the point where arcane university conventions concerning such
technical matters as IPR licensing, probably means an administrative functionary is best
suited to conduct decoding of administrative conventions. Or a senior organisation figure
needs to be complicit with the aims of ITIs and their market-facing responsibilities to
effect action in cutting red tape. In whichever dimension, the creation of some modest,
‘weak ties’ social capital in order to push forward the innovation agenda is necessary if
innovation is to be seen as ‘systemic.’ In their first 20 months, the ITIs have performed
valuable network-building functions. This comes about in two main ways: formally, by
the construction of membership methods of organisation of outside organisations
(e.g. universities in their corporate guise); firms (large and small) and individuals
(e.g. interested academics or entrepreneurs). The second way is more informally, through
one-to-one meetings with key actors in the ‘system’ such as Deans of University
Schools, investors and anchor firms to create trust, reputation and communication
channels (communities of practice) among innovation constituencies. Both ITI
TechMedia (individual and corporate) and Energy (corporate only) stressed the
importance of their ‘membership model’ for communicating with users.
A different model is under development at VINNOVA in Sweden. There, it is
internal ‘organisational knowledge’ concerns that have stimulated the creation of an
Intranet now transformed into an Extranet to regional innovation systems the agency is
constructing throughout Sweden. VINNOVA was itself in 2001 an organisational
innovation arising from a former technology transfer agency called NUTEK. The
latter was seen politically as having failed Sweden’s regional development and to be
over-wedded to an engineering culture influenced still by the linear innovation model.
Among other ‘single cell organism’ characteristics, this ‘linear’ thinking was reflected in
over-allocation of technology subsidies to the Stockholm area. VINNOVA’s task was
more complex, entailing construction of innovation systems in Sweden’s regions.
132 P. Cooke

It began with cluster building activities in Skåne, Uppsala and west of Stockholm.
However, due to inheriting NUTEK ‘linear’ and recruiting new ‘interactive’ innovation
professionalsthe new economic governance agency still contained two cultures and
related administrative ‘silos.’ Hence different parts of the agency did not necessarily
know what other parts were doing. However, an outsourced advisory recommendation
found favour to the effect that VINNOVA should construct a Digital Knowledge Flows
Platform (DKFP) as a knowledge management system for the internal and external use of
the organisation along lines first proposed by Hansen (2002). By late 2004, an intranet
system had been constructed and populated using a Geographical Information Systems
program format. Thus, the system contained locational data on all innovation nodes and
networks in Sweden and its regions, with detailed firm data layered beneath map-based
geographical information. On to these bases were added policy information, types of
measures and financing regimes. This database GIS is accessible by all VINNOVA
employees and successfully adapted so that knowledge may be added in structured form
by any authorised member of the workforce, thus acting as a policy-based knowledge
management system or DKFP. Subsequently and subject to data protection standards, it
was made accessible to actors external to VINNOVA, particularly in the new regional
innovation systems being formed. In time, management attention will have to be given to
an incentivisation scheme and knowledge management function to ensure the platform
attracts knowledge users and producers of consequence to Sweden’s regional and
national innovation systems.
To some extent, therefore, connectivity was the key barrier to be overcome internally
and externally in both the Scottish and VINNOVA systems. The assistance to
entrepreneurship signified by the ITIs in Scotland is paralleled by a heightened desire to
improve ‘talent’ retention by generating employment opportunities resulting from
successful entrepreneurship in innovative fields. To that extent, an advanced knowledge
management system may support the acquisition of inputs necessary to improving
the quality of complicit knowledge for regional innovation. However, it cannot substitute
for, merely augment the kind of rich communicative gains that accrue to face-to face
communication. As we have seen in numerous contexts discussed in this paper, a
well-functioning regional innovation system capable of contributing to regionally
constructed advantage has smooth knowledge communication flows within and across
communities of practice at its heart. In some cultures, historic trajectories have endowed
regions with good interpersonal communication skills and networking practices
(see Iammarino, 2005). But in others, for reasons of peripherality and low density
settlement patterns or an inclination towards economic ‘individualism’ if not
‘egocentrism,’ knowledge flows less unhindered than it might. In such circumstances,
‘virtual networks’ may enable innovation barriers to be eroded somewhat.

4 Conclusions

It is beginning to become evident that for regions to be able to construct advantage for
themselves, as distinct from institutional borrowing of possibly inappropriate policies
from elsewhere, they have to be much more outward-looking from their innovation
system ‘icebergs’ than hitherto. This is true not simply for regional innovation systems
but, in high technology sectors at a European scale too. Entrepreneurship is a weakness
in Europe, especially ‘knowledge entrepreneurship’.25 So much so in, for example,
 Regional innovation, entrepreneurship and talent systems 133

biotechnology that the UK’s leading biotechnology venture capitalist who only in 1999
was the strongest advocate that UK and European biotechnology businesses would soon
be producing abundant new therapeutic treatments, has now recanted. Performance had
been so disappointing thereafter that he foresaw a new ‘trend to America.’ This was
altering the ambitions of his own venture capital business as well as those of the UK
biotechnology industry. Hence, instead of creating major UK biotechnology companies,
the UK’s scientists and early-stage companies will act as a ‘conveyor belt of quality
medical projects to sell to America.’26 Thus, the companies created by the Merlin
Biosciences venture fund in future will employ US scientists and managers with firms
establishing a presence there at an early stage: ‘Designed and built in Britain to grow in
America,’ he continued. Such decapitation strategies, as they are termed in the industry,
are logical where IRIS architectures remain insufficiently entrepreneurial compared to
the ERIS set-ups in the USA, possibly prefiguring a future trajectory for European
science-driven research (Ward, 2005). This implies a strong regional platform for
knowledge exploration but global ‘entrepreneurship outsourcing’ regarding knowledge
exploitation and, presumably knowledge examination. New policy options will include
assisting the search function for members of the commercialisation-outsourcing ‘research
industry’. Policies instruments to recruit and import ‘serial entrepreneurs’ from ERIS
settings are a logical next step, but one that needs to be fully informed of the asymmetric
entrepreneurship ‘regimes’ prevailing not only between Europe and elsewhere, but
among the member-states that constitute it.
Secondly, the portrayal of the evolution of regional innovation systems studied
longitudinally in Figure 1 showed most systems had evolved a trajectory taking
them towards the ‘networking’ centre of the model, building constructed advantage
by improving communication capabilities. 27 Of key importance here are elements
stressed in the above analysis, notably recognition of institutional and governance
capabilities in regions, recognising variation in the quality of regional communication
and connectivity infrastructures, understanding the knowledge base strengths of the
region and presuming that true regional innovation system connectivity is not complete
in most regions (as the case of Scotland revealed). Action to construct regional advantage
utilising these perspectives is, in policy terms, an imperative as we have seen. The role of
innovation platform policies combining instruments across not only the above elements
but also entrepreneurship, talent and quality of life issues are imperative to control the
spectre of decapitation. In this, structural embedding and connectivity to distant
networks has become more of a priority than it was a decade or so ago when the idea of
28
regional innovation systems (mainly IRIS) was in its infancy.
This, in turn, has clear policy implications as has been shown. Piccaluga (2005)
suggests that it is methodologies rather than recipes that policy makers nowadays seek
as the idea of the ‘learning region’ recedes into the past. Too often regional learning
sought to ‘clone’ successful substantive examples from elsewhere, usually inappropriate.
Thus, the notion of policy ‘platforms’ that include an array of instruments appropriate to
constructing regional advantage by enabling firms to be highly knowledgeable and with
global connectivity but also integrated with different policy levels is unavoidable.29 Such
policies must interact, as we have seen, with issues such as compensating for market
failures, enhancing the quality of regional talent pools for global sourcing and
outsourcing, and policies for enhancing regional knowledge flows also included.
Policy auditing such that ‘conventional wisdoms’ (like the attractiveness of academic
spinouts as a regional innovation model) are challenged in light of performance impact is
134 P. Cooke

essential and sometimes prosaic pathways such as growing existing firms rather that
growing business birth rates reprioritised. In the platform toolkit are necessities like
mission policies supporting targeted finance for research into cutting-edge technologies
(as with the Scottish ‘Science Strategy’ and supporting ITIs for commercialisation) as
well as the more pervasive, diffusion policies associated with innovation platform and
framework support. Equally, a balance between firm and ‘milieu’ strengthening has to be
struck. These take a variety of forms and vary in relevance and impact according to
regional characteristics. Where a regional university has world-class capabilities support
to network it into a client base is indicated. Where there are SME cluster strengths but no
such knowledge centres, building regional competence centres is implied. Where SMEs
exist but are isolated, network brokerage has been shown to yield innovation payoffs in
such countries as Denmark and Norway (Cooke, 2002). Talent is in short supply nearly
everywhere, so policies to assist its formation, recruitment and retention are obviously
valuable policy contributions. But this needs to be done in such ways as to engage with
entrepreneurship to moderate the ‘decapitation’ solution or alternatively, retheorise
European regional advantage around strategies further to strengthen the world-class
core of the European research industry and build new markets for inward global
knowledge outsourcing. That the latter strategic option has never appeared on the
European policy radar is testimony to a collective failure of imagination of truly
continental proportions. Above all, the analysis and evidence mobilised in this paper
point to a future in which innovation, talent-formation and entrepreneurship have to be
considered in triplicate to construct regional advantage in ways that intersect profitably
with regional, national and global innovation imperatives.

Acknowledgements

Much of the thinking contained in this paper arose from engagement with the issues of
regional constructed advantage mooted to me first by John de la Mothe of PRIME
at the University of Ottawa Management School, thereafter explored in policy terms
in the European Commission on Constructing Regional Advantage DG Research
commission which I chaired. I would thank all members but especially Dimitri Corpakis
and Jean-Marie Rousseau for their initiative. Some data arose from the ESRC research
project ‘Collective Learning in Knowledge Economies: Milieu or Market?’ Grant No.
RES-000-23-0192. The sponsors and my colleagues Carla De Laurentis and Oliver Ehret
are acknowledged for their assistance. I gave numerous conference presentations on
these and related issues in 2005. I am grateful to Bjorn Asheim, Ed Bergman, Ron
Boschma, Meric Gertler, Åsa Lindholm Dahlstrand, Andrea Piccaluga, Maureen
McKelvey and Javier Revilla Diez for their interest and advice. Finally, this paper
improved in clarity at key points because of useful referee comments. None bears any
responsibility for the result.

References
Akerlof, G. (1970) ‘The market for ‘lemons’: qualitative uncertainty and the market mechanism’,
Quarterly Journal of Economics, Vol. 84, pp.488–500.
Best, M. (1990) The New Competition: Institutions of Industrial Restructuring, Cambridge: Polity.
 Regional innovation, entrepreneurship and talent systems 135

Brenner, T. (2005) ‘Innovation and cooperation during the emergence of local industrial clusters:
an empirical study in Germany’, European Planning Studies, Vol. 13, pp.921–938.
Brown, J. and Duguid, P. (2001) The Social Life of Information, Boston: Harvard Business School
Press.
Carlsson, B. (2006) ‘Innovation systems: a survey of the literature from a Schumpeterian
perspective’, in H. Hanusch and A. Pyka (Eds). The Elgar Companion to Neo-Schumpeterian
Economics, Cheltenham: Edward Elgar.
Chesbrough, H. (2003) Open Innovation, Boston: Harvard Business School Books.
Colletis (1993) ‘An analysis of technological potential and regional development processes in
Rhône –Alpes’, European Planning Studies, Vol. 1, pp.169–180.
Cooke, P. (1985) ‘Regional innovation policy: problems and strategies in Britain and France’,
Environment and Planning C, Vol. 3, pp.253–267.
Cooke, P. (1992) ‘Regional innovation systems: competitive regulation in the new Europe’,
Geoforum, Vol. 23, pp.365–382.
Cooke, P. (Ed) (1995) The Rise of the Rustbelt, London: UCL Press.
Cooke, P. (1998) ‘Introduction: Origins of the concept’, in H. Braczyk, P. Cooke and
M. Heidenreich (Eds). Regional Innovation Systems, London: UCL Press.
Cooke, P. (2001) ‘New economy innovation systems: biotechnology in Europe and the USA’,
Industry and Innovation, Vol. 8, pp.267–289.
Cooke, P. (2002) Knowledge Economies, London: Routledge.
Cooke, P. (2004) ‘Introduction: Regional innovation systems – an evolutionary approach’, in
P. Cooke, M. Heidenreich and H. Braczyk (Eds). Regional Innovation Systems, 2nd edition,
London: Routledge.
Cooke, P. (2005a) ‘Regional asymmetric knowledge capabilities and open innovation: exploring
‘Globalisation 2,’ a new model of industry organization’, Research Policy, Vol. 34, pp.1–22.
Cooke, P. (2005b) ‘Rational drug design, the knowledge value chain, and bioscience megacentres’,
Cambridge Journal of Economics, Vol. 29, pp.325–341.
Cooke, P. (2005c) ‘Contrasts between biotechnology and ICT clustering: a UK study’, Presented at
the Regional Studies Association conference, Regional Growth Agendas, 28–31 May,
Aalborg.
Cooke, P. and Morgan, K. (1994) ‘The creative milieu: a regional perspective on innovation’,
in M. Dodgson and R. Rothwell (Eds). The Handbook of Industrial Innovation, Cheltenham:
Edward Elgar.
Cooke, P. and Schall, N. (2006) ‘Schumpeter and varieties of innovation: lessons from the rise of
regional innovation systems research’, in H. Hanusch and A. Pyka (Eds). The Elgar
Companion to Neo-Schumpeterian Economics, Cheltenham: Edward Elgar.
Cooke, P., De Laurentis, C., Tödtling, F. and Trippl, M. (2006) Regional Knowledge Economies,
Cheltenham: Edward Elgar.
Cooke, P., Heidenreich, M. and Braczyk, H. (Eds) (2004) Regional Innovation Systems,
2nd edition, London: Routledge.
Dierickx, I. and Cool, K. (1989) ‘Asset stock accumulation and sustainability of competitive
advantage’, Management Science, Vol. 35, pp.1504–1511.
Doeringer, P., Evans-Klock, C. and Terkla, D. (2005) ‘Management cultures and regional
development: high performance management and the location of new manufacturing plants’,
European Planning Studies, Vol. 13, pp.815–830.
Florida, R. (2002) The Rise of the Creative Class, New York: Basic Books.
Foley, S. (2005) ‘Can Merlin’s chief cast his spell on investors again?’ The Independent, 16 July,
p.40.
Foray, D. and Freeman, C. (1993) Technology and the Wealth of Nations: The Dynamics of
Constructed Advantage, London: Pinter.
Geertz, C. (1973) The Interpretation of Cultures, New York: Basic Books.
136 P. Cooke

Gertler, M. (1995) ‘‘Being There’: Proximity, organization, and culture in the development and
adoption of advanced manufacturing technologies’, Economic Geography, Vol. 71, pp.1–26.
Gertler, M. (1996) ‘Barriers to technology transfer: Culture and the limits to regional systems of
innovation’, EU-RESTPOR Conference, 19–21 September, Brussels.
Granovetter, M. (1992) ‘Problems of explanation in economic sociology’, in N. Nohria and
R. Eccles (Eds). Networks and Organisations: Structure, Form & Action, Boston: Harvard
Business School Press.
Hansen, M. (2002) ‘Knowledge networks: explaining effective knowledge sharing in multi-unit
companies’, Organisation Sciences, Vol. 13, pp.232–248.
Helpman, E. (Ed) (1998) General Purpose Technologies and Economic Growth, Cambridge: MIT
Press.
Iammarino, S. (2005) ‘An evolutionary integrated view of regional systems of innovation:
concepts, measures and historical perspectives’, European Planning Studies, Vol. 13,
pp.497–520.
Isaksen, A. (1997) ‘Regional clusters and competitiveness: the Norwegian case’, European
Planning Studies, Vol. 5, pp.65–76.
Krugman, P. (1995) Development, Geography and Economic Theory, Cambridge: MIT Press.
Longhi, C. and Quéré, M. (1992) ‘Local systems of innovation: Analytical and empirical
discussion’, Paper to ‘Systems of Innovation’ Conference, October, Bologna.
Malmberg, A. and Maskell, P. (1997) ‘Towards an explanation of regional specialisation and
industry agglomeration’, European Planning Studies, Vol. 5, pp.25–42.
Malmberg, A. and Maskell, P. (2002) ‘The elusive concept of localisation economies: towards a
knowledge-based theory of spatial clustering’, Environment and Planning A, Vol. 34,
pp.429–449.
Marshall, A. (1920) Principles of Economics, London: Macmillan.
Norton, R. (2000) Creating the New Economy: the Entrepreneur and the US Resurgence,
Cheltenham: Edward Elgar.
OECD (2003) Science and Technology Outlook 2003, Paris: OECD.
OECD (2004) Science and Technology Outlook 2004, Paris: OECD.
Orlikowski, W. (2002) ‘Knowing in practice: enacting a collective capability in distributed
organising’, Organisation Science, Vol. 13, pp.249–273.
Owen-Smith, J. and Powell, W. (2004) ‘Knowledge networks as channels and conduits: the effects
spillovers in the Boston biotechnology community’, Organization Science, Vol. 15, pp.5–21.
Piccaluga, A. (2005) ‘Typology of regional innovation policies and tools’, Presentation to
Constructing Regional Advantage Commission Meeting, Brussels, DG Research.
Sapsed, J., Gann, D., Marshall, N. and Salter, A. (2005) ‘From here to eternity? The practice of
knowledge transfer in dispersed and co-located project organisations’, European Planning
Studies, Vol. 13, pp.831–852.
Späth, L. (1985) Wende in die Zukunft, Stuttgart.
Van den Biesen, J. (2004) ‘University-industry relations and innovation strategy in Philips
worldwide: an R&D outsourcing approach’, Presentation to EU Conference ‘The Europe of
Knowledge 2020: a Vision for University-based Research and Innovation’, Liège Convention
Centre, 25–28 April.
Van der Linde, C. (2003) ‘The demography of clusters – findings from the cluster meta-study’,
in J. Bröcker, D. Dohse and R. Soltwedel (Eds). Innovation Clusters and Interregional
Competition, Berlin: Springer.
Ward, M. (2005) ‘Strategy: decapitation’, BioCentury, Vol. 13, pp.A1–A4.
Wright, M., Robbie, K. and Ennew, C. (1997) ‘Serial entrepreneurs’, British Journal of
Management, Vol. 8, pp.251–268.
Zucker, L., Darby, M. and Armstrong, J. (1998) ‘Geographically localised knowledge: spillovers or
markets?’ Economic Inquiry, Vol. 36, pp.65–86.
 Regional innovation, entrepreneurship and talent systems 137

Notes
1
This terminological shift is not mere semantics. Information is unlike knowledge in that it has no
intrinsic meaning until knowledge is able to interact with it. Meaningful knowledge facilitates
action. The analogy is the train timetable which only supplies knowledge when the journey
with meaning for the passenger is derived from the passive information set presented. Action
may then ensue by catching the appropriate train.
2
For a parallel exercise, see Longhi and Quéré (1992).
3
For example, in former East Germany and Switzerland. It is a notable cultural specificity of the
Steinbeis knowledge transfer model that it does not take root outside German-speaking
countries or areas of countries. This is examined by Meric Gertler (1996) and Gertler (1995).
4
Epistemic communities are professional associations the discourse of which to outsiders is by no
means transparent. Hence, despite ‘triple-helix’ presumptions, there are often cognitive
barriers among civil servants, academics and entrepreneurs that need transcending, as indeed
there are between, for example, chemists and biologists, within academe. The latter barrier is
often cited as one reason why pharmaceuticals firms with origins in fine chemistry are
generally less competent in biotechnology, from which an increasing number of drugs derive,
than small biopharmaceuticals firms. Similarly, institutional conventions associated with
specific countries may not be transferable except into other countries with similar institutional
conventions that enable adaptive learning to occur.
5
For a full inventory of the assets of the Rhône–Alpes innovation system at the time, see
Colletis (1993).
6
Wrong footnote numbers in a fascinating quest to find the world’s oldest cluster, Van der Linde
(2003) selects Oyonnax in the Rhône–Alpes region. ‘Plastics Valley’ as it is also known
“…can trace its history back to the year 630 when the village was given a monopoly to make
combs. Wooden combs were followed by celluloid combs….when combs fell out of fashion in
the 1920s and 1930s, the unemployed craftsmen turned to the production of plastic toys and
eyewear lenses….a further innovation was the adoption of injection moulding in 1936….
By 1998 600 firms…workforce of 12,700 produced €1.8 billion worth of plastic toys and
eyewear…’ (p.140).
7
Späth (1985).
8
Wrong footnote numbers Philips is now headquartered in Amsterdam.
9
This includes the remarkable DSP Valley ‘platform’ which connects internationally leading edge
Digital Signals Processing exploration knowledge from the Catholic University of Leuven,
Belgium and advanced engineering systems exploration work at the Technical University of
Aachen, Germany, to examination and exploitation knowledge resources at nearby Philips in
Eindhoven, Netherlands. Entrepreneurial firms from both universities perform key roles in
translational research between exploration and exploitation in this cross-border regional
innovation system, containing numerous clusters displaying ‘related variety’. It is supported
by ‘platform policies’ in Leuven and Aachen sub-regions that contain instruments relating to
entrepreneurship, talent-formation, risk investment, cluster networks and quality of life issues.
10
Wrong footnote number it is also noteworthy that regions in Figure 1 give a tendency to move
towards ‘networking’. This is because dynamic competition through innovation (as distinct
from static competition through price cutting) engenders a requirement for greater interaction
by firms with other firms and organisations. Aware of this tendency policy actions tend to
follow suit or in some cases, anticipate it through engaging in partnership or networking
organisational and firm interactions for innovation support (e.g. as in DSP Valley).
11
Serial start-up companies are a strong indicator of the presence of mature entrepreneurship in a
region. This is because serial entrepreneurs have scarce capabilities to establish, often
successfully, innovative ventures which are of interest to them as successive revenue earning
investments not because of any particular attachment to technology, market or company
purpose. Wright et al. (1997) identify three types of serial entrepreneurial behaviour namely
venture repeaters, organic serials and serial dealmakers, distinguished by the motivations and
the methods used to develop their ventures.
138 P. Cooke
12
This was publicised, along with Philips’ new ‘Sense & Simplicity’ marketing slogan eschewing
historic emphases on technological sophistication in marketing, in a presentation by Philips
R&D chief Van den Biesen (2004).
13
A sceptic might say this exposition of ‘open innovation’ is rather uncritical, for example, of
problems of coordination implied by such extended outsourcing practices. However,
coordination of information but also of knowledge is arguably one of the great efficiency and
effectiveness gains from ICT. That is not to say that issues of implicit–explicit knowledge
translation are simply ‘solved’ in this model. However, open innovation means what it says in
that the neglected intermediary ‘translational’ form referred to in the DSP Valley case, namely
‘complicit’ knowledge, is brought frontstage (see next section). Indeed this is partly the
point of seeking locations in ‘global talent pools’ where complicit knowledge may be
accessed for translational research from specialised local knowledge markets and naturally,
exposed to testing and trialling wherever, possibly R&D HQ, is deemed appropriate
(see Cooke, 2005a,b).
14
Dierickx and Cool (1989a).
15
Malmberg and Maskell (1997).
16
Malmberg and Maskell (2002).
17
Cooke (2005a) where it is shown, schematically, that the complicit part of knowledge
transformation in biotechnology involves at least fifteen different ‘communities of practice’
before implicit knowledge of a therapeutic treatment may become an innovation utilised in
patient healthcare.
18
Cooke (2005c).
19
Marshall (1920).
20
Isaksen (1997) also shows the marine engineering cluster near Stavanger outperformed the
Norwegian average for employment and turnover growth. Hence clustering (and ‘related
variety’) in this and some, but not all, other industries was associated with greater domestic
and global competitiveness.
21
Interviews were conducted with ITI heads as part of the ESRC project ‘Collective Learning in
Knowledge Economies: Milieu or Market?’ Grant No. RES-000-23-0192.
22
Thus, as a case in point, the OECD STI Outlook 2004 reports that the following are the R&D
Priorities of several OECD members: Canada – Biotechnology, ICT, Nanotechnologies; J
apan – Life Sciences, ICT, Nanotechnology; S. Korea – Electronics and Biotechnology;
Mexico – ICT, Biotechnology; New Zealand – Biotechnology, ICT and Creative Industries.
Meanwhile in the USA, more ‘pervasive’ R&D Priorities have been set such as ‘Networking
IT’ involving large-scale networking, high confidence software systems, software
design productivity, social, economy and workforce applications and human–computer
interaction. Other ‘pervasive’ platform priorities include, complex systems, climate, water and
hydrogen, homeland security and long-term nanotechnology applications, for example, in
information processing.
23
The notion of clusters as ‘quasi-monopolies’ is not entirely new, Dierickx and Cool (1989) and
Krugman (1995) having suggested as much. However, a few have explored this idea
empirically. In a comparative research project on ICT and Biotechnology clusters and firm
performance under different ‘varieties of capitalism’ (UK and Austria; Cooke, 2005c;
Cooke et al., 2006) firms were asked a battery of questions about collaboration (market and
non-market). They were also asked if they perceived they were located in a (localised lateral
and vertical linkages) ‘cluster’. For ICT firms, 56% (71 of 126 non-collaborator respondents)
perceived themselves to be located in clusters as predefined. In Biotechnology it was 22%.
We deduce these to be ‘excluded’ seekers of knowledge spillovers under quasi-monopoly
conditions in which valuable localised knowledge nevertheless ‘leaks’ informally, as observed
by Owen-Smith and Powell (2004).
24
What is the difference between ‘epistemic communities’ and ‘communities of practice’.
Fundamentally, the former are distinguished by their professional certification, interests and
semantic discourses, whereas the latter are more differentiated by their relational
 Regional innovation, entrepreneurship and talent systems 139

embeddedness. Recall Brown and Duguid’s (2001) telling illustration whereby Silicon Valley
electronics engineers in different firms had more in common cognitively and in terms of
practical conventions than electronic engineers in the same firm (Xerox) located in Texas.
25
This means creating businesses (including university centres of excellence) with markets for
a broad array of specific kinds of scientific, technological (or cultural) knowledge.
Research-driven, such businesses are already present and R&D outsourced to them in
increasing numbers in the USA and especially in leading technology sectors. We may speak
therefore of a ‘research industry’ having evolved, precisely the kind of industry in which the
West may continue to keep a competitive edge not least due to its historic sunk costs in higher
education and research.
26
Foley (2005) a quote from Sir Chris Evans, head of Merlin Biosciences, envisaging a way out of
the UK’s entrepreneurship deficit in biotechnology. This will be achieved by outsourcing high
grade R&D to US companies to provide the complicit know-how to transform it into codified
therapeutic treatments aimed at the US market. This is a good example of the rise of a
specialist ‘research industry’ in an economy, that is modest in its biotechnology
commercialisation achievements. Of course, time will tell if this happens and whether it is
confined only to biotechnology.
27
EU DG Research established a commission (chaired by this author) to investigate processes by
which regional constructed advantage around innovation, entrepreneurship, talent (and culture,
including good quality environment) interacted. Philippe Larédo presented a paper in which
he stressed connectivity quality as a sine qua non of accomplishment in this regard.
Contemporaneously, Andrea Piccaluga presented for discussion a paper typologising regional
innovation policies. Both papers informed these concluding remarks.
28
Relational and Structural Embedding are distinguished by Granovetter (1992) to denote
‘community of practice’ or ‘bonding’ social capital in the former and more institutional,
‘bridging social capital’ in the latter, as solutions to ‘epistemic community’ dissonances.
As Doeringer et al. (2005) showed, both are vital elements of accomplished regional
innovation systems.
29
The concept of ‘platform’ is common in technology studies, where, as noted, sensors have a
potentially pervasive role with multiple applications. This is where economic growth from
surplus profit is earned through a Schumpeterian ‘recombination’ category of innovation
based on a single invention (Cooke and Schall, 2006). More difficult and itself innovative is
the idea of policy platforms. They must be supple too, a feature deriving from their systemic
inclusion of policy networks representing key stakeholders, and their combination of
instruments representing multiple governance constituencies and ‘related variety’ of industry
mix. Implementation is assisted enormously by utilisation of GIS-based DKF platforms, an
analogy being with the practices of GIS/GPS precision farming, where variable seeding and
fertilising intensities are controlled according to prevailing soil humidity and fertility
conditions.