R E G IO N A L

---~ O N O M IC S
Second edition published 2016
by Routledge
2 Park Square, Milton Park, Abingdon, Oxon OX14 4RN
and by Routledge
711 Third Avenue, New York, NY 10017 ZYXWVUTSRQPONMLKJIHGFEDCBA
R o u tle d g e is a n im p r in t o f th e T a y lo r CBA
& F r a n c is G r o u p , a n in fo r m a b u s in e s s

© 2016 Roberta Capello

The right of Roberta Capello to be identified as author of this work has been
asserted by her in accordance with sections 77 and 78 of the Copyright, Designs
and Patent Act 1988.
A11rights reserved. No part of this book may be reprinted or reproduced or
utilised in any form or by any e1ectronic, mechanical, or other means, now known
or hereafter invented, including photocopying and recording, or in any
information storage or retrieval systern, without permission in writing from the
publishers. In memory of my brother Edoardo and my father Andrea,
T r a d e m a r k n o tic e : Product or corporate names may be trademarks or registered two eminent professors in Engineering
trademarks, and are used only for identification and explanation without intent ro
infringe.
First published 2007 by Routledge
B r itis h L ib r a r y C a ta lo g u in g in P u b lic a tio n D a ta
A catalogue record for this book is available from the British Library
L ib r a r y o f C o n g r e s s C a ta lo g in g -in -P u b lic a tio n D a ta
Capello, Roberta, 1962-
Regional economics I Roberta Capello.-Second Edition.
pages cm
Includes bibliographical references and index.
1. Regional economics. 1. Title.
HT388.C35 2015
338.9--dc23
2015020123
ISBN: 978-1-138-85587-8 (hbk)
ISBN: 978-1-138-85588-5 (pbk)
ISBN: 978-1-315-72007-4 (ebk)

Typeset in Sabon
by Apex CoVantage, LLC
9 Territorial competitiveness and
endogenous developmentvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA
Innovation and proximity

9.1 The endogenous sources of competitiveness:

innovation and proximity
Thus far, we have examined the role of space as a generator of locational advantages -
lower production and transaction costs, and a more efficient use of resources - that
enables firms to achieve higher levels of productivity and profit.
However, the effects of space on economic activity do not consist solely in improve-
ments to the sta tic efficiency of production processes (that is, an increase in firms'
revenues or a decrease in their costs); they are also manifest in the innovative and
creative capacity of firms. In this case, space is a so urce ofZYXWVUTSRQPONMLKJIHGFEDCBA
d y n a m ic e { fic ie n c y . Areas
with high concentrations of economic activity enjoy easy information exchange, fre-
quent face-to-face encounters, the presence of research and development activities and
advanced services, an availability of skilled labour, co-operativeness facilitated by
shared rules and codes of behaviour, and local social capital, which facilitate and
incentivize innovation by the firms located within those areas.'
These features are easily present in urban areas, which ha ve for this reason always
been recognized as the natural sites of innovative activity, the 'incubators' of new
knowledge; cities are the principal centres of research, given their large pools of exper-
tise, and the availability of advanced services (finance and insurance) ready to carry
the risk of any innova ti ve activity. Yet it is indisputably al so the case that certain non-
metropolitan areas of small size display an innovative capacity that persistently out-
strips that of other geographical areas, and they achieve levels of innovation sometimes
greatly disproportional to their manufacturing weight. They thus testify to the pres-
ence of some form of increasing returns to the concentration of innovative activity.
Cases in point are Silicon Valley in California, 'Route 128' in the Boston area, Badén-
Württenberg in the South of Germany, Jutland in Denmark, Srnáland in Sweden and
Sophia-Antipolis close to Nice, to cite only some examples.
Understanding these phenomena became of particular interest in the 1980s. In those
years, under the impetus of profound technological changes, innovation carne to be
considered the driving force of economic development, and knowledge the key factor
in local economic success. Instead, the uneven spatial distribution of innovative activ-
ity was taken to be the primary cause of regional imbalances. In periods when there
are evident signs of the hypermobility of labour and capital, the most immobile of
factors are knowledge and the intangible elements connected with culture, skill and
innovative capacity; it is on these elements that the competitiveness of local systems
depends.
 In n o v a tio n a n d p r o x im ity 237

From what has just been said, one understands why identification of the endoge- pro cess based on co-operation and the exchange of knowledge among actors. In most
nous, local conditions determining an area's innovative capacity became the mosr of the theories, co-operation is understood to be a result of an explicit decision by
important aim of regional development theories developed in the 1980s. These theo- economic actors to co-operate, and the different concepts of proximities help identify
ries differed sharply from the studies on the spatial diffusion of innovation discussed the criteria on the basis of which partners are chosen. Only the m ilie u in n o v a te u r
in Chapter 7; their primary aim was no longer to interpret innovation processes theory explicitly conceptualizes the exchange of knowledge as a spontaneous phenom-
through exogenous factors, but to identify the local endogenous determinants of inno- enon, which takes place through the socialization of knowledge at local level, even
vations. Their emphasis on elements endogenous to the innovative process fully justi. against the will of local actors. In this respect, the process of knowledge accumulation
fies their inclusion in this chapter. in the m ilie u in n o v a te u r theory differs from the interactive learning concept, and is
For these various theories, the endogenous determinants of innovation are increas- defined as a process of 'collective learning'.
ing returns in the form of dynamic location advantages deriving from:ZYXWVUTSRQPONMLKJIHGFEDCBA
Given the abundance of proximity concepts, it is necessary to understand the degree
of complementarity that exists among them, and the usefulness of the concept of physi-
• among firms, which facilitates the exchange of
s p a tia l, g e o g r a p h ic a l p r o x im ity cal proximity for interpretation of knowledge exchanges amid the plethora of new
tacit knowledge: this characterizes reflection by economic geographers concerned approaches. As we shaIl explain (Section 9.6), the theoretical bases of the different
to explain the concentration of innovative activities (Section 9.2); concepts of proximity differ substantially. Moreover, as we shall see, when some
• r e la tio n a l p r o x im ity among firrns, defined as interaction and co-operativeness approaches interpret the mechanisms behind proximity effects at locallevel, they refer
among local agents, the so urce of collective learning processes and socialization again to a concept of physical proximity.
to the risk of innovation (i.e. territorialized relations among subjects operating As for the theories of endogenous development associated with agglomeration
in geographical and social proximity): this was the approach taken by territorial Iorces, we do not criticize the qualitative nature of the approaches presented here; on
economists in explaining the dynamics of local systems in terms of local innova- the contrary, we think that they enrich the economic theory by highlighting the role
tive capacity (Section 9.3);2 of intangible e1ements (knowledge, learning, relational and social capital) in determin-
• in s titu tio n a l p r o x im ity , taking the form of rules, codes and norms of behaviour ing local competitiveness. Moreover, some of these theories present a new approach
which (i) facilitate co-operation among actors and therefore the socialization to the interpretation of local competitiveness. They in fact no longer embrace the
of knowledge and (ii) assist economic actors (individual people, firrns and local traditional functional approach, characterized by deterministic cause-effect relation-
institutions) to develop organizational forms which support interactive learning ships whereby the presence of a certain degree of knowledge in the region mechanically
processes: this aspect was emphasized by more systemic approaches seeking leads to innovation. They instead assume a relational approach, according to which
to understand the evolution of complex systems like the innovative system probabilistic e1ements - envisaged as the ways in which economic actor s perceive the
(Section 9.4); economic reality, react to external stimuli, and are capable of co-operative and syner-
• c o g n itiv e p r o x im ity among economic agents, interpreted as the existence of a gic behaviours - come into play in explaining when the existence of knowledge really
common knowledge base that guarantees mutual understanding among actors leads to innovation. These elements enrich the interpretation of the real world, and
characterized by complementary knowledge, as suggested by evolutionary eco- open the way to a more profound and sophisticated interpretation of economic
nomic geography in its explanation of the formation of c1usters of innovative development.
firms (Section 9.5). Therefore, from these theories it is possible to develop a new concept of 'regional
innovation patterns', a concept able to interpret the different modes through which
As we shall see, the development of these theories marks the overcoming of the regions innovate on the basis of the existence of the local preconditions for the genera-
simple view of pure geographical proximity as the explanatory e1ement of knowledge tion of knowledge and relationality among economic actors, both internal and exter-
exchange. During the mid-1980s with the theory of the 'm ilie u in n o v a te u r ', and in nal to the area (Section 9.7). It is on these 'regional innovation patterns' that modern
the 1990s with the French proximity school,' new and more profound analyses were innovation policies should be developed (Section 9.8).
added to the interpretation of local knowledge exchange, summarized in diverse
concepts of proximities - relational, cognitive, organized, social, technological -
without avoiding a certain confusion and an overIap among concepts. So me of these 9.2 Knowledge spillovers: geographical proximity
concepts were elegantly inserted into regional development theories - this is the case That innova ti ve activity has a natural tendency to concentrate in space has been
of relational, institutional and cognitive proximities - and it is around these con- confirmed by numerous empirical studies. Using both input indicators (e.g. spend-
cepts of proximities and on the theories on which they are based that this chapter is ing on research and development) and output indicators (e.g. number of patents)
organized." of innovative activity, these studies show that innovation is concentrated in central
The feature shared by all the approaches considered is that each concept of proxim- and metropolitan areas. Moreover, in all the industrialized countries, analyses of
ity is analysed in its capacity to reduce uncertainty associated with innovative activity the location of high-tech firms reveal marked polarization effects due to the pro-
and the solution of the co-ordination problem among actors acting individually. The nounced preference of these firms for central locations with strong sectoral
existence of proximity allows the development of 'interactive learning', a learning specialization."
 - · _ · _ • • . . •_. _ r • • • • ~ . _ _ "0"_- r o
•••••• _ _ , •• .1

Explanation of the phenomenon is straightforward: concentrated location facilitates and specialized knowledge. Once again, the results show that expenditure on local
exploitation of technological and scientific knowledge developed by research centres R&D is, for the majority oí sectors, more significant than expenditure on exter-
and universities; it gives easier access to the tacit uncodified knowledge required for nal R&D, and that diversified rather than specialized knowledge is important for
imitation and reverse engineering; and it ensures the ready availability of skilled labo ur local innovative capacity.
and advanced services.
Moreover, the complex and systemic nature of innovative processes explains their However, the theory can be criticized on various grounds. First, it should be borne
cumulative character: c1usters of incremental innovations follow an initial radical in mind that research and development expenditure and number oí patents are highly
innovation that marks out a 'technological trajectory' along which knowledge grows selective indicators of innovative capacity. Both capture only product innovations, that
and develops within well-defined technological boundaries. At locallevel, demand is, breakthroughs often associated with the innovative activity of large firms. They
for and the supply of innovative factors interact and mutually reinforce each other. entirely neglect the process innovation, the creative imitation and the reverse engineer-
Advanced firms enrich the surrounding environment by diffusing their technological ing that characterize the innovative processes of small firms.
and organizational expertise, while the surrounding environment simultaneously Even more dubious is the concept of space assumed by the theory. This space is
sustains their activity. The outcome is a cumulative polarization of research and purely geographical, a physical distance among actors, apure physical container of
innovation activities which reinforces the natural tendency for innovation to con- spillover effects which come about - according to the epidemiologicallogic adopted -
centrate in space. simply as a result of physical contact among actors. Important consequences ensue.
The role of agglomeration economies, both urban and sectoral, in explaining the First, this view is unable to explain the processes by which knowledge spreads at local
concentration of innovative activity was demonstrated long ago by Marshall. But level, given that it only envisages the probability of contact among potential innova-
interest in dynamic agglomeration economies (the agglomerative advantages that fos- tors as the source of spatial diffusion. Second, it concerns itself only with the diffusion
ter innovation by firms) has grown considerably in recent years, as recognition has of innovation, not with the processes of knowledge creation. It thus imposes the same
gained ground of the importance of innovation for the competitiveness of local limitations as did Hagerstrand's pioneering model in regard to the spatial diffusion of
systems. innovation; the diffusion of knowledge means adoption, and adoption means more
TheZYXWVUTSRQPONMLKJIHGFEDCBA
th e o r y o f te c h n o lo g ic a l s p illo v e r s developed in the 1990s linked the spatial innovation and better performance." Thus ignored, however, is the most crucial aspect
concentration of innovative activities with the increasing returns that concentrated of the innovation process: how people (or the context) actually learn. This is the aspect
location generates on those innovative activities themselves. Cross-fertilizations, of overriding interest not only for scholars but also, and especially, for policy-makers,
dynamic interactions between customers and suppliers, synergies between research should they wish to explore the possibilities of normative action to promote local
centres and local production units occur within circumscribed geographical areas like development.
highly specialized metropolitan areas. They do so as the result of the rapid exchange
of information and transmission of tacit knowledge made possible by face-to-face
9.3 Collective learning and the m ilie u in n o v a te u r :
encounters. In a concentrated location, the beneficial effects of a firm's research and
relational proximity
development activities are not confined within the boundaries of firrns, they 'spill
over' into the surrounding environment, to the advantage of innovative activity by
9 .3 .1 L o c a l s y n e r g ie s a n d r e la tio n s
other firms.
A large number of empirical analyses, mainly econometric, have successfully mea- In the 1980s, an international group of scholars set out to analyse the phenomenon of
sured the technological spillovers and the knowledge advantages enjoyed by spatiaIly the spatial concentration of smaIl firms. Their conclusion was that social interactions,
concentrated firms. Now briefly outlined are two of the methods employed to measure interpersonal synergies, and coIlective action among actors - in short, what they called
these effects.f 'relational proximity' - are the factors that account for the greater innovative capacity
oí spatially concentrated smaIl firms, and of the areas in which they are located.'? This
a) estimation of an aggregate knowledge production function at regional level, in
current of thought thus brought space as the generator of dynamic efficiency into the
order to verify the existence oí technological spillovers; or in simpler econometric
central focus of analysis on territorial development.
terms, to verify the existence of differing effects exerted by research and devel-
For this theory, economic and social relations among local actors condition the inno-
opment (R&D) activities, conducted within and without a region, on its patenting
vative capacity and economic success of specific local areas termed 'm ilie u x in n o v a -
activity.? The results confirm the existence oí spillovers from innovative activity,
te u r s ' .11 Synergies among actors are enhanced by spatial proximity and economic and
in that the significance of the parameter associated with local R&D is greater
cultural homogeneity, and thus produce dynamic advantages for small firms beca use
than that of the parameter for external R&D;8
they underpin processes of collective learning and socialization of knowledge.
b) estimation oí a disaggregated knowledge production function for individual local
Economic and social relations take two different forms in a m ilie u :
sectors that separately includes not only expenditure on local and external R&D,
but also expenditure on R&D by the same sector and by different ones. The
• a set of mainly informal, 'untraded' relationships - among customers and sup-
purpose is to determine the differing impacts on innova ti ve activity of diversified CBA
pliers, among private and public actor s - and a set of tacit knowledge transfers
240 D iv e r s ifie d -r e la tio n a l s p a c e vutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA
ln n o v a tio n a n d p r o x tm tty ¿41

that take place through job-mobility chains and inter-firm imitation processes. T a b le 9.1 Functions of the local milieu
These informal relationships have been widely studied b~ the French 'proximity
school', and they have recently been labelled 'untraded mterdependencies,.12 CBA
, C o n d itio n s G e o g r a p h ic a l p r o x im ity R e la tio n a l p r o x im ity
• more formalized, mainly trans-territorial co-operation agreements - among firms F u n c tio n s
among collective agents, among public institutions - in the field of technologicai -
R e d u c tio n o f Information collection/selection Information transcoding
development, vocational and on-the-job training, infrastructures and services u n c e r ta in ty Vertical integration within Selection of decision
provrsion, 'filieres' routines
Local signalling (collective Risk sharing among
marketing) partners
Relationships of the former type constitute the 'glue' that creates a m ilie u effect.
they are complemented by the latter, more formalized, kinds of relationship, whid~ R e d u c tio n o f Information collection Reduction of control
c o o r d in a tio n c o s ts Reduction of transaction costs costs through trust and
can be interpreted as 'network relations' proper. Both sets of relationships can be
(á la Williamson) loyalty
viewed as tools or 'operators' that assist the (small) firrn in its competitive endeavour Ex-ante co-ordination of Social sanctions on
enhancing its creativeness and reducing the dynamic uncertainty intrinsic to innova: day-to-day decisions (á la opportunistic behaviour
tion processes. Marshall) Ex-ante co-ordination
The partners in trans-territorial networks are selected single economic units _ in strategic decisión-
making
enterprises, banks, research centres, training institutions or local authorities - for
D u r a b le s u b s tr a te fo r Labour turnover within the Co-operation on
which location is only one co-ordinate among the many that serve to identify the unit.
c o lle c tiv e le a r n in g milieu industrial projects
At first glance, therefore, these networks merely link different economic actors and Imitation of innovation Tacit transfer of
have no necessary relation with space. But when the location of a unit takes on sig- practices knowledge
nificant meaning, inasmuch as it reveals a set of relations which generate territorial Public/private
development and identity (e.g. Apple at Cupertino, Silicon Valley), and when these partnerships in complex
development schemes
network relations start to multiply, they do indeed become territorial. When carefully
observed, the identity of the local m ilie u often prevails over the identity of the indi- Source: Camagni and Capello (2002)
vidual partner, which highlights the importance of the territorial aspect; the strategic
importance of links with a company in Silicon Valley resides more in the opening of customers and suppliers based on loyalty and trust. These relations produce a codified
a 'technological window' in Silicon Valley than in gaining access to that specific com- and tacit transfer of knowledge between customers and suppliers that triggers pro-
pany's know-how.':' cesses of incremental innovation and specific technological trajectories. Relations in
the local labour market likewise perform an important role in the local production
9 .3 .2 C o lle c tiv e le a r n in g and n e tw o r k c o -o p e r a tio n system beca use high turnover of skilled labour within the area and scant external
mobility cross-fertilize knowledge among firms and upgrade workers' skills. Finally,
'Relational capital' is defined as the set of norms and values that govern interactions firrn spin-offs - independent firms created by workers previously employed by a local
among people, the institutions where they are incorporated, the relationship networks firm - also participate in the knowledge socialization process.
set up among various social actor s and the overall cohesion of society. Relational capital The accumulation of knowledge in large firms is ensured by the presence of R&D
is therefore explaining the intensity of social interactions, interpersonal synergies and departments; and it is permanent beca use large firms are long-lived and develop their
collective action among local actors; that is, relational proximity. The latter has the own internal capabilities and cultures. By contrast, small firms have very short life-
same role in m ilie u theory as spatial proximity has in the knowledge spillover theory, cycles, with the consequence that they are unable to develop a solid stock of firrn-
in that it generates dynamic advantages taking the following forms (see Table 9.1):14 specific knowledge. This difficulty is remedied by the m ilie u and by the relations within it,
which guarantee continuity of knowledge through labour market stability, high people
• collective learning and socialization processes; mobility within the area, and stable relations between custorners and suppliers.
• reduction in the risk and uncertainty associated with the innovation process; In m ilie u in n o v a te u r theory, therefore, collective learning is the territorial counter-
• the e x -a n te co-ordination of routine and strategic decisions made possible by pan of the learning that takes place within firms. In large firms, knowledge and infor-
reduced transaction costs. mation are transferred via internal functional interaction among the R&D, production,
marketing and strategic planning departments.'! In m ilie u x , and in local small firrns
These functions are performed in a large firm by its R&D department, and rhey are systems, this function is performed by the already-mentioned high level of people
facilitated by internal diversification and complexity. A small firm finds the sarne [une- mobility, by intense innovative interactions between customers and suppliers, and by
tions in a highly specialized territory - as now explained.
firm spin-offs (Table 9.2).
Learning in a m ilie u takes place in spontaneous and socialized manner within the M ilie u theory f1anks these channels of learning available to firms with a third and
local labour market through forms of s ta b le a n d e n d u r in g collaboration betw een complementary one: learning through 'network co-operation' (Table 9.2). Through
 242 D iv e r s i{ ie d -r e la tio n a l s p a c e
In n o v a tio n a n d p r o x im ity 243
T a b le vutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA
9.2 Preconditions and channels for learning processes in innovative milieux
or simply to integra te priva te investment decisions), is a further element enhancing the
P r e c o n d itio n s C o n tin u ity innovative process in a m ilie u . Such co-ordination generally suffers from the avail-
D y n a m ic s y n e r g ie s
C o n te x ts (c h a n n e ls )
ability of limited and costly information, and frorn the possible existence of opportu-
F ir m s R&D functions nistic behaviour. The presence of the m ilie u reduces these costs beca use it enables
Functional interaction IN T E R N A L
Tacit transfer of knowledge
information to circulate more easily; it facilitates the taking of co-ordinated decisions
L E A R N IN G
T e r r ito r y Low mobility of through proximity and social homogeneity/cohesion; while it discourages opportunis-
High mobility of the
the labour force labour force within the
C O L L E C T IV E tic behaviour by fostering trust and threatening social sanctions. This last social/psy-
L E A R N IN G
outside the m ilie u chological element is crucial: it derives frorn the sharing of common values and of
m ilie u similar codes of behaviour, and it acts positively by developing trust and loyalty.
Stable linkages Co-operation for Conversely, it develops rapid processes of isolation and punishment for opportunistic
with suppliers innovation with behaviour.17
and customers suppliers and customers The influence exerted by Marshallian district theory on this approach is evident: the
Local spin-offs m ilie u theory reiterates the importance of geographical proximity, but even more so
N e tw o r k in g Stability as a Transfer of knowledge via L E A R N IN G of social and cultural proximity, in guaranteeing forrns of stable and enduring co-
consequence of co-operation
the complexity
THROUGH operation in small firm areas. For industrial district theory, these forrns of co-operation
of strategic N E T W O R K IN G give rise to a 'community market', the form of production organization which ensures
alliances the static efficiency of firms. For m ilie u theory, co-operation generates processes of
Source: Camagni and Capcllo (2002) knowledge socialization, and it reduces th e risk associated with innovation, and col-
lective learning - that is, factors of dynamic efficiency.
In recent years, econometric empirical analyses have corroborated the theory. In the
case of three m ilie u x in Italy, a production function was estimated using data collected
at individual firm level in which efficiency parameters of the production factors were
strategic alliances and/or non-equity co-operation agreements, firms acquire sorne of
connected to:
the strategic assets that they require externally, thus avoiding the costs of developing
thern internally. This knowledge-acquisition process stands midway between internal
• for labour: effects of collective learning, these being identified in the intensity
learning and collective learning, in that the firm comes into contact with the outside
of local spin-offs, and appreciation of the stability and quality of the locallabour
but still maintains a set of selected and targeted relationships. This form of learning
market;
assumes an important role in m ilie u in n o v a te u r theory beca use it permits local
• for intangible capital: effects of 'industrial atmosphere' and collective learning,
knowledge - which is produced by socialized and collective processes liable to isola-
these being identified in the importance to the individual firm of specialized
tion and lock-in - to enrich and innovate itself. Only through the co-operation with
knowledge internal to the local area, and th e lesser importance of acquiring
external firms that ensures an influx of new knowledge can a m ilie u avoid death by
knowledge from outside.
entropic uniformity. It is with this conceptual tool that th e theoreticians of the m ilie u
in n o v a te u r interpret the growth of small firms areas, among them the Marshallian
industrial district. T h e results showed that labour productivity is subject to increasing returns (given
the small average size of firms) that are substantially reinforced by the presence of
collective learning processes. Conversely, (intangible) capital productivity is subject to
9 .3 .3 B e y o n d c o lle c tiv e le a r n in g a n d n e tw o r k c o -o p e r a tio n decreasing returns, but is greatly augmented by an increase in th e appreciation and
use of local specialized knowledge (Figure 9.1).18
However, collective learning is not the only dynamic advantage generated for local
In terms of economic theory, the m ilie u in n o v a te u r approach has recently been
firms by the m ilie u , with its assets of relational capital. A further factor facilitating
indirectly validated by stylized analytical models a la Romer and Lucas.!? The rigidly
firms' innovative capacity is the reduction of th e uncertainty that accompanies innova-
neoclassical and aggregate forrn of these endogenous growth models distorts neither
tive processes. In large firms, the functions of information-gathering, the codification
the hypotheses nor the intrinsic logic of the m ilie u theory - which testifies to the latter's
of knowledge and th e selection of decision-making routines _ all of which are geared
ability to depict the endogenous economic laws underpinning the dynamic of local
to reducing sta tic and dynamic uncertainty - are performed by the R&D departme~t,
economic systems.
or by the planning unit. In the case of a m ilie u in n o v a te u r , they are undertaken tn
Finally, it should be pointed out that the m ilie u in n o v a te u r theory remedies the limi-
socialized and collective manner by the m ilie u itself, in which information rapidly
circulates beca use of geographical and collectiveCBA p r o x im ir y .t«
tation intrinsic to the theory of knowledge spillovers; it explains the channels through
which knowledge disseminates, not in terms of pure probability of contact, but rather
Finally, the reduction of the costs of e x -a n te co-ordination among decision-making
in those of well-evidenced econornic-territorial phenomena - supplierlcustomer rela-
units, and the facilitation of 'collective action' (undertaken to furnish collective goods
tions, high local labour turno ver, and spin-offs. The theory accordingly returns
 In n o v a tio n a n d p r o x im ity 245

a)CBA b) territorial factors to centre-stage in analysis of an area's endogenous innovative capac-

iry, and it reinstates space to the active role in the economic dynamic which the theory
Labour Labour
produetivity of the Marshallian industrial district attributed to it within a framework of pure sta tic
produetivity
efficiency.

9.4 The 'learning regions' and regional innovation systems:

institutional proximity

~ 9 .4 .1 T h e 'le a r n in g r e g ío n ' th e o r y

Labour The theory of the m ilie u in n o v a te u r has been paralleled by the international develop-
Labour market stability ment of wide-ranging analysis of the endogenous factors at the basis of local innova-
tive capacity. This approach has shifted its attention to institutional aspects, and
specifically to the set of social, economic and cultural rules embedded in a territorial
e) setting. It originated with the Danish school of Aalborg, and in the works of its
d)
founder, the economist Bengt-Áke Lundvall, and it has subsequently been widely
Labour adopted, mainly in the UK and the USA.20
Produetivity
produeti vi ty The main components of this approach can be summarized as follows. The prin-
of intangible
capital cipal resource of modern economies is knowJedge. Consequently, the principal pro-
cesses on which an economy's competitiveness depends are learning and the
acquisition of knowledge. Moreover, the complexity and systemic nature of innova-
tion, and the brevity of the product life-cycle characteristic of technoJogical change
in recent years, entail that learning is an interactive process. Put otherwise, learning
springs from co-operation and interaction between firrns and the local scientific
system, between different functions within the firm (between production and
research and development, between marketing and research and development),
Intangible
between producers and customers, and between firms and the social and institu-
Spin-offs
eapital tional structure. The feedbacks, interdependencies and complementarities among
the various functions internal to the firrn, and between the firm and external actors,
required by the innovative process evince the need for co-operative and interactive
e) f) forms of organizationallearning. Finally, innovation is increasingly the result of an
informallearning process, based on direct experience or that of others, which comes
Produetivity of Produetivity about through activities focused on finding solutions to specific technological, pro-
intangible of intangible ductive or market problems.
capital eapitalZYXWVUTSRQPONMLKJIHGFEDCBA
The consequence of these various features is that the innova ti ve process is strongly
localized: it results from the variety of traditions, norms, habits, social conventions
and cultural practices that constitute what has been called 'institutional rhickness'."
Innovation therefore cannot be understood properly unless it is examined within the
socio-cultural and institutional context in which it takes place. In areas where there is
'institutional proximity' - meaning the set of norms, codes and rules of behaviour
which help economic actors (people, individual firms, public and private institutions)
to adopt forms of organization that facilitate interact learning - the innovative process
Knowledge Knowlcdge comes about more rapidly and gives competitiveness to the economic system.22
internal to the external to thc
A 'Iearning region' is in this sense:
area area

F ig u r e 9.1 Collective learning and factor productivity • a regio n in which norrns of social and institutional behaviour support interactive
Source: Capello (1999b) learning: the horizontal organization of corporate functions, co-operation and
246 D iv e r s ifie d -r e la tio n a l s p a c e vutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA In n o v a tio n a n d p r o x im ity 247
agreements among firms, and co-operation between firms and research cen! of a complex syst~m o.f feedbacks ~nd retroactions among act?rs, and t~at ~he success
facilitate the exchange of knowledge and support the learning process' f ' novation resides m the capacrry to merge new technologIcal/orgalllzatlOnal sol u-
• a region with an 'organized market' in which implicit and generall~ h ~ ID with the potential demand for new products and production processes.
rules of behaviour guarantee the tacit exchange of information and the e s ~' tIOns h . lvsi h lati
1 fact the RIS t eory centres ItS ana ySIS on t e re ation
b etween two su b -systems
'. reatl,
of knowledge. These come about through an mteractrve and crea ti ve le . of ~cal ~ctors: the syste~ tha.t .creates ~nd diffu~es knowledge (the 'infr.astructure
proc~ss cent~ed o.n the i:nplemen.tat:on of ~ew products, a~d ne:-v techno~;: tem'), consisting of uruversities, public and pnvate research laboratones, on the
solutions. Likewise, a destructive learnmg process (a creanvs disruptiCBA I
syse hand, and the system of local firms, potential adopters (the 'business system') on
teaches the system to abandon obsolete and inefficient technological trajecto~O rhe other,zs The success of the local innovative activity lies, according to this theory,
and guides it through the profound .tra~sformation brought abour by ea, ~ rhe match and in the constant synergies between the new and crea ti ve technological
technological - all the more so organizational - changes imposed by a 01 I~lutions proposed by the 'infrastructure system' and the real technological needs of
technological trajectory.P ~ocalfirms. This view is based on local institutional e1ements, in particular norms and
behavioural codes, that support co-operation, already highlighted in the 'learning
In short, a 'learning region' is a socio-economic system able gradually to dev,
region' theory.
forms of interactive learning. It is on this 'learning ability' that a region's competiti· The RIS theory brings to the fore institutional proximity as a factor that reduces the
ness depends. It is accordingly a concept that identifies the condition necessary for risks and uncertainty associated with the innovative process. This result is achieved
economic system's competitiveness as a process (learning) more than a state (the without using economic-territorial elements that may emphasize the local nature of
of knowledge)." the process. Moreover, the supporters of the RIS theory strongly underline the irnpor-
Despite its necessary abstraction, which makes its empirical application impossi tance of the reinforcement of both the sub-systerns that compose an RIS: inefficiency
the concept of 'Iearning region' has gained general consensus, not just in a partí and lack of development of one of the two sub-systerns is interpreted as the source of
scientific community (that of Britain and North America) but also at institutional l weaknesses in the local innovative capacity. In this framework, innovation policies
given the European Union's need to devise new policy instruments with which to must be devoted to remedying this weakness."
port regional cohesion. However, the results obtained when the concept of '1 In regard to these norma ti ve suggestions, some words of caution are necessary:
region' is translated into regional economic policies are rather perplexing: the inn innovation may arise from external knowledge, and it may even be the result of an
tions proposed concern the creation of education and training services, incentives activity not necessarily based on research and development. The imposition of devel-
learning, the sharing of successful experiences in creating organizational forma oping formal knowledge-creation activities in all regions means pushing all of them
support interaction, and financial aid to firms undergoing corporate restructuriog: towards the same model of innovation, a strategy now widely recognized as
of which are interventions in support of weak regions which are well known unsuccessful.27
already applied in the past. It should also be borne in mind that the 'regional innovation system' was born of
Also it should be stressed that there is an apparently major weakness in the the concept of 'national innovation system"." The possibility of shifting the concept
Although it envisages a system of homogeneous socio-economic and institutional
&om one geographical level to another testifies to its necessarily aspatial nature, and
ditions in the region, and interaction and co-operation among actors, it is neve l e intrinsic impossibility of deducing the endogenous e1ements that underpin pro-
markedly aspatial. Nothing in the theory explains how and why these relatio~ s of territorial innovation from a theory like this one.
necessarily be local; nor does it explain what territorial conditions must be ID
for the 'organized market' to arise; or what territorial factors fuel the procese o'
active learning. .s Evolutionary economic geography and the concept
These shortcomings are all the more evident when one considers that the of 'related variety': cognitive proximity
of the 'Iearning region' is derived from that of the 'learning economy', and ~wards the end of the 1990s, a new stream of thought arose, taking the name of
concept of 'learning economy' is in its turn used to denote a 'n a tio n a l system
,~olutionary economic geography'. Its distinctive feature was that it centred interpre-
01

vation' where the set of institutional rules and norms allow, strengthen and e
on.of the dynamics of local areas on analysis of the birth and death of firms in a
forms of interactive learning. The fact that the concept can be shifted among
~ncal-evolutionary perspective. The innovative and locational choices of firms
territoriallevels of analysis demonstrates its aspatial nature. e analysed in a context of bounded rationality and interpreted within a theory in
a st
tr . with the assumption of perfect information of neoclassicallocation theories,
9 .4 .2 R e g io n a l in n o v a tio n s y s te m s lbWlth the inductive approach of institutional economic geography.
ess k evolutionary nature of this theory leads to the description of innovation and
Strongly rejecting the Schumpeterian idea that innovation is a linear proc. . nowledge development as resulting from a creative process of discovery devel-
of dif~erent and successive temporal phases of creation an~ transformattO;) O.
.:rou~d existing competences, within specific technological paradigms, and along
edge i n t o a tradable idea, the theory of regional innovanon systemS (fI h.traJectories.29 Limited by bounded rationality, firms are strongly influenced by
the opposite view: that innovation is the result of an interactive and non- Inl IStory, which influences both their innovative activities and their location
248 D iv e r s ifie d -r e la tio n a l s p a c e vutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA In n o v a tio n a n d p r o x im ity 249

choices. For this reason, new firms, usually spin-offs of other local firms, limit the kind separates the learning process from the economic and territorial context, and
uncertainty of their choice by locating in the same area; already-existing firms sho w imposes a step backward with respect to the theory of m ilie u in n o v a te u r , which high-
an evident location inertia due to the long and stable relationships that they have lighted the elements in the local context that make it possible to generate a collective
established with local suppliers and customers. The result of locational decision- learning process.
making processes of both new and already established firrns is a concentration of firms
belonging to the same industry in the local area.
9.6 General rernarks on the concepts of proximity
For the first time in the history of spatial c1usters formation, industrial concentration
is not interpreted as a so urce of location advantages, but as a result of evolutionary The complementary aspect of the di verse elements at the basis of knowledge transfer
processes that follow highly structured and organized routines intended to reduce the emerges from the different theories presented: from context conditions in the form of
uncertainty that accompanies decisions in bounded rationality conditions.P social and institutional rules and governance that push towards co-operation, through
Spatial concentration (or its absence) is not only the result of a historical industrial behavioural and subjective (relational) elements that facilitare the involuntary exchange
process but also determines its future evolutionary trajectories. In fact, the probability of knowledge, to cognitive elements linked to a constructive co-operation based on a
that local firms can survive depends, in this perspective, on their capacity to exploit background of common knowledge. The presence of these proximities generates dif-
the information present in the area. This capacity, in its turn, depends on the existence ferent positive externalities that reinforce knowledge transfer: a reduced risk of oppor-
of a common knowledge base within the industry. runistic behaviours, the limitation of uncertainty, reduced transaction costs and
It is in this way of reasoning that the concept of cognitive proximity emerges as an common understanding of technological aspects (Table 9.3).
element crucial for explaining innovation capacity. In order to innovate through the Complementarities among the different concepts of proximity are also c1ear, not-
knowledge that exists at the local level, it is necessary for firms to be endowed with withstanding some conceptual overlaps. Knowledge transfer requires at the same
the complementary knowledge necessary to be creative and generate new and innova- time relational capacity among actors, norms and rules of behaviours, and mutual
tive technological solutions. AII this, however, must take place on a common knowl- trust (a condition guaranteed by the interaction between relational and institutional
edge base that guarantees a common language and mutual understanding among proximities). Relational capacity is reinforced by a cognitive map shared by actors
firms. In the literature, this condition is labelled 'related variety', and it is defined as a (interaction between relational and cognitive proximities). Finally, the exchange of
variety of interrelated technological solutions with a common knowledge basis." complementary knowledge within a common knowledge basis is facilitated by rules
Although the concept of cognitive proximity was developed to explain local context and social norms that punish free-rider behaviour (interaction between cognitive and
forrnation, it has been applied to all forms of co-operation among firms, also long-dis- institutional proximities) (Figure 9.2).
tance co-operation. Interregional knowledge follows. This is generally formed through It should be borne in mind that the effects generated by the various proximities
firms' networks and requires a cognitive proximity between firms to generate innovative exhibit positive and negative non-linearities, as happens in all synergic processes.
projects in co-operation. AIso in this case, the theory of cognitive proximity suggests that
the greater the technological variety between two regions within a larger macro-industry,
the greater the benefit that these regions obtain from the exchange of knowledge. T a b le 9.3 A cornparison arnong the different concepts of proxirnity
Through implementation of a 'related variety' indicator at the regionallevel based
Types of D e fin itio n C h a n n e ls of P o s itiv e e x te r n a litie s R is k s a s s o c ia te d
on patent activities and a disaggregation of technological c1asses (e.g. five-digit disag- a s s o c ia te d w ith w ith to o m u c h
p r o x im itie s k n o w le d g e tr a n s fe r
gregation) within a larger technological c1ass (e.g. two-digit disaggregation), a number p r o x im itie s p r o x im ity
of empirical analyses ha ve identified a positive relationship between the degree of
Relational High degree of Econornic- ln-voluntaryexchange Risk of lock-in
'related variety' and an area's growth rate."
relationaliry of territorial of knowledge within local
Despite the continuing success of the concept of cognitive proximity, some critical knowledge
local actors elernents Reduced risk of
reflections on its advantages and shortcomings are necessary. The concept has certainly (suppliers- opportunistic
the great advantage of overcoming the simple idea - first propounded by Hiigerstrand custorners behaviour and
and then re-launched by the knowledge spillover theory - that the pure contact prob- relationships; Iirnited uncertainty
ability among actors can explain an exchange of knowledge. Moreover, the cognitive spin-off,
specialized
proximity concept enriches the concept of 'absorptive capacity' introduced into the
labour rnarket)
literature to explain the differing capacities to exploit knowledge of actors localized
Institutional Rules and Macroeconornic Reduced transaction Institutional
in the same area, and interpreted rather poorly as pure technological advancernent-" environrnent costs inertia
behavioural
The concept of cognitive proximity encompasses more subtle cognitive elements: by codes cornrnon supporting co-
simultaneously imposing a knowledge complementarity and a common knowledge to aIllocal operation
basis, it identifies cognitive capacity on the basis of both the specific technological agents
Shared The right rnix of Cornrnon Risk of lock-in
knowledge of single actors and the common aggregate knowledge of the area. Cognitive
knowledge industries understanding of within industry
Even if we recognize the merits of this concept, it has an important intrinsic limita-
technological aspects knowledge
tion: that of reducing cognitive aspects tOj:he industrv dirnension.CBA A nersriecriv» of rl .
 ¿50 ihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA
D iversified-relational space xwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA
Innovation and proxim ity 251

recent hermeneutic approaches to local creativity, are physical places, identitarian

relationships and collective learning processes highlighted as the sources of continuous
Territorial relations facilitated
innovative processes.:" The natural loci for these processes are identified in urban
by common and at the same time areas. Only with an approach like this one can space be given an active role in innova-
complementary knowledge
tion processes.
among actors

Socio-cconornic 9.7 Regional patterns of innovation

and territorial
c1cmcnts Complcmentary
Recently, the idea that the sectoral dimension is not sufficient to explain the regional
supporting knowlcdgc within
coopcration acommon innovation capacity has been strongly advocated. When applied at regionallevel, in
knowlcdgc basis fact, the sectorallogic shows two kinds of limits. The first one refers to the fact that
Rclational proximity Cogoitivc proximity
rhis logic pushes towards the interpretation of formal knowledge as the main source
of innovation; instead, we have previously seen how different theories have identified
the source of knowledge in a variety of informal elements, like face-to-face meetings,
'Territorialized' informal co-operation, creativity, collective learning processes. The second limit refers
relations facilitated Cooperation to the idea that only knowledge stemming from local sectors is a source of local inno-
by a local context
Institutional proximity among actors with vation; instead, a large literature has highlighted the role of knowledge coming from
of common rules a common
protecting non- outside the region as a fundamental source of innovation.
knowledge basis
opportunistic and with no In order to overcome such limits and return the local characteristics to the centre of
behaviours opportunistic the explanation of innovation processes, a new concept has recently be en proposed,
behaviours
that of regional patterns of innovation. This concept interprets the different modes of
innovation as the result of the presence/absence of contextual conditions necessary to
Common social create knowledge and to translate knowledge into innovation.P
rules protccting
non--opportunistic In fact, the concept of territorial patterns of innovation is proposed and defined as a
bchaviours combination of territorial specificities (context conditions) that líe behind different modes
of performing the different phases of the innovation process. In particular 'territorial
patterns of innovation' consist in spatial breakdowns of variants of the knowledge - - >
invention - - > innovation - - > development logical path built on the presence/absence of
Figure 9.2 Advantages from the presence of different proximities territorial preconditions for knowledge creation, knowledge attraction, and innovation.
The concept of territorial patterns of innovation therefore lies on a logical sequence
between knowledge, innovation and economic performance; it is therefore drawn in the
Whilst the .simul~aneous presence of the various proximities may multiply the advan- abstract but consistent Schurnpeterian 'linear model of innovation', even if heavily criti-
tages .obtamed, r t may also multiply the risks. Accordingly, decreasing _ or even cized as unrealistic, and rooted in the idea of a rational and orderly innovation process.
negatrve - returns may aris. e. A ~ th ~onze . d b y t h e Am i 'liteu m
. The local conditions are integral part of the innovation mode, and are interpreted
novateur approach, the
re~ence of o.nly local relationships risks keeping knowledge within the same techno- in this approach both as material elements, in the form of functions for the creation
ogical paradigm, thus reducing local creativity and the achievement of external flows of knowledge (R&D laboratories and universities), and non-material, intended as the
of new relational capacity of local actors. The pure existence of knowledge creation functions
. '.knowledge.'. Codes an d ru les strong l y em b e dd e d m . the local society may gener-
ate iostuunonal mertIa Fi 11 t hi h . . ". . . is not sufficient to guarantee an innovation process to occur; and it is not even the
. hi . na y, a 00 Ig cogmnve p r o x ir m t y risks generatmg lock-in
wit In the same industry k l d ("r bl 9 3) 34' . necessary condition, since knowledge can be acquired from outside. In order to explain
h . . now e ge la e . . Innovative processes, especially
tose of a radical nature, may be reduced or even become null this last aspect, the present theoretical framework relies on the most recent theories
Although the different d e fi ruit io
i d s a concept of'. relational space they on proximities presented above.
. n s move towar
remam anchored to a m t . d hi . ' For what concerns the territorial specificities (context conditions) that are behind
. . e rrc an geograp IC VISlOnof space; and they are far from a
~odern v~slon able to .sy~thesize the concept of 'territory'. Only within the Marshal- each phase of the innovation process, this theory takes advantage of the vast and
la~ theones of I?cal districts and m ilieu innovateur can geographical space become a articulated literature that takes territorial elements into consideration in innovation
social space (which takes th~ name of 't:rritory') where socio-economic relationships, processes, namely theories:
competences and culture, history and identiry merge and th us genera t e compe riti I IVe
a d vantages for the local community. Only in the m ilieu inno t th d . concerning know ledge creation: hurnan capital and education in general, uni-
va eur eory, an m sorne
versities and R&D activities, presence of an urban atmosphere have been
 considered, in a variety of approaches, as the territorial preconditions for endog-
enous knowledge creation in the vast literature that was developed during the
1980s (Section 8.3);
• concerning ihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA
know ledge diffusion, exploiting all concepts of proximities presented
above (Section 9.5);
• concerning know ledge receptivity and the capacity to turn know ledge into
innovation: local interaction and co-operation in order to achieve reduction of
uncertainty (especially concerning the behaviour of competitors and partners)
and of information asymmetries (thus reducing mutual suspicion among part-
ners); trust, sense of be1onging, place-loyalry and social sanctioning in order to
reduce opportunistic behaviour are all territorial e1ements, typical of the innova-
tive m ilieux, that increase the capacity of a region to speed up innovation and
take full advantage of collective learning processes and entrepreneurial activity
(Section 9.3 ).37

Exploiting the different theories, the regional patterns of innovation approach pro-
vides a deductive framework to interpret how the different phases of the innovation
process are put together at the spatialleve1, and why some of them take place in certain
areas and others do noto
Among all possible combinations of innovation modes and territorial e1ements, the
'archetypal' ones may be indicated in the following, each of which reflects a specific
piece of literature on knowledge and innovation in space:

• an endogenous innovation pattern, where local conditions fully support the

creation of knowledge, its local diffusion and transformation into innovation
and its widespread local adoption. Given the complex nature of knowledge ----- ----,
11 1
creation nowadays, this pattern is expected to show a tight interplay among "Ü)'I I
¡.... II
I
I

regions in the form of international scientific networks. From the conceptual ~~~::]
..2 ~ ~ II p..
:
I
point of view this advanced pattern is the one considered by most of the exist- ~ p..,"Oh I I ~ Il) I

ing literature dealing with knowledge and innovation creation and diffusion ]"§ -5 : : ~ ~ .g
:
.~ ~.fi I I .~ ~ I

(Figure 9.3); & 5 ~ ~ ::J r]:

__ - - - _' '_ - __ 1
• a creative application pattern, characterized by the presence of crea ti ve economic
actors interested and curious enough to look for knowledge outside the region -
given the scarcity of local knowledge - and creative enough to apply external
knowledge to local innovation needs. This approach is conceptually built on
the literature on regional innovation adoptionJadaptation (Figure 9.4);
• an im itative innovation pattern, where the actors base their innovation capacity
on imitative processes, that can take place with different degrees of adaptation
of an already existing innovation. This pattern is based on the literature dealing
with innovation diffusion (Figure 9.5).

Figures 9.3-9.5 show in a srylized way the three regional patterns of innovation envis-
aged before. As these figures show, regional modes of innovation are more comp\ex =o
-s.
than the simple core-periphery distinction suggested at the end of the 2000s; the latter ~
was encouraging core regions to be the natural places for general purpose technolo -
gies, which can achieve a critical mass of scientists and knowledge able to achieve MLKJIHGFEDCBA
 Phases Territorial preconditions for Knowlcdge ourput BA T e r r ito r ia l preconditions fo r I n n o v a tio n E c o n o m ic e ffic ie n c y
knowledge creation ínnovation

MLKJIHGFEDCBA
Region j 1--- - - - -- - - - - - -,
I Basie knowledge I

: (General Purpose :
E d u c a tio n ,
I Teehnologies, GPTs) I
hum an c a p ita l, 1- ~

a c c e s s ib ility ,
urban 1- - - - - - - - - - - - - -,

e x te r n a litie s : Speeifie and applied :

I knowledge I
1- 1

Territorial creativity

R e g io n i

Colleetive learning

r--------------, Product
E c o n o m ic

-:>
: Speeifie and : and
effieieney
I applied knowledge I p rocess

ihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA
: : Capabilities innovation
1- 1

Entrepreneurship

Figure 9.4 A crearive application pattern

Source: Capello (2012)
 Phases Territorial Knowledge ourput Territorial Innovarion Territorial Economic
preconditions preconditions precondlttons for efficiency
for knowledge for innovation innovation adoption
creation

Reglonj MLKJIHGFEDCBA
--------------, Collective learning
Education, : Basic knowledge 1

human capital, 1 (General Purpose : Product

accessibility, : Technologies, GPTs) I and

urban •• - - - - - - - - - - - - _1 process
externalities - - - - - - - - - - - - - -1
innovation
: Specific and applied I Entrepreneurship
: knowledge :
•• 1

RegioniihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA
Territorial
attractiveness:
FDls

Product
and Economic
process efficiency
innovation

Figure 9.5 An imitative innovation pattern

Source: Capello (2012)
 lnnovation Aa n d proxlm tty L .;)/ xwvutsrqpo

" " hich is meant the technological fields in which reg~ons are
increasing returns to R&D, giving 'peripheral' regions the role of co-inventorsMLKJIHGFEDCBA
of t
echnologtcal dom am s, by w" 1 1"" hould be tailored to promote local Innova-
applications in their technological domain. Moreover, the approach in terms o f ecialized an d to w h iICh reglOna 1""
po r e te s s
h Id be tailored to the regions' tec
hn 1 " 1
o ogica
sp 45 R " nal po retes s ou "1" "
regional patterns of innovation highlights territorial elements (context conditions) in tion processes. egio 1 1" tion processes in their fields of specia izanon.
in ord romote oca innova " 1"
the explanation of innovation processes." These context conditions are much more domains m or er to p "" di "translating the a-spatial smart specia Iza-
articulated than the mere presence of R&D activities, which is generally applied to
" his ai the marn mgre ients m "1"" "1 1
la achleve t IS airn, "" b ddedness of innovanon po retes m oca
" tial settmg are em e " " 1
measure the absorptive capacity of a region.P tion concept into ~ s?a li ka es. relatedness supporting a process of strategic spec~a -
Supported by empirical evidence,"? this approach shows that the way towards knowledge-tran~mlsSlOn m ~ nd ectedness between sectors and firrns which
innovation and modernization is very different among regions, and it derives in a " d diversificatlOn of sectors, an conn
Ize d"ff
c1ear way from local specificities. The variety of innovative modes that the empirical enables knowledge to 1 use. " lization strategy c1aim, the originality (~nd strength)
analysis reports highlights how misleading is a common normative strategy for all As the experts of smart specl~ " roach to regional innovatlOn (competi-
" " this highly mnovatlve app 1 h
regions in Europe, such as the achievement of the 3 per cent target of R&D invest- of rhe concept 1les m "f prises some conceptual pil ars t at run
ments over GDP, suggested by the official documents of the EU Lisbon and Europe " ) policies. The concept m act com " " l
t1veness ditional regional innovation policies, name y:
2020 Agenda. On the contrary, thanks to its taxonomy, the approach of the regional counter to tra 1 1
innovation mode calls for ad hoc interventions with the aim of supporting, strength- " d ith R&D so that innovation policies should
ening and diversifying the virtuous aspects of each regional innovation process. It is • innovation is not merely assoCl~te WI~ 'd on R&D investments. This
n o t be focused primarily on hlgh-tec se~tors, al~
on these local innovative profiles that the European strategy of modernization and
" h' ize-hts all' innovatlon po ICy; , " k"
innovation (the so called 'smart innovation srrategy'<') finds a strong and appropriate goes agamst t e one-Slz e - t " tend to encourage a culture of pIC mg
"1 " " trategy do es no m hi
base for its implementation. • smart specia izanon s " her.j h t wards public-private partners ip
winners' on a sectoral basis; rat er, it pus eds1 0 " .
"al discovery' an earnmg,
processes o f 'entrepreneur"l h based on the self-discovery
9.8 Modern innovation policies: the smart specialization strategy • advocated in this regard IS a bottom-up a~?ro~~e old policy style calling for
of entrepreneurial capability, th~Sd sU~fer~edmgtrial development priorities. This
9.8.1 A new policy design and its advantages i" thods to 1 enti y m us f 1 1
centralize d P annI"ng" me l d d-driven because it is derived rom oca
The smart specialization approach was developed with the aim to find an explanation- policy approach IS m genera eman
and a consequent rational strategy - for the large R&D gap between Europe and some potentials and local needs; l that the logic and design of
key trading partners. The most straightforward reason for the knowledge gap was • the bottom-up nature of the POdlicYlsty e etns~:s local regional context, rarher
" iate for an re evant o l b d
outlined in the smaller share of European economy composed of high-tech, R&D inten- the po 1icy IS appropr "1 b dy It is therefore a true p ace- ase
sive sectors. A second reason for the gap was pointed out in the spatial dispersion of than being imposed by a supra-reglOna o .
d b the Barca Report; " "" " 1
the limited R&D efforts, generating insufficient critical mass and investment duplica- policy, as a d vocate Y "1 d al development pnontles IS a so
tions, inefficient resource allocation, and consequent weak learning processes.f • the endeavour to identify real ~ot~~t1a" an "re d at fostering growth which is
d licy pnontlzatlon aune 1" d i h
On the basis of this diagnosis, a rational and concrete proposal was put forward by a way to engen er a po - d which can be explicitly app le m t e
the 'Knowledge for Growth' expert group. It advocated differentiated policies for 'core' realistic for the context concerned, an
and 'periphery' regions, the former able to host laboratories and research activities on regional contexto
general purpose technologies (GPT), the latter oriented towards the identification of
their 'knowledge domain' in which to specialize and towards co-operation with exter-
9.8.2 The risks of the new policy design diti l Id-style
nal R&D providers ('co-application of innovation'}."! "" " hi a roach from the tra rttona o
In more recent formulations, the smart specialization strategy has been translated The above conceptual pillars dlstmgul~h t IS "pp d dern efficient and shareable,
"" Th f s are ínnovanvc an mo , " " "
to a regional setting - with problems, however - and extended to encompass aIl strate- innovation poltCles. ese eature desi d implementatlon of mnovatlOn
gies intended to build regional competitiveness through the design and implementatio n and they constitute a cultural leap" in the efslgn anR&D_based policy that over past
" " ving away rom an h
of innovation; where innovation is understood in a broader sense than mere R&D (as (and competitiveness) po 1icies, mo d i Ifici y At the sarne time, however, t e
d ll i f iliry an me ciencv, " l
it was in the original formulation) and as embracing creative industries, social and years has demonstrate a l"tS ragi 1 nd novelties that are not easy to put m pace,
service innovation, new business models, and practice-based innovation." new policy style "a~so clo~ta~sh~sfo:c~~:ir successful implementation to overcome the
The strategic importance of the concept consists especially in the fact that irnple- and require addltlO na t o g " l
h" pohcy sty e. 1 k
mentation of the smart specialization strategy is interpreted as one of the conditions risks that accompan~ t Ishnet t that local preconditions for innovation may be ac mg.
that a region must fulfil to receive funding from the European Regional Development A first risk is rel~te to t e ac the real ea acity of regions, especially lagging ?nes," to
Fund, the main funding stream within EU Cohesion Policy. Regions are required to The main criticallssue concerns that is, a bottom-up strategy for the identlficatlOn
develop their innovation strategies on the basis of their technological specialization, put in place a self-dis~~ver~:r:~;i~:. Laggi~g regions in general lack the key elements
or - in the words of the smart specialization strategy's experts - on the basis of their of their strengths an pp
necessary for a smart specialization strategy to be effective: they lack connectedness,
entrepreneurial spirit, size in terms of market potential, industrial diversity, quality of
local governance and a critical mas s of capabilities to develop collective learning pro-
cesses.MLKJIHGFEDCBA
A J I of these are elements that a successful entrepreneurial search process requires,
and the smart specialization strategy runs the risking going against objectives of regional
cohesion policy if it is not carefully monitored.
The second risk regards the difficult policy-prioritization. In the case of regions
where the potential for innovation exists thanks to the presence of a variety of sectors,
of entrepreneurial spirit and of enough human capital to spread the advantage of
individual innovative activities around the local area so as to genera te collective learn-
ing, the problem of the entrepreneurial search process lies in the criteria with which
to select (and co-ordinate) among the bottorn-up projects that may be proposed, and
the domains in which to concentra te public resources. At the theoretical level, the
concept of related variety - which refers to the variety of industries in a regio n that
are cognitively related (see Section 9.5) - has been identified as a possible way to
identify opportunities for regions to diversify into new industries. The higher the
degree of related variety, the more learning opportunities are available at the local
level, the more knowledge spillovers across industries occur, and the higher the regional
growth. On this reasoning, policies should support discoveries that can actually build
on, and are embedded in, existing related resources at regionallevel, and this is a basis
for policy-prioritization.f
Another risk is that of misallocation of public resources and unlikely local strategies.
A bottorn-up process of strategy design carries high risks of misallocation of public
resources whenever local interests and local political needs may set unfeasible indus-
trial targets and risky innovation strategies. One suggestion for dealing with this limi-
tation is to reason on a common policy design for regions with similar types of
innovation modes. To this end, a sound taxonomy of innovative regions is required,
a taxonomy which moves away from simple knowledge creation indicators (tradi-
tional patents and R&D indicators), and is able to capture the different innovation
modes that regions are actually developing thanks to the presence of specific local
preconditions for knowledge and innovation creation. In this regard, the regional
'innovation patterns' presented above are in my view a good way to build a useful
taxonomy of innovative regions; they may be found empirically in the way knowledge
and innovation are developed within individual regions according to the nature of
their traditional knowledge base and productive specificities, andJor are captured from
other regions via co-operation, the mobility of scientists and professionals, market
procurement and trans-regional investments. The identification of territorial patterns
of innovation leads to the suggestion of 'smart innovation policies'. These are defined
as policies able to increase the innovation capability of an area and to enhance local
expertise in knowledge production and use by acting on local specificities and on the
characteristics, strengths, and weaknesses of already-established innovation patterns
in each region.
Another risk associated with the smart innovation strategy is that of lock-in with
respect to local historical specialization. The self-discovery process goes against tech-
nocratic approaches c1aiming that they can define priorities, objectives and targets on
the basis of scientific techniques, and that they can identify which knowledge and
inter-industry spillovers should be implemented and supported. However, this process
is necessarily guided by routines and competences at the organizationallevel that make
search behaviour localized (Nelson and Winter, 1982), not only in the industrial sense
(i.e. within a specific technological paradigm), but also in the regional sense. It targets
existing regional specializations and pushes towards possible future diversification
processes. This strategy entails the risk of lock-in.
How can these risks be prevented? Are the proposals mentioned acceptable and
shareable? The debate on these questions is still open and calls for additional insights.
9.9 The concept of territorial capitalihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA
9.9.1 The definition and the taxonom y
This chapter and the previous two have concentrated on the elements considered to
be at the basis of local development. They are of diverse nature: material (infrastruc-
ture, presence of large firms) or non-material (knowledge, creativity, entrepreneurship,
social capital), public (transport and energy infrastructure) or private (financial and
productive capital), generated by endogenous (development of local creativity and
knowledge) or exogenous (multinationals, investments, investments of the public sec-
tor) processes.
The variety of elements considered to generate a local development pattern has
recently induced development of a synthesis concept labelled 'territorial capital', which
is defined as all local, tangible and intangible, endogenous and exagenaus, assets, of
public and private nature, that constitute the deoelopm ent potentials of an area.
The concept of 'territorial capital' was first proposed in a regional policy context
by the OECD in its Territorial Outlook, and it has been recently reiterated by D.G.
Regio of the Cornmission of the European Union:
Each Region has a specific 'territorial capital' that is distinct from that of other
areas and generates a higher return for specific kinds of investments than for oth-
ers, since these are better suited to the area and use its assets and potential more
effectively. Territorial development policies (policies with a territorial approach to
development) should first and foremost help areas to develop their territorial
capital Y
Launched in a scientific context by Roberto Camagni," territorial capital warrants
closer inspection in order to draw up a taxonomy of all potential sources of develop-
mento The proposed taxonomy is built upon two main dimensions, chosen so as to
identify the economic nature of each component of territorial capital and, conse-
quently, the laws of accumulation and depreciation of each component (Figure 9.6):
• rivalry, which makes it possible to identify whether the territorial capital asset
can be used only by an individual (private good) or by a specific group of people
(impure public goods - available to everybody, but there is rivalry in their use
since they are subject to -congestion and scarcity - or club goods, available for
a specific group of people that can make use of them without rivalry in their
use), or available to the whole community (public goods);
• m ateriality, which makes it possible to identify a good according to its physical
or intangible nature: tangible goods, intangible goods, and an intermediate c1ass
of mixed, hard-soft goods are identified.
260 D iversified-relational space xwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA Innovation and proxim ity 261

explained in detail in Chapter 8 - that are sources of agglomeration economies and

High rivalry Prívate fixed capital Rclational prívate Human capital:
stock - entreprcneurship
genera te growth for the entire cornmunity.
services operating on:
(prívate - external linkagcs for - creativity
goods) Pecuniary extemalities firms - privatc know-how
(hard) - transfer of R&D Pecuniary 9.9.2 Law s of accum ulation and depreciation of territorial capital
results extemalitics
TolI goods (excludability) Universily spin-otTs (soft) The taxonomy of the various components of territorial capital is not only a means to
eA i f summarize al! potential assets that can influence local development. The proposed
taxonomy allows identification of the specific economic nature of each component,
and the consequent accumulation and depreciation processes that accompany the life
(club Propríetary networks Co-operalion networks: Relational caoital:
goods)MLKJIHGFEDCBA - strategic alliances in - co-operation cycle of each asset. This is a fundamental aspect for defining the appropriate strategies
~ R&D and knowledge capability for use of these resources, ensuring their protection and their valorization in the long
"" Collective goods: - public/private partnerships - collective action
~ run. Private and public, as well as tangible and intangible, goods are subject to differ-
-landscapc in services and schemes eapability
•• (impure - cultural hcritagc Govemance of land and - collective
ent laws of accumulation and depreciation. The accumulation of material public
~ publie (privale 'ensembles') cultural resources competencies goods, such as infrastructure, but also cultural and natural resources owned by the
goods)
b h e government, largely depends on the quality of their governance: the public sector is
responsible for the crucial functions of control over these goods, the purpose being to
Resources: Agencies for R&D Social capital:
maintain their potential benefits for a long period of time, avoiding their depreciation
- natural transcoding - institutions and their destruction.
(publie
- cultural (punctual) - behavioural The accumulation in a local context of a private material good, like labour or capi-
Receptivily enhancing modcls, values
goods) tal, today depends on its local anchorage. In modern economies, in fact, the hyper-
Social overhead tools - trust, reputation
capital: Connectivity - associationism
mobility of production factors like capital and labour is avoided only if these factors
- infrastructurc a Agglomcration and d are strongly anchored to the local environment through the presence of other less
Low rivalry district economics g mobile factors, like knowledge and social and relational capital." The latter factors,
of an intangible nature, are embedded in the local society; impossible to transfer else-
Tangible goods Mixed goods Intangible goods
where, they become sunk costs.
(hard) (hard + soft) (soft)
As regards intangible private goods, like knowledge, these accumulate at the local
Materiality level through education processes, research investments, co-operation among firms,
individual and collective learning. In an indirect way, knowledge can develop through
Figure 9.6 A theoretical taxonomy of the components of territorial capital spillover effects from the research centre where it is formed because of the mobility of
Source: Camagni (2009) researchers and skilled technicians, as well as imitation and co-operation processes.
The accumulation over time of knowledge at locallevel depends closely on the conti-
nuity and persistence of the actor s that participate in its generation. Today, there are
high risks of knowledge 'de-cumulation' due to the local firms' outsourcing of phases
Thanks to these two dimensions, it is possible to go beyond the traditional classifica- of the production processes. In the short term, this can disrupt integration and synergy
tion of potential productive resources based on social capital, human capital, infra- processes within firms and within the local area, and impoverish the flow of techno-
structure and productive capital (fixed private capital), which are located at the four logical creativity and knowledge production.
corners of the matrix. In fact, the matrix shows an intermediate class, embracing club
and impure public goods, which characterizes the group of innovative elements ste~-
ming from the most recent theories. On the one side, we find networks, of mater.,al 9.9.3 Endow m ent and efficiency of territorial capital
nature in the case of ICTs, but also of non-material nature in the case of co-operatlv~ The vast and composite mosaic of local success stories and of dynamic evolutionary
networks and strategic alliances among firms for the production of new products an trajectories can be explained by diverse elements of territorial capital, endogenous and
services. On the other, we find public goods that are subject to congestion and to exogenous, qualitative and quantitative, traditionally functional (based on the presence
depreciation in the presence of free-rider behaviour. The peculiarity of this catego~ of production factors and of preconditions for local efficiency) or relational in nature.
of goods lies in the fact that it requires new forms of governance, inclusive :nd b'. Regions do not require the presence of all the above components of territorial capi-
partnership, that guarantee the maximum advantage for the members of the cl~ l- tal to develop and maintain over time a positive and dynamic development trajectory.
Moreover, a category of mixed goods emerges. This is characterized by both matena- The local endowment of specific assets of territorial capital results from the history of
ity and non-materiality, and its presence makes it possible to underline the importa nce the local area, and determines its productive specializations on which a strategic
of those complex territorial organizations like cities and industrial districts - as growth pattern is to be built.
L .ó L . U tu e rs m e d -re ta tio n a l ihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA
space xwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA Innovation and proxim ity 263

Most of the empirical analyses measuring the various components of territorial
capital at regionallevel show that there is a decisive difference between the endowment
of territorial capital elements and development levels; for example, the North of Italy
has a much higher endowment of all territorial capital assets than the Centre and the
South of Italy, this last possessing the lowest endowment of all territorial capital
assets." However, when the analysis is developed at a more disaggregated territorial
level, a different picture emerges. An analysis conducted at provinciallevel (NUTS3)
depicts the following situation:
• metropolitan provinces, which are endowed with above-average levels of territo-
rial capital components;
• provinces mainly endowed with intangible e\ements like social and relational
capital;
• provinces mainly endowed with non-material e\ements like transport, energy
and educational infrastructure;
• finally, provinces that lack an endowment of all territorial capital e\ements.
Provinces that belong to one of these clusters do not have a specific geographic loca-
tion; the soft elements of territorial capital, like social capital, are mostly present in
the northwestern part of ltaly, while the northeastern part has none. Central Italy,
generally presented as a uniform socio-economic area, has provinces that be\ong to all
four clusters. The provinces of the South of Italy, with the exception of the metropoli-
tan provinces, mostly be\ong to the fourth cluster.
Even more interesting is that the same analysis shows that the simple endowment
of territorial capital assets is unable to explain the rates of growth of those areas;
whilst the northern part of Italy has the highest endowment of territorial capital assets,
it does not record the highest growth rates. This testifies that what makes the differ-
ence in terms of growth is the efficiency with which these assets are used. The greatest
efficiency in the exploitation of territorial capital resides in the integration of tangible
and intangible elements, which reinforce each other."
As in the case of all economic resources, the efficiency of territorial capital assets
depends not on the endowment of single assets but on the presence of complementary
and synergic components, and on their balanced deve\opment: an idea that recalls the
balance of deve\opment theory (Chapter 4). The novelty is that today the interaction
takes place among non-material resources. Econometric analyses show that the mere
existence of knowledge does not explain regional growth trajectories; on the contrary,
it plays an important role in those European regions with high endowments of social
and re\ational capital.P
The synthesis of an area's success factors reminds us that identification of such
factors takes place within conceptual approaches that are extremely different from
each other. The traditional functional approach - also termed a positivist and cogni-
skilled labour; more to relational capital than to accessibility; more to local identity
than to the presence of important e\ements like quality of life and efficiency of the
economic systern.P
Such a rich concept is of great normative value, especially in a period when regional
policies are expected to be conceptualized on the basis of differentiated strategies spe-
cific to the local contexto As the 'Barca Report' of the European Union suggests, regional
policy must be a place-based policy built on the basis of the specificities and e\ements
of competitiveness of each single area through participatory and inclusive processes.i"
A conception of territorial capital that embraces and systematizes all the e\ements on
which competitiveness can rely, and that highlights the laws of accumulation and depre-
ciation, is therefore crucial for the appropriate design of these policies.BA
9 .1 0 Conc1usions
The chapter has surveyed the theories based on the idea that the role of space in eco-
nomic activity does not consist sole\y in improvements to the sta tic efficiency of pro-
duction processes (that is, an increase in firms' revenues or a decrease in their costs).
The advantages of a concentrated location of activities in space are also manifest in
the innovative and creative capacity of firms, and space becomes a so urce of dynam ic
ef(iciency. In these theories, the innovation capacity of local systems is made dependent
on local socio-economic conditions deeply embedded in the local area. Different con-
cepts of proximity (physical, re\ational, institutional, cognitive) have been brought to
the fore as elements explanatory of the innovation capacities of local firms, within a
rigorous microeconomic and micro-behavioural framework.
The next chapters will consider theories developed in more recent times which
include spatial aspects and the increasing returns that derive frorn them in macroeco-
nomic growth mode\s, but are only able to do so beca use they adopt a different con-
ception of space: that of 'diversified-stylized' space.
Appendix: indicators of 'related variety'
In formal terms, the indicator of 'related variety' is built as an entropy indicator.P
Applied to the concept of 're\ated variety', maximum entropy shows a situation of
equal distribution in a region of knowledge among the different technological classes
within a larger technological class to which they be\ong. More precise\y, the indicator
of 'related variety' is the sum of entropy (H ) internal to a broad technological class
(e.g. a two-digit class) (g ), weighted for the ~hare of patents in each class (Sg) on the
total number of patents present in a region from 1 to G:
G
RelatedVariety = ¿SgH g
g~l

tive approach - interprets the reality on the basis of deterministic, mechanical,

cause-effect relationships. Another approach has recently been deve\oped. It sug- Entropy within a digit class g (H g) is calculated as:
gests inter-subjective relationships more complex than the deterministic ones and
based on the ways in which economic actors interpret the reality, react to external
stimuli, and are capable of synergic and co-operative behaviours. This new approach 1
underlines that local competitiveness is linked more to trust and a sense of be\onging
than to a simple endowment of capital; more to creativity than the pure presence of
j{
where s represents the share of patents in one technological class (e.g. five-digit) on 9 What is the definition of a regional innovation pattern? What are the novelties
the total of patents contained in a larger technological class (e.g. two-digit). When the of this approach?
share of patents of a larger technological class (S ) is equal to the more detailed tech- 10 What is the smart specialization strategy about? How would you distinguish
nological classA (s), entropy oí class g is rninimum; knowledge is entirely confined the smart specialization strategy from previous innovation policies? What are
within one technological class belonging to class g and knowledge complementarity is the risks that still exist in relation to the implementation of this strategy?
null within the technological class g. If this happened in all classes g, the indicator of 11 How would you define a territorial capital? What are the dimensions on which
'related variety' would also be equal to zero, and it would signal the absence oí related a taxonomy of territorial capital elements have been produced and why? What
variety. By contrast, when the entropy indicator increases, the level oí 'related variety' is the usefulness of the territorial capital taxonomy?
increases as well.
More recently, a related interregional variety has been suggested. This index mea-
sures the degree of knowledge complementarity within a common knowledge base of Selected reading on empirical findings
two regions; and it has been used to determine the role of interregional cognitive
About innovation and regional developm ent
proximity in explaining scientific co-operations.f It is calculated as follows:ihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA
Aydalot P. and Keeble D. (eds) (1988), H igh Technology Industry and Innovative Environm ent,
Routledge, London.
c(S "·S )(1 )
L
1 g" ~ g"1
lnterregional Related Variety MLKJIHGFEDCBA
= L
(Is;" ~ s;"I) Breschi S. (2000), 'The Geography of Innovation: A Cross-sector Analysis', Regional Studies,
g=l S 6,-1
" Sg ,-2 ,= 1 vol. 34, no. 2, pp. 213-229.
Ciciotti E. and Wettmann R. (1981), 'The Mobilisation of Indigenous Potential', Commission
where r and r represent a pair of regions, S the share of patents in the largest tech- of the European Community, Internal D ocum entation on Regional Policy, no. lO.
nologic~l class~s g (present in a number from 1 to G), s the share of patents in classes Cooke P. and Morgan K. (1994), 'The Creative Milieu: A Regional Perspective on Innovation',
L belonging to class g, in a number from 1 to l. The index mea sures the technological in M. Dodgson and R. Rothwell (eds), The H andbook of Industrial Innovation, Edward
Elgar, Cheltenham, pp. 25-32.
complementarity between regions through the differences in the patent shares in the
Decoster E. and Tabariés M. (1986), 'L'innovation dans un Póle Scientifique et Technologique:
technological classes i: in particular, the larger the share of patents in classes i between
Le Cas de la Cité Scientifique Ile-de-France', in Ph. Aydalot (eds.), M ilieux Innovateurs en
two regions, the higher the technological complementarity between regions. Moreover,
Europe, GREMJ, Paris, pp. 79-100.
the index measures the common knowledge base of the two regions through the prod- Frenkel A. (2001), 'Why High-Technology Firms Choose to Locate in or near Metropolitan
uct of the patent shares g of the two regions. When the shares of the patents within Areas', U rban Studies, vol. 38, no. 7, pp. 11083-11101-
classes g of the two regions, having controlled for their relative size (measured as the Goddard J. and Thwaites A. (1986), 'New Technology and Regional Development Policy', in
difference in the share at the denorninator-"}, are high, regions show a large cornmon P. Nijkamp (ed.}, Technological C hange, Em ploym ent and Spatial D ynam ics, Springer Verlag,
technological base. Berlin, pp. 91-114.
Maggioni M. (2002), C lustering D ynam ics and the Location of H igh- Tech Firm s, Physica-
Verlag, Berlin.
Review questions Malecki E. and Varaiya P. (1986), 'Innovation and Changes in Regional Structure', in P. Nijkamp
(ed.), H andbook of Regional and U rban Econom ics, North-Holland, Amsterdam,
1 What is conceptualized by the knowledge spillover approach? What are the
pp. 629-645.
limits of this approach?
Oakey R., Thwaites A. and Nash P. (1980), 'The Regional Distribution of Innovative Manufac-
2 How would you define a 'm ilieu innovateur'? What are the genetic elements of
turing Establishments in Britain', Regional Studies, vol. 14, no. 3, pp. 235-253.
a milieu? Does the theory of the m ilieu innovateur overcome some of the lirnits Paci R. and Usai S. (2000), 'Technological Enclaves and Industrial Districts: An Analysis of the
of the knowledge spillover approach, and how? Regional Distribution of Innovative Activity in Europe', Regional Studies, vol. 34, no. 2,
3 What is conceptualized in the learning regions theory? How would you define pp. 97-114.
a learning region? What are the main strengths and weaknesses of the theory? Schatzl E.L. and Revilla Diez J. (eds) (2002), Technological C bange and Regional D euelopm ent
4 What is the role of space in the knowledge spillover approach, in the m ilieu in Europe, Springer Verlag, Berlin.
innovateur theory and in the learning region approach? Sternberg R. (1996), 'Reasons for the Genesis of High-tech Regions - Theoretical Explanation
5 What is meant by regional innovation systems? and Empirical Evidence', G eoforum , vol. 27, no. 2, pp. 205-223.
6 What does the evolutionary economic geography approach theorize? According
to this theory, what are the determinants of industrial specialization of an area? About m ilieux innovateurs
What makes the exchange of inforrnation easier according to this approach?
7 What is meant by cognitive, relational and institutional proximities? Are rhey Arnin A. and Thwaites A. (1986), 'Change in the Local Economy: The Case of the Northern
Region (UK)', in Ph. Aydalot (ed.), M ilieux Innovateurs en Europe, GREMI, Paris,
synonyms? How do they relate to the geographical proximity concept?
pp. 129-162.
8 Which role do es space play in the m ilieu innovateur theory, in the learning
Camagni R. and Capello R. (1998), 'Innovation and Performance in SMEs in Italy: The Rele-
regíon approach and in the evolutionary economic geography approach?
vance of Spatial Aspects', C om petition and C hange, vol. 3, pp. 69-106.
 Innovation and proxim ity 267

Camagni R. and Rabellotti R. (1997), 'Footwear Production Systems in Italy: A Dynamic Com-
parative Analysis', in R. Ratti, A. Bramanti and R. Gordon (eds),ihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA
The D ynam ics of Innova-
tive Regions, Ashgate, Aldershot, pp. 139-164.
Capello R. (1999), 'SMEs Clustering and Factor Productivity: A Milieu Production Function
Model', European Planning Studies, vol. 7, no. 6, pp. 719-735.
Capello R. (2001), 'Urban Innovation and Collective Learning: Theory and Evidence from Five
Metropolitan Cities in Europe', in M.M. Fischer and J. Froehlich (eds), K noioledge, C om :
plexity and lnnouation System s, Springer, Berlin, pp. 181-208.
Gordon R. (1993), 'Structural Change, Strategic Alliances and the Spatial Reorganisation of
Silicon Valley's Semiconductor Industry', in D. Maillat, M. Quévit and L. Senn (eds), Réseaux
d'innouation et M ilieux Locaux, GREMI, Neuchátel, EDES, pp. 51-72.
Lambooy J. (1986), 'Regional Development Trajectories and Small Enterprises: the Case Study
of the Amsterdam Region', in Ph. Aydalot (ed.), M ilieux lnnouateurs en Europe, GREMI,
Paris, pp. 57-78.
Tabaries M. (1992), 'Nouvelles PME et cité scientifique en Formation: Ile-de-France Sud', in D.
Maillat and J.-e. Perrin (eds), Entreprises lnnouatrices et D éueloppem ent Territorial,
GREMI, Neuchátel, EDES, pp. 23---40.
Boschma R., Minondo A. and Navarro M. (2012), 'Related Variety and Regional Growth in
Spain', Papers in Regional Science, vol. 91, no. 2, pp. 241-256.
Frenken A K ., van Oort EG. and Verburg T. (2007), 'Related Variety, Unrelated Variety and
Regional Economic Growrh', Regional Studies, vol. 41, no. 5, pp. 685-697.
Neffke E, Henning M. and Boschma R. (2011), 'How Do Regions Diversify Over Time? Indus-
try Relatedness and the Development of New Growth Paths in Regions', Econom ic G eogra-
phy, vol. 87, no. 3, pp. 237-265.
Todtling E and Trippl M. (2005), 'One Size Fits All? Towards a Differentiated Regional Innova-
tion Policy Approach', Research Policy, vol. 34, no. 8, pp. 1203-1219.
About regional patterns of innovation
Camagni R. and Capello R. (2013), 'Regional Innovation Patterns and the EU Regional Policy
Reform: Towards Smart Innovation Policies', G row th and C bange, vol. 44, no. 2, pp. 355-389.
Capello R. and Lenzi e. (eds) (2013), Territorial Patterns of lnnouation: An lnquiry on the
K noioledge Econom y in European Regions, Routledge, London.

About the sm art specialization strategy

About know ledge spillovers and regional developm ent
Boschrna R. (2014), 'Constructing Regional Advantage and Smart Specialisation', Scienze
Acs Z., Audretsch MLKJIHGFEDCBA
O , and Feldman M. (1994), 'R&D Spillovers and Recipient Firm Sizc', Regionalt - Italian [ournal of Regional Science, vol. 13, no. 1, pp. 51-78.
Review of Econm ics and Statistics, vol. 76, no. 2, pp. 336-340. Foray D. (2009), 'Understanding Smart Specialisation', in D. Pontikakis, D. Kyriakou and R.
Audretsch D. and Feldman M. (1996), 'R&D Spillovers and the Geography of Innovation and
van Bavel (eds), The Q uestion of R& D Specialisation, JRC, European Comrnission, Direc-
Production', Am erican Econom ic Review , vol. 86, no. 3, pp. 630-640. toral General for Research, Brussels, pp. 19-28.
de Groot H., Nijkamp P. and Acs Z. (2001), 'Knowledge Spill-overs, Innovation and Regional
Foray D., David P. and Hall B. (2009), 'Smart Specialisation: The Concept', in K notoledge for
Development', Papers in Regional Science, vol. 80, no. 3, special issue. G row th: Prospects [or Science, Technology and lnnouation, Report, EUR 24047, European
Maier G. and Sedlacek S. (eds) (2005), Spillouers and lnnouations - Space, Enuironm ent and Union.
the Econom y, Springer Verlag, Vienna. Foray D., David P. and Hall B. (2011), 'Smarr Specialisation: From Acadernic Idea to Political
Ruslan L. and Plasmans J. (2003), 'Mcasuring Knowledge Spillovers in the New Economy
lnstrurnent, the Surprising Career of a Concept and the Difficulties Involved in its Implernen-
Firms in Belgium Using Patent Citations', G lobal Business and Econom ics Review , vol. 5, tation', MTEI working paper, Novernber, Lausanne.
no. 1, pp. 75-99. Giannitsis T. (2009), 'Technology and Specialization: Strategies, Options, Risks', K notoledge
Econom ists Policy Brie], no. 8, available online, http://ec.europa.euJinvest-in-researchlpdfl
download_enlkfg.....policy_brieCn08.pdf?11111.
About regional innovation system s
McCann P. and Ortega-Argilés R. (2014), "The Role of the Smart Specialisation Agenda in a
Bacaria]., Borras Alomar S. and Fernández-Riba A. (2004), 'The Changing Institution Struc- Reformed EU Cohesion Policy', Scienze Regionali, ltalian [o u rn a l of Regional Saence, vol.
ture and Performance of the Catalan Innovarion System', in P. Cooke, M. Heidenreich and 13, no. 1, pp. 15-32.
H.-J. Braczyk (eds), Regional lnnouation System s, Routledge, London, pp. 63-90. Neffke F., Henning M. and Boschma R. (2011), 'How Do Regions Diversify Over Time? Indus-
Boekhold P. and de jager D. (2004), 'South-east Brabant: A Regional Innovarion System in try Relatedness and the Development of New Growth Paths in Regions', Econom ic G eogra-
Transition', in P. Cooke, M. Heidenreich and H.-J. Braczyk (eds), Regional Innouation Sys- phy; vol. 87, no. 3, pp. 237-265.
tem s, Routledge, London, pp. 44-62. Pontikakis D., Chorafakis G. and Kyriakou D. (2009), 'R&D Specialization in Europe: From
Cooke P. (2004), 'The Regionallnnovation System in Wales: Evolution or Eclipse?', in P. Cooke, Stylized Observations to Evidence-based Policy', in D. Pontikakis, D. Kyriakou and R. van
M. Heidenreich and H.-J. Braczyk (eds), Regional lnnouation System s, Routledge, London, Bavel (eds), The Q uestion of R& D Specialisation, jRC, European Commission, Directoral
pp. 214-233. General for Research, Brussels, pp. 71-84.
Maskell P. (2004), 'Learning the Village Economy in Denmark: The Role of Institutions and
Policy in Sustaining Competitiveness', in P. Cooke, M. Heidenreich and H.-J. Braczyk (eds),
Regional Innouation System s, Routledge, London, pp. 154-185. About territorial capital

Camagni R. (2009), 'Territorial Capital and Regional Development', in R. Capello and P.

Nijkamp (eds), H andbook of Regional G row th and D euelopm ent Theories, Edward Elgar,
About evolutionary econom ic geograpby
Cheltenham, pp. 118-132.
Boschma R. and Iammarino S. (2009), 'Related Variery, Trade Linkages, and Regional Growth Perucca G. (2013), 'A Redefinition of Italian Macro-areas: The Role of Territorial Capital',
in Italy', Econom ic G eography, vol. 85, no. 3, pp. 289-311. Rivista di Econom ia e Statistica del Territorio, vol. 2, no. 2, pp. 35-63.

Perucca G. (2014), 'The Role of Territorial Capital in Local Economic Growth: Evidence from
Italy',ihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA
European Planning Studies, vol. 22, no. 3, pp. 537-562.
Tóth B.I. (2014), 'Territorial Capital: Theory, Empirics and Critical Remarks', European Plan-
ning Studies, MLKJIHGFEDCBA
D O ! : 10.1080109654313.2014.928675.
Further reading
Camagni R. (1991), 'Local M ilieu, Uncertainty and Innovation Networks: Towards a New
Dynamic Theory of Economic Space', in R. Camagni (ed.), Innovation N etw orks: Spatial
Perspectives, Belhaven-Pinter, London, pp. 121-144.
Capello R. (1999),. 'Spatial Transfer of Knowledge in High-technology Milieux: Learning vs.
Collective Learning Processes', Regional Studies, vol. 33, no. 4, pp. 353-365.
Foray D., David P. and Hall B. (2009), 'Smart Specialisation: The Concept', in K now ledge [or
G row th: Prospects for Saence, Technology and Innovation, Report, EUR 24047, European
Union.
Frenken K., van Oort EG. and Verburg T. (2007), 'Related Variety, Unrelated Variety and
Regional Economic Growth', Regional Studies, vol. 41, no. 5, pp. 685-697.
Keeble D. and Wilkinson E (2000), H igh-technology C lusters, N etw orking and C ollective
Learning in Europe, Ashgate, Aldershot.
Lundvall B.A. (ed.) (1992), N ational System s of Innovation: Tow ards a Theory of Innovation
and Interactive Learning, Pinter, London.
Neffke E, Henning M. and Boschma R. (2011), 'How Do Regions Diversify Over Time? Indus-
try Relatedness and the Development of New Growth Paths in Regions', Econom ic G eogra-
phy, vol. 87, no. 3, pp. 237-265.
Torre A. and Wallet E (eds) (2014), Regional D evelopm ent and Proxim ity Relations, Edward
Elgar, Cheltenham.
Notes
1 For a critical survey of neo-Schumpeterian theories see Mouleart and Sekia, 2003.
2 The concept of proximity gave rise to a school of thought in France that analysed the rela-
tively greater importance for local development of organizational and cultural proximity
compared with geographical proximity, Caragliu, 2015. The French school's notion of
'organizational and cultural proximity' and the one proposed here of 'relational proximity'
have many features in common. On the French school of the 'economics of proximity' see
Bellet et al., 1993; Rallet and Torre, 1998 and 2005; Rallet, 2002; Torre and Wallet, 2014.
3 For the French school of proximity, see among others, Bellet et al., 1993; Rallet, 2002, Rallet
and Torre, 2005; Torre and Wallet, 2014. For the theory of m ilieu innovateur see Camagni,
1991. ' and to find an economic rationale for the manner in which they support innovative pro-
4 In the next sections of the chapter we shall not take into consideration organized, techno-
logical and social proximities. As regards social proximity, the concept was formulated in
the industrial district theory devoted to explanation of sta tic rather than dynamic efficiency,
and i t has be en presented in Chapter 8. As regards organized proximity, this is a very general
concept embracing different concepts of proximities at the same time, and therefore not
really useful for highlighting the effects of each type of proximity. Technological proximity
IS interpreted by the concept of cognitive proximity, and is therefore already taken into
consideration by that concepto
5 Wide-ranging empirical studies on innovative activity ha ve be en carried out in the UK by
the CUROS (Centre for Urban and Regional Development Studies) oí the University of
Ne~castle (see Oakey et al., 1980; G?ddard and Thwaites, 1986) and by the SPRU (Scienc~
Policy Research Unir) of the University oí Sussex (see Clark, 1971); in the USA by Malecki
(see Mal.ecki and Varaiya, 1986); and more recently also in Italy (see Breschi, 2000; Pacl
and Usai, 2000). Studies on the concentrated location of high-tech firms have been con-
ducted by Keeble on the UK, Sternberg on Germany, Ciciotti on Italy, Decoster and Tabaries
lnnovation and proxim ity 269
on Fra.nce, Malecki on the USA, Frenkel on Israel and Maggioni on a group of OECD
countnes. See Ciciott~, 1982; Decoster an~ Tabaries, 1986; Keeble, 1990; Sternberg, 1996;
Frenkel, 2001; Maggioru, 2002. For detailed studies on the role oí innovation in regional
development see, e.g., Cappellin and Nijkamp, 1990; de Groot et al., 2004; Ewers and
Allesch, 1990. For a theoretical and ernpirical analysis of spatial spillovers see Maier and
Sedlacek, 2005.
6 Among the numerous empirical studies on knowledge spillovers, to be mentioned in par-
ticular are the semlllal. work by]affe, 1989, which was íollowed by other studies: Acs
et al., 1994, who exarnined the differing abilities of small and large firms to exploit knowl-
edge spillovers; Audretsch ~nd Feldman,. 1996 and Feldman and Audretsch, 1999, who
distinguished between sCI~ntIfically diversified and specialized spillovers; and Anselin et al.,
2000, who defined the distance beyond which spillover effects disappear. For a critical
review of the rol e of knowledge spillovers in regional development see de Groot at al
2001. .,
7 Given expenditure on research and development as the input to innova ti ve activity (R&D)
and t~e nu~ber of I?atentsA(B ) as the output, the knowledge production function shows the
quantity of innovanve input required to obtain a certain amount oí innovative output:
B= f(R& D ) (8.4n)
Recent econometric tests ha ve shown the existence oí a certain simultaneity between R&D
andpatents. This evidences that applications for patents tend to be made very early on in
t!"temnovatrve process and are consequently les s indicative oí a capacity to produce innova-
trve output.
8 More refined methods, such as the inclusion oí 'lagged' variables taking different values
according to the geographical distance between the areas analysed, have recently been used.
The results oí the analysis do ~ot change: they confirm the existence oí technological spill-
overs through the greater significance oí university expenditure on R&D for the innovative
capacity oí areas geographically closer to where the university is located, finding that 50
miles IS the distance beyond which spillover effects disappear.
9 See Section 7.4.
1O T~e reference is to studies conducted by GREMI, Groupe de Recherche Européen sur les
Milieux Innovateurs, headquartered in Paris, whose members were scholars from all the
European countries. The group's research results were set out in a series oí publications,
most notably Aydalot, 1986; Aydalot and Keeble, 1988; Camagni, 1991; Maillat et al.,
1993; Ratti et al., 1997; RERU, 1999.
11 It should be stressed that 'the concept oí "innovative milieu" is necessarily abstract; the
mIlI~u must be considered an economic and territorial archetype more than an empirical
reahty. Its conceptualization in economic terms enables us to generalise some recent empiri-
cal findings showing the importance oí relational assets in the success oí some specific areas,
cesses. The characteristics oí the innovative milieu are never fully realised in real territorial
systems, however. The relationship between the presence oí these characteristics and the
mnovativc outcome has been verified in some empírical cases, and it is above all theoretically
iustified, But it can never be considered a precondition, either necessary or sufficient, for
mnovation; it is only an element which increases the probability oí an innovative outcome.'
See Camagni and Capello, 2002, p. 17.
12 See Bellet et al., 1993; Storper, 1995.
13 See Camagni and Capello, 2002, p. 18.
14 The concept oí relational capital is similar to that of social capital developed by Putnam.
See Putnam, 1993. Ithas been argued that the main difference between the two concepts is
that social capital exists wherever a local society exists, while relational capital consists in
the (rare) ability of actors to inter-relate their different skills interact with each other trust
each other, and co-operate even at a distance with other complernentary organizations. See
Camagni, 2001.
15 Since the concept of collective learning was first formulated by the GREMI group (see
Camagni, 1991), it has been used by numerous other authors. See Capello, 1999a; Keeble
270 D iversified-relational space xwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA
and Wilkinson, 1999 and 2000; Lawson and Lorenz, 1999. On the concept of cognitive
dimension of agglomeration economies, see Cappellin, 2003.
16 See Camagni, 1991.
17 This recalls the theory of the Marshallian industrial district and the rol e performed by social
and cultural homogeneity in producing forms of transaction regulation which deter oppor.
tunistic behaviour. See Camagni and Rabellotti, 1997; Arrighetti et al., 200l.
18 For details on the methodology used see Capello, 1999b. After this pioneering study, sub-
sequent analyses have also examined the effect of collective learning and the local atmo-
sphere on the innovative activity of firms. A collection of studies is published in Camagni
and Capello, 2002.
19 Romer's and Lucas's theories are set out in Chapter 11. For 'stylization' of the m ilieu inno-
vateur theory within a neoclassical endogenous growth framework see Capello, 2002b.
20 For the main studies produced by the Danish school see Lundvall, 1992; Lundvall and
johnson, 1994; Asheim, 1996; Edquist, 1997; Maskell and Malmberg, 1999; Malberg and
Maskell, 2002. For studies produced in Britain and North America see Cooke and Morgan,
1994; Morgan, 1997; Boekema et al., 2000; Cooke, 2002. The importance of institutional
factors for local growth is now so widely recognized that creating institutional performance
indicators for inclusion in macroeconometric growth models is considered to be essential.
See Stimson et al., 2005.
21 See Amin and Thrift, 1994. The term 'institution' should of course be understood in the
sense given to it by North's and Williamson's institutional economics, namely as a set of
societal norms and 'rules of the game' (North, 1990, p. 3). See on this also Wllliamson,
2002.
22 'Institutions are here defined as the sets of habits, routines, norms and laws that regula te
the relations between people and thus shape human interaction and learning': Lundvall and
johnson, 1994,p. 33.
23 This definition of 'learning region' has obvious link s with the theories already described in
Chapter 8. The concept of 'organized market' recalls that of 'community market' developed-
much more convincingly and in richer form - by industrial district theory: social rules and
norms regula te the market, making it more efficient and dynamic. The difference between
the 'community' market and the 'organized' market resides in their outcomes: the former
generates the factors that determine the co-existence and positive interaction between forms
of co-operation and competition; the latter genera te s a dynamic process of interactive learn-
ing. Moreover, the theory of learning regions resembles that of the m ilieu innovateur when
it emphasizes the importance of the 'destructive' learning that enables a regio n to abandon
an obsolete technological trajectory. The m ilieu innovateur theory, too, stresses the impor-
tance for the local system's dynamic of avoiding 'lock-in' to knowledge that may become,
like rules and norms of behaviour, 'barriers to exit' if the m ilieu must rapidly shift to a new
technological trajectory. See Bianchi and Miller, 1993.
24 Lundvall and Johnson point out that 'Iearning economy refers not only to the importance
of the scientific and technology system - universities, research organisations, in-house R&D
departments and so on - but also to the learning implications of the economic structure, rhe
organisational forms and the institutional set-up': Lundvall and johnson, 1994, p. 26.
25 See among others, Edquist, 1997, 2005; Cooke et al., 2000; Cooke et al., 2004; Asheim and
Gertler, 2005; Trippl, 2010.
26 See Todtling and Trippl, 2005.
27 See Section 9.7.
28 On the debate concerning whether the concept of 'national innovation system' can be used
to derive a 'regional' version in the form of a 'regional innovation system' see Howells,
1999; Acs et al., 2000; Fritsch, 2001.
29 In this perspective, the meaning of 'localized innovation' becomes clear: it is an innovatio n
that takes place along specific technological trajectories within a specific paradigm. For an
in-depth analysis of evolutionary theory, see Nelson and Winter, 1977; Dosi, 1982; Antonelh,
1989; Foray and Lundvall, 1996.
30 Supporters of this theory higWight in fact that in the first stages of formation of a conccn-
trated area, space is neutral in the locational decision-making process, since it does nor
influence location choices. In an evolutionary perspective, once firms have by chance located
Innovation and proxim ity 271
in one area, the latter becomes an attracting factor for new firms of the same industry, and
space moves from being a 'neutral space' to a 'real place'. See Boschma and Frenkel, 2006,
p.290.
31 The first author to deal with the concept of 'related variety' was Nooteboom (2000), but it
is thanks ro the Dutch school of evolutionary economic geography that the concept has
developed and received empirical evidence. For advanced studies on this concept, see Bos-
chma, 2005; Frenkel et al., 2007; Boschma and Iammarino, 2009; Boschma et al., 2012.
32 See Frenkel et al., 2007. For the way in which 'related variety' indicators are built, see the
Appendix in this chapter.
33 On the concept of 'absortive capacity', see Cohen and Levinthal, 1990.
34 For a detailed analyses of the negative aspects of proximity, see Boschma, 2005.
35 For the recent hermeneutic approaches to local creativity, see Cusinato and Philippopoulos-
Mihalopoulos, 2015.
36 On the concept of regional patterns of innovation, see Camagni and Capello, 2013; Capello
and Lenzi, 2013.
37 For the knowledge filter theory, see Acs et al., 2004.
38 This idea has opened the way to what has been la ter on called the 'smart specialization
strategy' of the European Union. The first version of this strategy suggested innovation poli-
cies differentiated between centre and periphery. Today, the same policy is more elaborated
in terms both of strategy design, and of geographical areas on which to implement different
strategies. See on this issue Foray, 2009; Foray et al., 2009; McCann and Ortéga-Argiles,
2014. See Section 9.8.
39 For the absorptive capacity concept, see Cohen and Levinthal, 1990.
40 Based on indicators covering all aspects of territorial innovation patterns (from local pre-
conditions, to intensity of knowledge and innovation creation) for all European regions, a
cluster analysis has identified six (and not three) territorial patterns of innovation, witness-
ing the complexity of the real world. The theoretical endogenous innovation model identifies
two ernpirical patterns, one associated with regions generating base knowledge, and another
with regions generating applied knowledge. The creative application pattern distinguishes
in the real world two groups of regions, applying external knowledge to internal formal
knowledge, the first one, and to informal knowledge, the second one. The imitative pattern
also shows two empirical groups of regions: those that imitate and those where innovation
does not take place. See Capello and Lenzi, 2013, chapter 7.
41 On the 'smart specialization strategy', see, among others, Foray, 2009; Foray et al., 2009;
McCann and Ortéga-Argiles, 2014.
42 See Pontikakis et al., 2009.
43 See Foray, 2009; Foray et al., 2009; Giannitsis, 2009.
44 See Foray et al., 2011
45 See Camagni and Capello, 2013; McCann and Ortega-Argilés, 2014.
46 See Frenken et al., 2007; Boschma 2011 and 2014; Neffke et al., 2011; Iacobucci, 2014.
47 See OECD, 2001 and European Commission, 2005, p. 1.
48 See Camagni, 2009.
49 To be clarified is the distinction between social and relational capital. While one can easily
argue that social capital exists when a society exists, relational capital may be kept separare
and be defined as that part of social capital which refers to the co-operative action of each
individual, measured through the bilateral and multilateral agreements that are developed
among actors, both within and outside the local area, facilitated by a friendly and trusting
atmosphere based on shared behavioural rules and values. See Camagni, 2009.
50 For empirical analysis of territorial capital, see Perucca, 2013 and 2014; on Hungarian
regions, see Tóth, 2014.
51 For a detailed analysis of the data and methodology adopted, and the results obtained, see
Perucca, 2014.
52 For data, methodology and results, see Capello et al., 2011.
53 See Camagni, 2009.
54 See Barca, 2009.
55 The entropy principie makes it possible, under imperfect information conditions, to deter-
mine the most probable condition of a system formed of a large number of elements
272 D iversified-relational space xwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA

(molecules) corresponding to its equilibrium condition (maximum entropy). The entropy

principIe, in fact, describes the process of a system (e.g. a gaseous system), oriented only in
one direction, exposed to an external solicitation (e.g. a variation in temperature) towards
PartBAI V
an order system. This is the case of Prigogine's example of an iron bar exposed to a source
of heat at one of its extremes, in which temperature distributes in an ordered way. When
the external so urce of heat ceases, temperature distributes uniformly along the bar, and the T h e o r ie s o f r e g io n a l g r o w th
system moves from a molecular order to a disorder situation, and therefore from a low to
a high probability condition. See Camagni, 1992a, p. 105.
56 See Capello and Caragliu, 2012. D iv e r s ifie d -s ty liz e d sp ace
57 Thanks to the denominator, the greater the difference in the class between the two regions,
the lower the cognitive proximity. The denorninator, in fact, makes it possible to adjust high
levels of the numerator due to extreme cases in which one of the two regions concentrates
its patent activity in only one technological class.