1

WORKSHOP DOCTORAL 2014

DETERMINANT FACTORS OF CREATIVE CLUSTERS IN SPAIN, PORTUGAL AND ITALY

PhD candidate: Daniel Sánchez Serra

Supervisor: Rafael Boix Domènech

Abstract. This paper examines the factors determining the spatial clustering of creative industries in

Europe by using plant-level microdata. The paper proposes a model tailored to differentiate the effect of

general-economic and specific-creative forces on the localization of creative industries. The model is

experimentally applied to Spain, Portugal and Italy. The results show that traditional external economies

affect the location of creative industries in Europe, complemented by the effect of specific creative

externalities. These findings offer a novel insight into the determinants of location of creative industries

and provide some empirical basis for the design of policies that may boost the capacity of territories for

creativity and innovation, in line with the objectives set out by the European Commission.

1. Introduction

The spatial clustering of firms is one of the core research questions of urban and regional economic

studies. Acs and Varga (2002, p. 134) for instance, underline that a central research issue in economics is

to explain why economic activities tend to be concentrated in certain places while in other places they

remain relatively underdeveloped. During the past two decades, firm clustering has also become relevant

for sub-national policies (Malmberg and Maskell, 2001, p. 4). Indeed, governments (local and regional)

from developed economies have introduced and implemented policies aiming at facilitating the

emergence of clusters as well as at supporting existing clusters (Karlsson, 2008, p. 1).

During the last decade creative industries have been one of the main topics of urban policy and research.

The importance of creativity has been highlighted in fields such as cultural geography, sociology, urban

planning, innovation and economic development, and at the same time it has given place to the

development of new concepts, such as creative economy (UNCTAD 2008), creative city (Landry 2000) or

the creative class (Florida 2002). To date, however, only a limited number of studies have tried to explain

the spatial and industrial conditions that help to understand the clustering trends of creative activities.

The current interest in agglomeration has old roots and traditions. Each of these traditions is related to

different advantages generated by spatial proximity which motivate firms to locate close to other firms.

Authors from the traditional location theory such as Marshall (1890), Weber (1909), Ohlin (1933) and

Hoover (1937) and authors from the clustering theory such as Gordon and McCann (2000) and Boshma

and Frenken (2009) helped researchers to understand why firms tend to cluster in the space. Recently a

limited group of researchers have renewed the interest on the understanding of the factors that explain

2

why creative industries, in particular, tend to be geographically concentrated (Hanson 2000; Tshang and

Vang 2008; Vang 2005, 2007; Lazzeretti et al. 2008, 2012). Indeed, there is a need to understand if the

multiple types of externalities that contribute to explaining the spatial concentration of the economic

activity in general can also help to explain the spatial organisation of creative industries in particular

(Vang 2007). Authors such as Tschang and Vang (2008, p.3) suggest that traditional approaches provide

only a partial explanation of the determinants that might affect creative industries.

In this framework, the purpose of this paper is twofold: (i) to review different definitions proposed for

creative industries and the models suggested in the literature to explain their spatial concentration; and (ii)

to test through 3 case studies (Spain, Portugal and Italy) several determinants of creative industry

concentration.

This paper is organized as follows. The second and third section presents a literature review of the

definition of creative industries and their spatial organization over space. The fourth section reviews the

main traditional and clustering theories used to explain the advantages associated to the concentration of

traditional industries. The fifth section presents the creative industries data source the concentration of the

creative industries. in Spain, Portugal and Italy. The sixth section develops the Count Data Model used

for the analysis, the econometric estimations and the variables used. The seventh section presents the

variables under study and the eight section present the main results. section 9 presents the conclusions and

policy recommendations.

2. Defining creative industries

Defining the creative industries’ sectors is not an easy task. There has been an extensive debate in the

literature regarding what activities should be included within creative industries. Several researchers and

international organisations have identified some creative sectors, but there is not yet a list of activities

universally accepted. One of the main reasons of the difficulties for identifying creative sectors has been

the constant technological evolution of the creative sector.

The term "creative industries" has its origins in Australia in the beginning of the 90's. More concretely,

the term was originally coined in 1994, in a cultural and economic policy document entitled "Creative

Nation" (KEA 2006, p.46). The term gained more popularity among policy makers when the United

Kingdom government (Department of Culture, Media and Sport) set up the Creative Industries Task

Force in 1997. This Task Force helped to promote the role played by creative industries at the heart of the

sub-national economic strategies (O'Connor 2007, p. 41) with the objective to turn UK into "the world's

creative hub".

Recently a number of different models have been put forward in the literature to define creative industries

(NESTA 2008, p. 3; UNCTAD 2008, p.10). Galloway and Dunlop (2006, p. 35) identify five structural

3

characteristics that help to define creative industries: creativity, intellectual property, symbolic meaning,

use value, and methods of production.

a) Creativity: The Anglo-Saxon definitions of creative industries are based on the concept of

innovation and creativity, ranging from the technological to the most artistic innovation

(O'Connor and Xin 2006, p. 272; Brandellero et al. 2006, p. 3). Marshall (1980/1963, p.124)

emphasised that although people do not have the capability to create material, they can produce

value when they give useful forms to things. In other words, the knowledge generation that

comes from the creation of new ideas, new technologies or new business models, is an intrinsic

capability of people (Florida 2005b, p. 32-34). In this line, the UK Department of Culture, Media

and Sport (2001, p.5) defines the creative industries as those "which have their origin in

individual creativity, skills and talent and which have a potential for wealth and job creation

through the generation and exploitation of intellectual property". However the assumption that

any activity that involves creativity will be part of a creative industry has some shortcomings, as

highlighted by O’Connor (2000, p. 10). The author rightly argues that creativity can be applied

to other industries which cannot be included under the umbrella of creative industries, such as a

city cleaning service.

b) Intellectual property: As highlighted by UNCTAD (2010, p.6) this model focuses on industries

that are involved on the production and distribution of copyright goods1. Similarly, Power and

Nielsen (2010, p.28) highlight that goods and services produced in creative industries are

generally defined as intellectual property2 subject to copyright

3. In this line, Howkins (2001, p.

xii-xiii) defines creative industries as those that produce or deal with copyright, patents,

trademarks and designs. This argument was also included in the UK government’s approach to

define creative industries when saying that creative industries are those that "generate and

exploit intellectual property" (DCMS 2001, p.5). However, Galloway and Dunlop (2006, p.36)

underline that many industries which are not creative industries generate intellectual property

(i.e. academia). Therefore, defining creative industries exclusively on the basis of their capacity

to generate intellectual property seems to be inappropriate as well.

c) Symbolic meaning: Another approach to define creative industries is based on the idea that

creative industries process4 and transfer popular cultural value (UNCTAD 2010, p.6). This is

based on a shift of the concept of culture from describing individual intellectual and artistic

cultivation to a set of attitudes, beliefs, customs, values and practices shared by a group of

individuals (EU 2006, p. 44; Galloway and Dunlop 2006, p.38). Several authors have followed

this approach to define creative industries. O'Connor (1999, p.5), for instance, defines creative

1 In this sense, inventors own the products of their creativity and are entitled, by intellectual property rights, to exercise both

economic and moral rights over these products (Gowers 2006, p. 11-12). 2 Intellectual property rights can be regarded as a collection of legal protection rights given to creators to protect their ideas or information having commercial value (Hansen et al 2003, p. 4). 3 The rational idea behind the implementation of these rights on the definition of creative industries is the following: knowledge and

ideas can be easily consumed or copied given that they are partially public goods. However, since their development could be expensive, without a system of exclusive rights there would not be much incentives to pursue innovative activities (Pro inno 2007,

p. 26) 4 Products bear the symbols of the territories in which they are produced giving rise to the notion of “idiosyncratic” products (OECD 2005, p.8)

4

industries as "those activities which deal primarily in symbolic goods - goods whose primary

economic value is derived from their cultural value ". Similarly, Garnham (1987, p.55) defines

creative industries as those that "produce and disseminate symbols in the form of cultural goods

or services". In this line the UNESCO (2005, p.6) defines creative industries as "those which use

creativity, cultural knowledge and intellectual property to produce products and services with

social and cultural meaning". However, one of the shortcoming of this model highlighted by

Galloway and Dunlop (2006, p.40) is that some goods that have symbolic meaning might not be

always produced within creative industries (i.e. painting).

d) Use value: Another approach is based on the fact that industries that use the creative output of

other creative industries in their production process are important actors for the spread of

creativity in society (Galloway and Dunlop 2006, p.39). Following this trend, Caves (2000, p.vii)

defines creative industries as suppliers of a range of products that "we broadly associate with

cultural, artistic, or simply entertainment value". In this line, Potts (2009, p. 9) highlights that

creative industries help other sectors to organise, adapt and retain innovations. Additionally,

UNCTAD (2010, p.6) underlines that this model has been the basis for classifying creative

industries in Europe. Nevertheless, Flew (2002, p.13) has observed that an important

shortcoming of this approach is that is has an extremely wide coverage and therefore includes

almost any industrial activity.

e) Methods of production: Galloway and Dunlop (2006, p.43) consider that creative industries are

characterised by a combination of industrial-scale production and cultural content. In this line,

Towse (2003, p. 170) defines creative industries as those which "mass-produce goods and

services with sufficient artistic content to be considered creatively and culturally significant".

However, Galloway and Dunlop (2006, p.43) point out that the production method is not a

sufficient foundation on which to base a definition of creative industries. Indeed, defining

creative industries on the basis of how they produce can misrepresent what is being produced. As

a result Towse observed that following this model, the creative arts sector would be excluded

from the concept of creative industries since this sector does not typically use industrial-scale

methods of production.

This study takes the symbolic meaning method to classify creative industries. Asheim et al. (2005), based

on the symbolic knowledge model, provide a clear definition of what are creative industries and which are

their main characteristics. According to them, originally creative industries have been restricted to artistic

or cultural activities, however knowledge creation is increasingly important in all segments of economic

activity (Asheim et al 2005, p. 11). Traditional creative industries have experienced a wide modernisation

process (O'Connor and Xin 2006, p. 274) associated to technological change which promote the creation

and exchange of new knowledge. Nowadays, the creative sector5 is characterised by a diversity of

economic activities (Baumount and Boiteax-Orain 2005, p.8), which range from traditional arts and

cultural heritage to more technological and service-oriented activities (UNCTAD 2010, p. 7; 2008, p. 4).

Cultural industries such as media (film making, publishing, music, etc.), advertising, design and fashion

5 Pratt (1997, p. 1959) observes that the term sector is used in this context by certain authors to refer to a group of activities that are linked in the chain of production, as the term filière traditionally used in France.

5

are considered to be based on a symbolic knowledge base mainly characterised by the following elements

(Asheim and Vang 2005, p. 29-30; Asheim et al 2007, p. 11):

a) Innovation in these industries is mainly produced by a recombination of existing knowledge.

Creativity and innovation requires that the new knowledge is combined with the current

knowledge generating new ideas (OECD 2008, p.22).

b) The innovation process of these industries is mainly market oriented since their main purpose is

to release products that maximise the return on their investment.

c) Products tend to be ephemeral since they attract the consumer's attention for a limited period of

time; their product is mainly meant to entertain.

d) The knowledge used in the innovation processes of these industries is mainly tacit knowledge,

which is normally linked to the habits and norms learned in specific communities. This

knowledge is usually incorporated and transmitted through aesthetic symbols, photos, videos,

graphs, designs, artefacts, sounds and narratives. Abilities to cope with this strong semiotic

content are therefore required.

e) These are project-based industries where a group of agents work together with the aim of

producing a new product. Knowledge exchange takes place mainly through informal

interpersonal interaction in the professional community (face-to-face).

3. Spatial organization of creative industries

Recent studies have analysed the spatial organization of creative activities. They emphasise that creative

activities are not distributed uniformly across space (Scott 2005; Cook et al. 2007; Florida 2008; Florida

et al. 2008; Cook and Lazzeretti 2008; Lazzeretti et al. 2008; Boix et al. 2012). Several authors point out

that activities with a high propensity to innovate tend to be more clustered than manufacturing industries

(Feldman 2000, p. 378-379; Scott 2000a, p.327) due to their own characteristics.

Boix et al. (2013, p.10) point out that symbolic knowledge and knowledge spillovers tends to be locally

sensitive. Industries such as media, advertising, design and fashion with a symbolic knowledge base, are

mainly based on tacit knowledge. This knowledge is normally linked to the habits and norms learned in

specific social groups and which are exchanged mainly through informal interpersonal interaction in the

professional community (face-to-face).

And indeed spatial proximity matters in the innovative or creative process (Cooke et al 2007, p.30).

Marshall (1980/1963, p.226) points out that in a knowledge-dense context, tacit knowledge6 can benefit

from spatial proximity. Indeed, these kinds of knowledge spill overs occur mostly among geographically

proximate individuals and organisations (OECD 2008, p.10). Several scholars see this as a dynamic

process called "knowledge spiral" (Nonaka and Takeuchi 1995; OECD 1996; Becattini 2005, pp. 52-53;

6 Tacit knowledge cannot be codified or expressed by words. Thus, the only way this knowledge can be transmitted is by personal

contact or socialisation processes (Montuschi 2001a, pp. 14-15). Codes or language are not developed well enough to permit the

transmission of this knowledge (Gertler 2003, p.78). Polanyi (1966, p. 4,136) summarises this by saying "we can know more than we can tell".

6

Cooke et al 2007, p.29): the inventions as well as the organisational or process improvements achieved by

a company located in a territory are made explicit, and then shared, analysed and adopted by the rest of

the companies located in the same territory. Such knowledge is more effectively transmitted in a local

context where there is proximity between individuals with a common social context (OECD 2008, p.8).

Thus, as Audrersch and Feldman (1996, p.634, 637-639) note, industries which are more knowledge

oriented will be expected to be more concentrated, given the need of transmitting tacit knowledge

informally trough face-to-face interactions and repeated contact (Pratt 2004, p. 122; Audretsch 1998,

p.21; Von Hipple 1994; Audretsch 2003; OECD 2008, p. 28).

New ideas generated from the spatial concentration of creative activities, give place to more creativity

intensive locations (Maskell and Lorenzen 2004; Cook et al. 2007). Creativity will thus be concentrated in

creative clusters (Le Blanc 2000; Lazzeretti et al. 2008), Industrial Districts and cities or Metropolitan

areas. Building on this observation, De Propris et al. (2009) define a creative cluster as “a place that

brings together (a) a community of ‘creative people’ (Florida 2002) who share the same interest in

novelty but not necessarily in the same subject; (b) a catalyzing place where people, relationships, ideas

and talents can spark each other; (c) an environment that offers diversity, stimuli and freedom of

expression; and (d) a thick, open and ever-changing network of interpersonal exchanges that nurture

individuals’ uniqueness and identity”.

4. Location theories

People and firms are not homogeneously distributed over space. Indeed, the spatial distribution of the raw

materials, productive factors, technology and the demand of final goods force firms to select their location

(Capello, 2004, p. 41)7. The spatial clustering of firms has been one of the core research questions of

urban and regional economic studies. Acs and Varga (2002, p. 134) for instance, underline that a central

research issue in economics is to explain why economic activities tend to be concentrated in certain

places while others remain relatively underdeveloped.

Industries demonstrate that creative firms agglomerate in specific places in order to benefit from

advantages8 generated from geographical and sectorial proximity. More specific theories have been

developed over the last century that help to understand what are the advantages behind the concentration

of creative industries over space.

7 Classic theories such as David Ricardo or Adam Smith, did not pay attention to space as an important element to understand the

economic development of firms (Capello, 2004, p. 42). 8 Even if the agglomeration of industries in specific locations can also generate negative externalities, this paper only focuses its analysis on the advantages.

7

4.1. Global framework

The Industrial Location Theory is the oldest branch of regional economics having the objective of

studying the static economic mechanisms that explain the organization of the activity9 in the territory

(Capello, 2004, p. 23). According to this theory, the two main economic forces that determine the

organization of the economic activity over space are: reduction of transport costs and increase of

productivity.

a) Reduction of transport costs: Alfred Weber in 1909, based on the seminal contribution of von

Thünen (1826), identified the reduction of transportation costs as the main determinant of

industrial location. Indeed, under several assumptions10

, Weber showed that the optimal location

of the firm is based on the minimization of the transportation costs between the raw materials

and the final product market. This theory is considered as the foundation of modern location

theories11

. Actually, Weber’s theory shed some light on the spatial location of firms where the

frequency of delivery of goods and services is high, and also on the location of industries where

the transportation of the raw material is really costly (i.e. heavy industry), particularly in the first

half of the twentieth century.

Creative industries are defined by Asheim and Vang (2005, pp.29-30) and Asheim et al. (2007,

p.11) as characterized by the fact that innovation is mainly produced by a recombination of

existing tacit knowledge. Thus, even if organization and transfer of tacit knowledge does not

imply transportation costs issues, creative industries require spatial proximity to other agents,

who are holders of tacit knowledge.

b) Increase of productivity: Agglomeration economies explain the tendency to spatial concentration

based on the increase of efficiency and thus reduce production costs. Indeed, agglomeration

economies can be defined as all economic advantages that firms can benefit from in concentrated

locations. This concept could be defined as the result of the combination of the conceptualization

developed by four authors: Marshall (1890/1963, pp.222-225) uses two essential elements

(natural resources and internal and external economies) to explain the production, which can be

interpreted as location factors. Weber (1929/1968, pp.124-173) introduces the concept of

“factors of agglomeration” understood as transport costs advantages, to refer to the elements that

cause a dense industrial localization on the territory. Hoover (1937/1971, pp. 90-91) clarifies and

extends the concept of “concentration economies” building on Ohlin (1933, p. 203). These

9 Location theory explains the economic mechanism of the localization of firms but also of the residential activities and the configuration of urban systems (Capello, 2004, p. 42). 10 Weber’s theory was based on the following assumptions: (1) perfect competition exists in the market, (2) there are perfectly

mobile factors of production, (3) firm’s technology exhibits constant returns to scale, (4) raw material and output markets are fixed at certain specific points, (5) production factors are available in unlimited supply, (6) transport costs are proportional to the weight

of the goods and distance to the markets. 11 Indeed, Weber’s location theory was extended by several authors in the second half of the twentieth century, such as Isard (1959), Moses (1958), Sakashita (1987), Sheih and Mai (1997). For more details see, Zhang (2002, pp. 3-5).

8

external economies are usually divided in two sub-categories (Hoover 1937): localization and

urbanization economies.

The localization economies (Marshall 1890, p. 222) are the advantages derived from the

concentration in a particular location of specialized companies, suppliers and workers. In

preliminary stages of the development of a technology or a new product or service, fast and

direct contact with other concerned actors, such as specialized suppliers, workers and firms, is

needed (Prager and Thisse 2009, p.43). Creative industries are project-based industries where a

group of agents work together with the aim of combining existing tacit knowledge and produce a

new product. Thus, creative industries can benefit from the location of firms near specialized

companies or suppliers since this enhances professional association and knowledge exchange.

Cook et al. (2007, p.31) and Audretsch (1998, p.18) claim that since knowledge is created and

shared more efficiently at local proximity, firms based on a combination of existing knowledge

will have a high propensity to cluster over space. In this line, Carlton (1983, p.446) found that in

industries where sophisticated technology is needed, the presence of concentrated technical

expertise from the same industry or related industries is crucial. Additionally, location of

educated people and the necessary infrastructures for their formation (such as universities and

other educative institutions) plays a fundamental role in the location and performance of

companies, especially for those where individuals with high levels of human capital constitute a

primary input to the production process (Arora et al., 2000). In fact, the externalities generated

by the concentration of human capital in a place can be seen as a reason for the clustering of

economic activities (Glaeser, 2000; Fritsch and Stützer, 2008) as well as the generators of new

ideas, and the attractors of new creative firms (Lazzeretti et al., 2009). Indeed, creative industries

are characterized by generating products that tend to be ephemeral since they attract the

consumer’s attention for a limited period of time. In this sense, creative industries need to adapt

their production and thus their labour force to the market needs. Thus, spatial proximity of

creative industries to the labour market might facilitate efficient matching between labour supply

and demand.

Urbanization economies are advantages derived from the urban environment factors or

characteristics, which are directed in an indistinct way (without coming necessarily from the

same productive sector) to all the economic activities that are located in it (Camagni 2005, pp.

24 and 34). In this line, Chinitz (1961, pp. 281-282) presented the economic variety (or

diversity) as determinant of economic concentration.

Turok (2003, p.562) underlines that the city size as well as the density of the economic agents of

a territory determine the importance of the benefits that creative firms could gain from their co-

location. Indeed, the innovation process of the creative industries is mainly market oriented.

Thus, creative industries will cluster in particular locations to take advantage of close proximity

to concentrations of customers.

9

4.2. Clustering models

Since the early 1990’s, industrial clusters received considerable attention by both policy makers and

researchers. Literature on industrial clusters has focused its attention on the causes of a non-random

spatial concentration of economic activity in space. Over the specialized literature several typologies of

clusters can be summarized in three theories: Gordon and McCann; Evolutionary Economic Geography

models and National cluster policy initiatives.

a) Gordon and McCann (2000, pp. 516-521) provided a comprehensive assessment of various

typological frameworks used in the literature for the analysis of industrial clusters. The authors

adopting a transactions-costs perspective from the relations between the firms within the cluster,

such as transportation or communication costs, identify three distinct types of industrial clusters:

pure agglomeration, the industrial complex, and the social network.

The classic model of pure agglomeration, refers to the external economies of scale or scope that

benefit firms located in the same area. These externalities can arise from three different sources

(Gordon and McCann 2000, p. 516): First, firms benefit from access to a more extensive labour

pool, helping firms to find skilled labour force and thus to maximize the job matching by

adjusting labour needs according to the market conditions. Second, firms benefit from access to

a large range of specialized industries and suppliers. Third, firms benefit from the exchange of

knowledge between specialized and concentrated firms, such as in the filière. Glaeser et al.

(1992, p. 1127) also suggest that these knowledge externalities are often shared through the

inter-firm movement of highly qualified people. Such external economies are essentially

economic externalities that are derived from a geographical proximity between economic agents.

Thus, co-location of creative industries might increase their opportunity to benefit from a skilled

labour pool, to trade with specialized suppliers as well as to cooperate with other specialized

firms to overcome market uncertainties.

The second model of clustering is the industrial complex, which is characterized mainly by

stable trading relations between firms in the cluster (Gordon and McCann 2000, pp. 518-519). In

this line, Rosenfeld (1992) has demonstrated the importance of fostering cooperation and

collaboration in industrial environments were multiple small firms coexist. According to

Boschma and Iammarino (2009), “related variety” is understood as industrial sectors that are

characterized by complementary competences. The concentration of these elements in the same

place could facilitate the generation of a dense and varied network of agents that foster economic

and social collaboration, enhancing knowledge transfer through cross-fertilisation mechanisms

and promoting innovation (Lazzeretti et al. 2011; Lorenzen and Frederiksen 2008, p.171). It is

important to note that, some authors have shown that the access to a diversified pool of firms

will not have the same effect as a pool of diversified related firms and industries (Porter 2000, p.

259). According to Lazzeretti et al. (2012, p. 1246) related variety promotes creativity due to

spillover processes of innovation in other sectors. In this line, Porter (1990, p. 52) notes that

innovation will be fostered in geographically concentrated clusters of small firms due to: (1)

10

strong local rivalry, that requires firms to distinguish themselves through creativity, and (2) the

changing final product demand which requires cooperation and collaboration among firms with

complementary products (marketing, research, among others) to act rapidly and turn

opportunities into real products and thus maintain the cluster reputation.

The third form of clustering is the social network that facilitates cooperation between firms

(Gordon and McCann 2000, pp. 519-521). According to authors such as Malecki (1994),

Camagni (2008) and Bergman et al. (1991), cooperation requires the presence in the territory of

social networks and relationships of trust. In the creative domain, creative industries are

characterized by the need of flexible production units in order to change their process and

product configurations according to the unstable and changeable needs of the market. In order to

operate in this way, Scott (2006, p. 5-6) highlights that creative industries are generally

connected to dense networks of specialized and complementary firms, which require high levels

of trust to allow for the flow of information and ideas between them. Scott (2006, p. 6) also notes

that these networks of creative firms are frequently dominated by large firms.

b) Over the past decade, economic geography has been influenced by evolutionary thinking giving

place to Evolutionary Economic Geography (EEG) models. EEG aims to explain the uneven

distribution of economic activities underlying the industrial dynamics of firms (Boschma and

Frenken 2009, 2). Boschma and Frenken (2010, p.6) underline four elements that explain the

spatial concentration of firms: first, self-reinforcing and irreversible dynamic processes; second,

path dependency on early decisions in the formative stage; third, location choices and fourth,

market competition driven by scale economies at the firm level. Additionally, according to this

theory, spatial clustering of firms is the result of spatial historical conditions that contribute to

the creation, maintenance and transmission of established organizational procedures, underlying

the importance of related variety. In this line, Berg and Hassink (2013) contributed to explaining

the spatial distribution of creative industries by using an evolutionary economic prospective.

Indeed, they found that five essential elements to explain this spatial trend are path dependence,

lock-ins, path creation, related variety and co-evolution.

c) Cluster policy initiatives have been used in several European countries as a platform to increase

innovation and thus to contribute to sustainable growth. In recent years, national and regional

authorities have started to see creative industries as important elements of the economic

performance of their territories, and thus have started supporting initiatives through industrial

policies (NORDEN 2006, p. 8). For instance, cluster initiatives such as the VinnVäxt in Sweden,

have provided funding for knowledge-intense Clusters initiatives since 2002, strengthening thus

the linkages between local nodes of knowledge an innovation. This model is also known as

“triple helix” and the main idea is to enable effective cooperation between companies,

institutions of research/high education and public administration. One of the main aims of this

program is to create attractive environments in order to attract national and international

companies and researchers.

11

4.3. Specific creative models

Authors such as Tschang and Vang (2008, p. 3) suggest that traditional approaches only provide a partial

explanation of the determinants that might affect creative industries. Other authors have incorporated in

the literature other elements that could help to explain the tendency of creative firms to concentrate over

space.

Currently, the academic debate on the determinants of creative industries is shifting from a business to a

more people-oriented approach (Selada et al. 2010, p. 5). According to Turok (2003, p. 562) these

amenities are important elements to attract and retain highly skilled workers, which tend to be extremely

mobile. Residential or worker amenities are exogenous goods or services that could increase the

attractiveness, value or comfort of a specific place. Based on Boix et al. (2013, p. 3), these alternative

explanations can be summarized taking into consideration the particular characteristics of creative

industries clusters: existence of cultural infrastructure, presence of ‘soft’ factors, access to gatekeepers,

presence of patronage, proximity to political power, location of ‘star’ artists and creative class, existence

of a particular identity and a place brand and image.

a) The presence of cultural infrastructure and the proximity to political power have been

highlighted by several authors such as Sivitanidou (1999, p.9), Viladecans (2002, p. 9), van Oort

et al. (2003, p. 516), Reardon (2009, pp. 13-16) and Selada et al. (2010, p. 7-10). According to

them, it is possible to identify 3 categories of non-productive amenities that affect residents and

workers utilities: (i) Governance, understood as the leadership and management of places as well

as the coordination of different actors and innovative and creative policies; (ii) Natural and

historical- cultural amenities, understood as the natural, architectonic, archeological heritage,

the urban landscape and image, the climate and the public spaces among others. Indeed, Kourtit

et al. (2013, p. 4) highlight that the presence of historic authenticity in a place, such as cultural

heritage, contributes to the creation of an appropriate urban location that favors creative minds.

(iii) Good access to economic activities and cultural facilities, understood as the structures

essential for the health, social wellbeing and economic prosperity of local communities.

b) Several authors highlight also the existence of a particular identity, place brand and image and

the importance of ‘soft’ characteristics to explain spatial concentration of creative firms. Indeed,

it is in these places, according to several authors, where creative people prefer to live

(Sivitanidou 1999, p. 25; van Oort et al. 2003, p. 521). In this sense, Markusen, Hall and

Glasmeier (1986) and DeVol (1999) highlight the impact of quality of life on the spatial

distribution of innovative firms. According to them, places with accessible natural environments

can facilitate the attractiveness of certain places to creative firms and employees. Additionally,

the level of openness, diversity and opportunity to work (tolerance), also highlighted by Jacobs

(1961 and 1969), Florida (2002), Saxenian (1994), Bounken (2009, p. 189) and EIS (2008, p. 11)

is also another characteristic that creative people valuate positively. Additionally, Bourdieu

(1960) defines social capital as the set of actual or potential resources related to a long lasting

12

network or relationships among a set of individuals. In this sense, the presence of public or semi-

public spaces such as bars, restaurants can help people to meet (Murphy and Redmond 2009, p.

73) and thus to facilitate social interactions.

c) Additionally, the presence of people working in creative occupations can attract other kinds of

talent and creative firms (Clifton and Cooke 2007, p.23). The externalities generated by the

concentration of human capital or creative class in a place can be seen as a reason for the

clustering of economic activities (Glaeser, 2000; Fritsch and Stützer, 2008) as well as creative

industries (Lazzeretti et al. 2009). Assmo (2010, p. 314) shows that creative actors are crucial for

the development of new creative and cultural firms and products. In this context, knowledge

exchange takes place mainly based on cognitive proximity, through informal interpersonal

interaction in the professional community (face-to-face). Face-to-face interactions between

different actors will facilitate spillovers of information, know-how and technology by imitation

or learning (Suarez-Villa and Walrod 1997; Globerman 2001; Cook et al. 2007), and as a result

new ideas will emerge. As it is emphasized by Lucas (1988, p. 38) and Hanson (2000, p.480)

creative professionals such as musicians or actors may learn from other colleagues working in

the same environment techniques that will improve their performance.

d) Finally, patronage and gatekeepers can contribute to the diffusion of knowledge inside the

cluster. In this line, Scott (2006, p. 6) stresses that in creative environments, free exchange of

information between members of a network is of critical importance for the development of new

production processes. However, he also points out that low levels of trust can impede the

knowledge flow between the members of the network. Scott highlights that industrial

associations or private-public partnerships can sometimes solve knowledge spillover failures in

competitive environments.

5. Location of creative industries in European Local Labour Systems

5.1. Territorial level of analysis: Local Labour Systems

The Spanish, Portuguese and Italian territories are mainly organized in four administrative levels (NUTS

1 or major socio-economic regions, NUTS 2 or basic regions, NUTS 3 as small regions or provinces and

local administrative units (LAU) defined as municipalities.

These administrative levels do not capture neither the economic nor the social interaction area. NUTS 2

were used by Power and Niélsen (2010) to provide initial evidence of the clusters of creative industries in

Europe. However, as it becomes evident from the mapping, these units are excessively large to capture

the real processes of clustering and give only a preliminary idea of the concentration. In fact, as pointed

out by Lazzeroni (2010), the regional and provincial scale would seem too broad and diversified to

represent the real economic area while the municipality level does not capture all the spillovers that occur

in a creative cluster since its spillovers usually extend to neighboring municipalities.

Local functional units such as the local labour systems (LLS) have the advantage over the administrative

boundaries to better portray current social and economic conditions, because their boundaries are made

13

according to commuting data (such as commuting flows from home to work). For that reason, several

researchers have used these territorial units in their location analysis. Indeed, Overman and Puga (2010)

use TTWA (Travel to work areas) to analyse the manufacture establishment location in the UK. Similarly,

Lazzeretti et al. (2008) and Boix et al. (2012) use labour markets (or systems) (LLS)12

as the territorial

unit for the study of the processes of creative clustering in Europe.

5.2. Data source

A growing number of researchers have used Bureau van Dijk´s firm-level dataset’s in recent years to

analyse the spatial location of economic activities, including international studies such as Abramovsky et

al. (2008) or Driffield and Menghinello (2010). Based on UNCTAD (2008, pp.286-287) and Lazzeretti et

al. (2010, p.26) the creative sector analysed in this study includes a range of economic activities:

Advertising, Architecture, Arts and antique markets, Crafts and performing arts, Creative R&D, Design,

Fashion, Film, Heritage, Jewellery, Music, Pothography, Publishing, Printing, Radio and television,

Software and Toys and Games. Table 1 provides the 4 digits for the NACE industries included in the

creative sectors analysed in this study.

Table 1. Creative industries (NACE rev. 1.1)

Advertising & related services

Heritage / Cultural sites

7440 Advertising

9251 Library and archives activities

9252

Museums activities and preservation of

historical sites and buildings

Architecture

9253

Botanical and zoological gardens and nature

reserves activities

7420

Architectural and engineering activities and

related technical consultancy

Jewellery (manufacturing)

Arts and antique markets/trade

3621 Striking of coins

5212 Other retail sale in non-specialized stores

3622

Manufacture of jewellery and related articles

n.e.c.

5248 Other retail sale in specialized stores

5250 Retail sale of second-hand goods in stores

Musical instruments

5263 Other non-store retail sale

3630 Manufacture of musical instruments

Crafts/Performing arts/Other visual arts

Music / Sound recording industries

9231

Artistic and literary creation and

interpretation

2231 Reproduction of sound recording

9232 Operation of arts facilities

9233 Fair and amusement park activities

Photography

9234 Other entertainment activities n.e.c.

7481 Photographic activities

Creative R&D

Publishing

7310

Research and experimental development on

natural sciences and engineering

2211 Publishing of books

7320

Research and experimental development on

social sciences and humanities

2212 Publishing of newspapers

2213 Publishing of journals and periodicals

Design / Specialized design services

2214 Publishing of sound recordings

7487 Other business activities n.e.c.

2215 Other publishing

Designer fashion

Printing

1771

Manufacture of knitted and crocheted

hosiery

2221 Printing of newspapers

1772

Manufacture of knitted and crocheted

pullovers, cardigans and similar articles

2222 Printing n.e.c.

12 Boix and Galletto (2006a and 2006b), following the ISTAT (2005 and 2006) methodology identified 806 Local Labour Systems in Spain.

14

1810 Manufacture of leather clothes

2223 Bookbinding

1821 Manufacture of workwear

2224 Pre-press activities

1822 Manufacture of other outerwear

2225 Ancillary activities related to printing

1823 Manufacture of underwear

1824

Manufacture of other wearing apparel and

accessories n.e.c.

Radio and television (Broadcasting)

1830

Dressing and dyeing of fur; manufacture of

articles of fur

9220 Radio and television activities

1930 Manufacture of footwear

Software, computer games and electronic publishing

Film / Motion picture & video industries

7221 Publishing of software

2232 Reproduction of video recording

7222 Other software consultancy and supply

2233 Reproduction of computer media

7260 Other computer related activities

9211 Motion picture and video production

9212 Motion picture and video distribution

Toys and games (excluding video games)

9213 Motion picture projection

3650 Manufacture of games and toys

Source: Based on UNCTAD (2008) and Lazzeretti et al. (2010)

5.3. Data quality

As underlined by Driffield and Menghinello (2010, p. 17) there is a source of bias concerning the use of

Bureau van Dijk’s database that needs to be mentioned. In relation to the use of firm level data as a proxy

of local units (establishments) data, the magnitude of the bias between the real number of establishments

and the establishments provided by the database is related to the presence of multi-plants firms and the

geographical scale of the territorial unit of analysis. However, given that ORBIS relies on country level

national sources, this bias is assumed to be limited. Indeed, comparing ORBIS data with establishment

data from SBS (Structural Business Statistics from Eurostat) at national level it has been observed that

ORBIS accounts for around 20% of the industries registered in Spain, Portugal and Italy, and this share

has been observed also among sectors and subsectors.

In this research, the use of ORBIS is justified by two main reasons:

a) This paper is an exploratory analysis applied to the case of Spain, Portugal and Italy. The final

analysis will include data for other European countries considered in the thesis, namely France

and the United Kingdom. Obtaining territorial administrative data disaggregated by creative

industrial sectors and homogeneous across countries is a difficult task. As it has also been

observed by Driffield and Menghinello (2010, p. 4) data provided by official statistics normally

present significant confidentiality and data quality constraints. The coverage of firm-level data in

EU countries in the ORBIS database allows to overcome these limitations.

b) Furthermore, additional individual-firm data provided by the ORBIS database such as firm size,

will also be used in the analysis. Despite its limitations, ORBIS can be considered one of the few

sources that contain this kind of data.

5.4. Concentration of creative industries in Spain, Portugal and Italy

The territorial distribution of creative economic activities in Spain, Portugal and Italy is obtained from the

information provided by the ORBIS database. Based on the address provided by this database, it has been

possible to geocode around 638,062 economic activities of all productive sectors in Spain in 2009,

275,657 in Portugal and 827,273 in Italy. Among these, around 10% of the firms can be classified as

creative industries (73,464 creative industries were identified in Spain, 34,358 in Portugal and 72,117 in

Italy). Figure 1 presents the creative industries identified in Madrid, Lisbon and Rome.

15

Figure 1. The location of creative industries in Madrid and Barcelona

A) Madrid (city) B) Lisbon (city) C) Rome (city)

Source: Based on ORBIS data

In this study, creative industries plants have been computed for each of the LLSs in Spain, Portugal and

Italy (806, 83 and 686 respectively). Figure 2 shows that most of these plants are concentrated around the

LLSs of major capital cities, highlighting a strong spatial concentration of creative industries over space.

Additionally, figure 3 presents the share of firms (total and creative) in the 2 major cities in each country

showing that this share is larger for creative industries that for all industries in all 6 cases. This

differential is even larger in cities like Madrid or Roma.

Figure 2. Number of creative industries by LLS in Spain, Portugal and Italy (2009)

Source: Based on ORBIS data

16

Figure 3. Share of total and creative firms in a selection of cities (2009)

Source: Based on ORBIS data

6. Model

Ellison and Glaeser (1997, p.892) suggest a location model based on the existence of natural advantages

and externalities or inter-firm spillovers inside the same industry. This model assumes an industry divided

in N business units, which choose in a consecutive way their location among the M areas in which the

territory is divided. In this case, and to make the model tractable, the authors take only one company to

illustrate the model. Thus, the kth business will maximize its profits through their decision to locate vk

inside the area i, by the following function:

kikiiki vvg ),...,(loglog 11 (1)

where i is a random variable reflecting the probability of locating in area i (as influenced by observed

and unobserved area characteristics), vj is the location of the business j, while ki is the random

component.

Equation (1) shows that the profits derived from the location of a business are related with two elements.

Firstly, they are related to an average measure of the territory profitability (general-economic factors),

and secondly, to a random variable that collects idiosyncratic elements of the industry (specific-creative

forces). The authors suggest a simple parametric specification of this model.

kl

kilikliki ue ))(1()log(log (2)

Where kle is the Bernouilli random variable equal to one with probability 0 that indicates whether a

potential valuable spillover exists between each pair of plants, and liu is an indicator for whether plant l

is located in area i (vl=i), and ki , again, is a random component independent from kle .

17

Discrete choice models are used to analyse the location from the perspective of the firm. Researches done

following these models are particularly focused on the individual elements of firms as determinants of the

location of each firm, such as dimension of the firm or the sector to which the firm belongs (Manjón and

Arauzo-Carod 2006). However, one of the main drawbacks of this empirical approximation is the

difficulty to calculate the likelihood function when there are so many location alternatives, which is so

common at a local level (Arauzo-Carod 2007, pp.4-5).

According to Guimaraes et al. (2003), a possible solution could be to apply Count Data Models which

allow to use large data sets (the number of alternatives in a Conditional Logit Model equals the number of

observations in a Count Data Model). Thus the increment of alternative locations when analysing the

phenomenon at a local level is not a major problem using a Count Data Model. Moreover, null

observations (territorial units that do not locate any industry over the analysed period) do not imply

modelling problems in Count Data Models (unlike Conditional Logit Models).

Thus, the count models allow to analyze the localization of creative industries from the geographical

space chosen (municipality, region or non administrative territory). The characteristics of the territory

analysed (telling apart general-economic and specific creative forces) will affect the probability to be

chosen as the location of a company. Since this paper aims at providing evidence of the determinants of

location of creative firms in the LLSs in Europe from a territorial perspective a count model will be used.

6.1. Count Data Models

Figure 4 displays the histogram of the frequency of the dependent variable (number of creative firms by

LLS) in the Spanish, Portuguese, Italian and in all three countries together. As it is also observed in the

industrial location literature (Arauzo-Carod et al. 2010, pp.692-696), the distribution of creative

industries in Spain, Portugal and Italy appear to be highly skewed. Indeed, there are many LLS which

have few or no creative industries.

Figure 4. Frequency of the number of creative industries by LLS (2009): histogram

Spain Portugal

Italy Spain, Portugal and Italy

01

02

03

04

0

Fre

que

ncy

0 5000 10000 15000 20000CI

01

23

45

Fre

que

ncy

0 2000 4000 6000 8000 10000CI

18

Source: Based on ORBIS data

a) Poisson model

Such industry distribution has to be properly taken into account in the model´s specification and

estimation. Several researchers have opted for an easy solution which is to do the regression of the

dependent variable as a linear combination of several explanatory variables (transforming the dependent

variable with a neperian logarithm). However, count data can potentially result in skewed distributions

cut off at zero, making unreasonable to assume that the response variables and resulting errors follow a

normal distribution. Thus, the problem of non-linearity should be handled through non-linear functions

that transform the expected value of the count variable into a linear function of the explanatory variables.

In this study linear and count data models will be used in the econometric section.

The most popular specification of Count Data (such as the number of creative industries localized in a

LLS) is probably the Poisson model. This model assumes that the probability of observing a count

location (an industry i (such as creative industry) in territorial unit j (such as municipality or LLS))

can be written as a function of specific location characteristics of the territory that affect firms’ spatial

profit function.

( ) ( ) (3)

where denotes the vector of location characteristics that affect the profit functions of firms and act as a

location determinant.

Mathematically, if is the realisation of the aleatory variable based in a Poisson with a parameter

(ratio of occurrence of event of interest). Given a vector of explanatory variables , the density function

of , will have the form:

( | )

(4)

In which the most common representation of the conditional mean is:

01

02

03

04

0

Fre

que

ncy

0 5000 10000CI

02

04

06

08

0

Fre

que

ncy

0 5000 10000 15000 20000CI

19

[ ] ( ) (5)

Where is the parameter vector to be estimated and is a vector of municipality attributes that affect

profit functions of firms.

b) Poisson assumptions

The Poisson regression models are the common starting point for count data analysis. However, count

data might exhibit some characteristics that might violate some of the Poisson assumptions. The use of

Poisson regression in the presence of any of these futures (ex. overdisperion or excess of zeros) may leed

to a poor fit, loss of efficiency and incorrect reported standard errors.

The first assumption refers to the “excess of zeros” problem. Poisson Models can deal with situations

where the dependent variable is characterized by a large number of observations whose value is zero13

.

However, some adjustments need to be done in the model when this number is excessive. Table 2 shows

the % of LLSs where a zero number of creative firms has been identified. Indeed, 31 LLS in Spain, 5 in

Portugal and 6 in Italy do not concentrate any creative industry. These LLS represent a 3.9% in Spain, 6%

in Portugal and 1% in Italy. Thus, these results suggest that there is no need to use other Count data

models such as Zero-Inflated Poisson Model (ZIPM) given the fact that Poisson or Negative Binomial

Models can deal with situations where the dependent variable presents a high number of zero

observations.

The second assumption is generally called “equidispersion”, which implies that the mean and the variance

should be equal. However, unobserved heterogeneity might lead to overdispersion due to the failure of the

assumption of independence of events which is implicit in the Poisson Model. In this line, Arauzo-Carod

(2007, p. 199) points out that industrial location generally violates this assumption, due to the large

concentration of certain firms in few locations. Indeed, as it can be observed in Table 2, the distribution of

creative industries in Spanish, Portuguese and Italian LLSs displays a greater variance than the mean

(ranging from almost 3,000 to 5,000 times larger than the mean in Portugal and Spain respectively).

Table 2: Descriptive statistics: Dependent variable

Spain Portugal Italy All

Mean 91.15 413.95 105.13 114.25

Standard deviation 680.96 1415.28 542.17 688.83

MIN 0 0 0 0

MAX 16664 10361 11086 16664

% of zeros 3.85% 6.02% 0.87% 2.67%

Source: Based on ORBIS data

The presence of overdispersion motivates the use of different distributions than the general Poisson

regression. Indeed, the Negative Binomial model is a generalized extension of the Poisson model which

does not impose equivalence between the mean and the variance. Thus it includes a dispersion parameter

to accomodate overdispersion problems resulting from unobserved heterogeneity distribution of the

13 Given that count models show how many times a location (LLS) has been chosen by a creative firm, the LLSs with no creative firms are relevant for the analysis. Indeed independent variables in these

locations will explain why these territories have not been chosen by any creative industry.

20

dependent variable. This problem can also be addressed by estimating Poison Models using Huber-White

Sandwich linearized estimator, which will be used in this study.

c) Generalized Linear Models

The restrictions observed regarding the general Poisson model can be more appropriately overcome by

the loglinear error distribution suggested by Preston (1948) which has the following functional form:

( | )

(6)

( ) (7)

One of the advantages of assuming loglinear distribution errors is that they provide a flexible framework

to deal with count data models. This model has been proved in the literature to be a feasible alternative to

the Negative binomial model.

Equation (7) will be then represented by the following formulation:

(8)

where is a normal distributed error term

7. Variables

The data used in this paper refer to Spain, Portugal and Italy. The data include one dataset accounting for

the location of creative industries (dependent variable) and another dataset about the territorial

characteristics of LLS in Spain (independent variable).

Econometric studies normally analyze the effect of the explanatory variables on the dependent variable.

However there is the possibility that the dependent variable has simultaneously an effect on the

explanatory variables (Kennedy 2003, p.401). In order to avoid the simultaneous causation bias the

dependent variable has been computed at time t, whilst all explanatory variables in the model are defined

at time t-1. The use of explicative variables established in the initial year of the period might reduce the

problem.

7.1. Dependent variable

According to the previous section, the dependent variable is the location of creative industries in each of

the 1575 LLSs in Spain, Portugal and Italy, which was drawn up using data from the ORBIS database

(2009).

7.2. Independent variables

Independent variables have been divided into two groups according to the theoretical section of this

paper.

a) Localization economies have been addressed by several indicators which approximate the advantages

derived from the concentration in a particular location of infrastructure and organization of the industry

(Marshall 1890, p. 222), specialized workers and suppliers (that belong to the same industry or sector of

production). The following indicators are proposed:

21

- The industrial complex of the creative industry is computed by the average size of the creative

industries in each LLS. This indicator will help to understand if the presence of small or large

firms in the territory is an important determinant of the creative industrial concentration.

, where L is the employment (Jobs) and F is the number of firms, i is the creative industry and j

is the LLS. Employment and number of firms has been extracted from ORBIS database year

(2009) due to lack of business demography data for certain countries14

.

- The market competition is approximated by inverse of a Herfindhal index of the market shares of

creative industries by size. Indeed, this indicator will help to understand if concentration of

creative industries is mainly driven or not by a diversity of firm size competitors in the creative

industry production process.

∑ [(

)

]

where s is the size of the industry which can be small, medium, large and very large (<15

employees, >=15, >=150 and >=1000 respectively). The number of firms has been extracted

from ORBIS database year (2009) due to lack of business demography data for certain

countries15

.

- The proxy for specialized suppliers have been approximated by the inverse of a Herfindhal index

inside the productive chain. This index indicated the relative degree of homogeneity in the

distribution of the employment among sectors in creative industries.

∑ [(

)

]

, where L refers to the employment in creative industries using NACE 3 digits from the census

(year 2001)16

.

- The knowledge and information spillovers generated inside the creative industries cluster can be

approximated by the Location Quotient (LQ) of employment in creative industries by LLS.

(

⁄ )

(

⁄ )

14 Note that only 66% of the creative industries provided by ORBIS contain information about the number of jobs. All information available was used to construct this indicator. 15 Note that only 66% of the creative industries provided by ORBIS contain information about the number of jobs. All information

available was used to construct this indicator. 16 Creative industries at NACE 4 digits have been converted to NACE 3 digits. Due to this conversion, several sectors at 4 digits

were not included in this indicator. The creative sectors included in this indicator are: 177 Manufacture of knitted and crocheted

articles ; 181 Manufacture of leather clothes ; 182 Manufacture of other wearing apparel and accessories ; 183 Dressing and dyeing

of fur; manufacture of articles of fur ; 193 Manufacture of footwear ; 221 Publishing ; 222 Printing and service activities related to

printing ; 223 Reproduction of recorded media ; 362 Manufacture of jewellery and related articles ; 363 Manufacture of musical

instruments ; 365 Manufacture of games and toys ; 525 Retail sale of second-hand goods in stores ; 722 Software consultancy and

supply ; 726 Other computer related activities ; 731 Research and experimental development on natural sciences and engineering ;

732 Research and experimental development on social sciences and humanities ; 742 Architectural and engineering activities and

related technical consultancy ; 744 Advertising ; 921 Motion picture and video activities ; 922 Radio and television activities ; 923

Other entertainment activities ; 925 Library, archives, museums and other cultural activities

22

, where L refers to the employment in creative industries using NACE 3 digits from the census

(year 2001)

- Related variety has been approximated by the number of overlaps of creative clusters by sector

, Creative clusters were obtained from Boix et al. (2011) where clusters where made for the

following creative sectors: architecture, cinema and music, trade, design, edition, photography,

engineering, fashion, heritage, performing arts, publishing, R&D, radio and television, software

and videogames.

b) Urbanization economies are advantages derived from the urban environment factors or characteristics

to all the economic activities that are located into it. These advantages can derive from two different

sources.

- Ohlin-Hoover’s potential size of the local market has been approximated by the population

density has been approximated by the employment density, where the denominator is the

urbanized land in km2.

, where A is the total land and P refers to total population, both values extracted from the census

(year 2001)

- Economic density (Hoover and Vernon 1959; Ciccone and Hall 1996, p. 54) has been

approximated by a proxy of the economic capacity of the population. This indicator is the result

of dividing the number of people with jobs over the total population of the LLS.

, where L refers to the total number of employees and P refers to total population, both values

extracted from the census (year 2001)

- Economic diversity of the productive structure of the LLSs (Chinitz 1961, pp. 281-282) have

been computed by using the inverse of a Hirschman-Herfinddhal index of employment diversity

at two diversity, also used by Paci and Usai (2005). Higher values indicate less diversity.

∑ [(

)

]

, where L refers to employment in 60 industrial sectors at NACE 2 digits obtained from census

(year 2001)

c) Specific creative forces: have been approximated by a set of indicators aimed at measuring the four

components previously developed.

- The cultural infrastructure indicator can be approximated by the Location Quotient (LQ) of

spaces of buildings related to cultural heritage and cultural services by LLS.

(

⁄ )

(

⁄ )

, where P refers to total population while I refers to all spaces or buildings related to cultural

heritage and cultural services. Cultural heritage sites have been obtained from the UNESCO

World heritage website (year 2013) while cultural services (museums, religious buildings,

23

cinemas, theatres, libraries, universities and schools) have been obtained from the Open Street

Map website (year 2013)

- Political power variable is constructed with a dummy variable that identifies the capital regions

(Comunidades Autónomas).

- Tolerance has been interpreted as the percentage of foreign workers to the total number of local

jobs (Florida, 2005)

, where FB is foreign born workers and Employment refers to the census year (2001)

- Talent is measured using the ration of human capital (Lucas, 1988 and Florida, 2005) over the

population above 24 years old, also used by Paci and Usai (2005).

, where H refers to the total population with at least tertiary education attainment while P24

refers to total population above 24 years old. Data is obtained from census (year 2001)

- Quality of life has been taken into consideration by computing a proxy of the accessibility to

places. This indicator is adjusted by the share of the length of streets in the LLS over the total

length of these streets in the country

, where streets refers the length of lively streets, footways, cycle ways, pedestrian and residential

provided by Open Street Map database (year 2013)

- The social capital indicator can be approximated by the Location Quotient (LQ) of public, semi-

public and private spaces that might facilitate social interaction of people in LLS.

(

⁄ )

(

⁄ )

, where P refers to total population and S refers to public, semi-public and private spaces in LLS.

These private services (coffee shops, pubs, nightclubs, restaurants, fast food, attraction, backery,

supermarket) and public or semi-public services (police stations, postbox and postoffices,

hospitals, pharmacy, fire stations, kindergarten, kiosks, parking, court houses, prisons, public

telephones, train and metro station and airports) have been obtained from the Open Street Map

website (year 2013)

- Patronage indicator was estimated based on the Location Quotient (LQ) of funding institutions in

LLS.

(

⁄ )

(

⁄ )

24

, where P refers to total population and S refers to funding institutions in LLS. These finance

institutions (bank offices and bank ATMs) have been obtained from the Open Street Map

website (year 2013)

8. Econometric analysis

The results of the empirical model are presented in table 3, 4 and 5, one for each of the three countries

analyzed in this study. A two-step strategy was followed to present the estimations. First, the analysis

started by estimating three separate regression groups in order to test separately the contribution of

different determinants to the clustering of creative employment (localization, urbanization and specific

creative conditions). Secondly, a general model, including all variables in partial regressions as estimated.

Each regression group presents the results of running a White Robust estimation and a Generalized Linear

Model (GLM) called Poisson log-linear model according to the configuration of dependent variable (log

and count of creative industries).

OLS regressions were computed and equations satisfied the normality assumption (Jarque-Bera test).

Heteroskedasticity (Breusch-Pagan test) was corrected by using an OLS White Robust estimation

(generic heteroskedasticity) in order to model the variance. Both results (OLS and OLS robust) were very

similar and suggest little effect of heteroskedasticity on the results. Apart from that, the condition number

and VIF did not show collinearity between explanatory variables. In the annex it is presented the

correlation between dependent and independent variables for Spain, Portugal and Italy.

Poisson log-linear regressions have been also used in this paper for the analysis of the determinants of

concentration of creative firms in LLSs. As it has been observed in the previous section, they are

considered to be a good model for count data models and model has been proved in the literature to be a

feasible alternative to the Negative binomial models.

8.1. Partial regressions

As expected, localization economies have a Localization economies have a good explanatory capacity of

creative industry clustering (OLS robust regressions present a R2 ranging from a 34% in Portugal to a

50% in Italy). However, regressions present similarities and differences in terms of the localization

determinants that affect clustering of creative firms by country. Indeed, competition, filiere and related

variety are positive correlated to the creative industries clustering in Spain and Italy. These results show

the importance of have a diversity of firm size in the creative industrial structure as well as a related

specialized industries or clusters of firms. Regarding the differences, on the one side, even if the

coefficients of an average size of creative industries are relatively small, regressions show that size of

creative industries affects negatively the creative industries cluster in Spain while this variable is

positively related to creative industries cluster in Portugal and Italy. In other words, creative industry

clusters in Spain are more related to small creative industries than in Portugal or Italy. On the other side,

location quotient of creative occupation is positively correlated in Spain (coefficient of 0.7) while this

coefficient is negatively correlated with creative industries concentration in Portugal (coefficient of -0.7).

Thus, these results suggest that Portugal creative industry clustering will be more determinate from the

concentration of related creative industries rather than a larger relative share of creative employment.

Urbanization economies regressions present an even larger explanatory capacity to explain creative

clustering (OLS robust regressions present a R2 ranging from 42% in Spain to 57% in Italy). Additionally

regressions show large and positive coefficients for the diversity variable. This indicates that the presence

of a diversified industrial structure is important to explain creative industry clustering. The other two

urbanization economies variables namely density and capacity even though they present positive

coefficients there are small. These results suggest that creative industrial clustering ware not much

affected by the size of the market and the economic capacity of people to buy new products.

Results show that the specific creative conditions have a larger explanatory capacity to explain creative

concentration than the previous externalities presented (especially in Spain and Portugal). Indeed, high

levels of education attainment, good access to patronage and tolerance presents to be the components that

more important in the three countries to explain the creative industry clustering, there are others that are

25

also positively correlate but their effect is smaller (access to political power and a good quality of life).

The regressions also present that a larger presence of cultural infrastructure do not seems to be a clear

determinant of the clustering of creative firms. This result is in line with previous studies (see for instance

Lazzeretti et al. 2009). Similarly, results suggest that presence of social life services such as bars,

restaurants, among others that can facilitate the interaction of individuals and as a result the share of

creative knowledge but also can contribute to the attraction of creative individuals do not seems to be a

clear determinant of creative industries clustering in Spain, Portugal and Italy

8.2. General regressions

In general terms it is worthy to say that these results are significantly consistent showing a strong capacity

of explanation of the findings. Indeed, OLS robust model show that the fit (R2) of these regressions are

around 70%. Additionally, similarly to what it was observed in the partial models, variables such as size,

access to cultural infrastructure and social life do not seems to be a clear determinant of creative

concentration in Spain, Portugal and Italy. On the contrary, competition, related variety, diversity and

education are presented to be the most important determinants of creative industry clustering in Spain,

Portugal and Italy.

Table 4. Spanish regression results partial and general regressions

Note 1: Dependent variable is the absolute number of creative industries by LLS and its logarithmic transformation (source ORBIS

2009)

Note 2: OLS robust estimation parameters should be interpret as direct elasticities while GLM parameters as change in the mean

response as x increases by 1 unit.

Note 3: Values in parenthesis are p-values. * p<0,1; ** p<0,05; *** p<0,01

Partial regressions General regressions

Localization economies Urbanisation economies Specific creative conditions

OLS GLM OLS GLM OLS GLM OLS GLM

(robust)

(family

poisson, llink

log) (robust)

(family

poisson, llink

log) (robust)

(family

poisson, llink

log) (robust)

(family

poisson, llink

log)

Dependent variable

ln(#creative

industries)

#creative

industries

ln(#creative

industries)

#creative

industries

ln(#creative

industries)

#creative

industries

ln(#creative

industries)

#creative

industries

Constant 0.673 *** 1.477 *** -1.554 *** -1.488 *** 0.277 ** 2.035 *** -1.495 *** -0.099 ***

(0.007) (0.000) (0.000) (0.000) (0.032) (0.000) (0.000) (0.05)

Size -0.02 *** -0.003 *** -0.019 *** -0.011 ***

(0.000) (0.000) (0.000) (0.000)

Competition 0.5 *** 0.433 *** 0.222 0.096 ***

(0.009) (0.000) (0.135) (0.000)

Filiere 0.191 *** 0.271 *** 0.033 0.066 ***

(0.000) (0.000) (0.11) (0.000)

Related variety 0.558 *** 0.348 *** 0.118 * 0.136 ***

(0.000) (0.000) (0.054) (0.000)

knowledge spillovers 0.746 *** 0.363 *** 0.698 *** 0.52 ***

(0.000) (0.000) (0.000) (0.000)

Density 0.000 *** 0.000 *** 0.000 * 0.000 ***

(0.002) (0.000) (0.089) (0.000)

Capacity 0.054 *** 0.066 *** 0.024 *** 0.02 ***

(0.000) (0.000) (0.001) (0.000)

Diversity 0.191 *** 0.237 *** 0.095 *** 0.067 ***

(0.000) (0.000) (0.000) (0.000)

Cultural infrastructure -3E-02 -5E-01 *** -0.028 -0.16 ***

(0.175) (0.000) (0.237) (0.000)

Political power 0.046 0.926 *** -0.223 0.136 ***

(0.892) (0.000) (0.404) (0.000)

Tolerance 0.046 *** 0.034 *** 0.041 *** 0.036 ***

(0.000) (0.000) (0.000) (0.000)

Quality of life 0.001 * 0.000 *** 0.000 0.000 ***

(0.05) (0.000) (0.145) (0.000)

Social life -0.175 *** -0.147 *** -0.156 *** -0.185 ***

(0.000) (0.000) (0.000) (0.000)

Education 0.254 *** 0.207 *** 0.158 *** 0.155 ***

(0.000) (0.000) (0.000) (0.000)

Patronage 0.07 ** 0.077 *** 0.057 * 0.076 ***

(0.033) (0.000) (0.08) (0.000)

R2 0.3529 0.4157 0.5405 0.6343

Log likelihood -31631.14 -78528.94 -31394.26 -15223.57

AIC 78.50 194.87 77.92 37.82

BIC 54329.97 148112.20 53869.59 21581.75

OBSERVATIONS 806 806 806 806 806 806 806 806

26

Table 5. Portuguese regression results partial and general regressions

Note 1: Dependent variable is the absolute number of creative industries by LLS and its logarithmic transformation (source ORBIS

2009)

Note 2: OLS robust estimation parameters should be interpret as direct elasticities while GLM parameters as change in the mean

response as x increases by 1 unit.

Note 3: Values in parenthesis are p-values. * p<0,1; ** p<0,05; *** p<0,01

Partial regressions General regressions

Localization economies Urbanisation economies Specific creative conditions

OLS GLM OLS GLM OLS GLM OLS GLM

(robust)

(family

poisson, llink

log) (robust)

(family

poisson, llink

log) (robust)

(family

poisson, llink

log) (robust)

(family

poisson, llink

log)

Dependent variable

ln(#creative

industries)

#creative

industries

ln(#creative

industries)

#creative

industries

ln(#creative

industries)

#creative

industries

ln(#creative

industries)

#creative

industries

Constant 1.144 5.32 *** -0.52 -0.244 *** 1.64 *** 3.67 *** -2.401 -0.667 ***

(0.604) (0.000) (0.666) (0.000) (0.003) (0.000) (0.145) (0.000)

Size 0.081 0.039 *** 0.013 0.029 ***

(0.303) (0.000) (0.736) (0.000)

Competition 1.579 0.026 3.102 ** 1.205 ***

(0.42) (0.677) (0.009) (0.000)

Filiere 0.13 -0.08 *** -0.009 -0.099 ***

(0.132) (0.000) (0.885) (0)

Related variety 0.505 *** 0.339 *** -0.202 0.015

(0.000) (0.000) (0.27) (0.116)

knowledge spillovers -0.693 * -0.174 *** -0.552 ** 0.003

(0.063) (0.000) (0.009) (0.879)

Density 0.000 ** 0.000 *** 0.000 0.000 ***

(0.013) (0.000) (0.634) (0.000)

Capacity -0.011 0.054 *** -0.078 ** 0.034 ***

(0.759) (0.000) (0.03) (0.000)

Diversity 0.417 0.222 *** 0.341 *** 0.194 ***

(0.000) (0.000) (0) (0.000)

Cultural infrastructure 0.000 0.000 *** 0.378 ** 0.094 ***

(0.274) (0.000) (0.014) (0.000)

Political power -0.611 -1.305 *** -0.473 0.321 ***

(0.188) (0.000) (0.438) (0.000)

Tolerance 0.003 -0.189 *** 0.299 *** 0.074 ***

(0.972) (0.000) -0.002 (0.000)

Quality of life 0.001 *** 0.001 *** 0.001 ** 0.000 ***

(0.003) (0.000) (0.044) (0.000)

Social life -0.686 *** -0.303 *** -0.555 *** -0.183 ***

(0.000) (0.000) (0.000) (0.000)

Education 0.373 *** 0.274 *** 0.198 * 0.062 ***

(0.000) (0.000) (0.055) (0.000)

Patronage 0.82 *** 0.436 *** 0.538 *** 0.096 ***

(0.000) (0.000) (0.000) (0.000)

R2 0.3442 0.5244 0.5435 0.7765

Log likelihood -11489.13 -8977.607331 -8509.901122 -3468.152545

AIC 276.99 216.4243 205.2506 83.95548

BIC 22158.44 17126.55 16208.81 6160.663

OBSERVATIONS 83 83 83 83 83 83 83 83

27

Table 6. Italian regression results partial and general regressions

Note 1: Dependent variable is the absolute number of creative industries by LLS and its logarithmic transformation (source ORBIS

2009)

Note 2: OLS robust estimation parameters should be interpret as direct elasticities while GLM parameters as change in the mean

response as x increases by 1 unit.

Note 3: Values in parenthesis are p-values. * p<0,1; ** p<0,05; *** p<0,01

9. Conclusions

The main purpose of this paper was to contribute to the broad topic of geographical concentration of

creative industries. Departing from theoretical and empirical literature on localization of creative

industries, this paper provides and explanatory approach of the location determinants of creative

industries in Spanish LLSs.

One of the main contributions of this paper is the use of micro-level data on creative industries to identify

the location of creative firms in Spanish, Portuguese and Italian LLSs. Findings show a high

concentration of creative industries around major capital cities, such as the LLSs of Madrid, Rome and

Lisbon. Additionally, micro data show a larger concentration of creative industries in the two largest

cities in each of the three studied countries.

Another contribution of this paper is the construction of an explanatory economic model (count

regression model) to investigate the distinct characteristics that bring a particular LLS to have more

creative industries. Additionally to the traditional approaches of externalities (Urbanization, Localization,

economies) this research also observes that more tailored creative forces are a crucial components that

Partial regressions General regressions

Localization economies Urbanisation economies Specific creative conditions

OLS GLM OLS GLM OLS GLM OLS GLM

(robust)

(family

poisson, llink

log) (robust)

(family

poisson, llink

log) (robust)

(family

poisson, llink

log) (robust)

(family

poisson, llink

log)

Dependent variable

ln(#creative

industries)

#creative

industries

ln(#creative

industries)

#creative

industries

ln(#creative

industries)

#creative

industries

ln(#creative

industries)

#creative

industries

Constant -0.567 * 2.054 *** -0.579 *** -0.062 *** 0.856 *** 2.628 *** -1.949 *** -0.297 ***

(0.023) (0.000) (0.000) (0.002) (0.000) (0.000) (0.000) (0.000)

Size 0.002 0.001 *** -0.003 -0.013 ***

(0.65) (0.000) (0.264) (0.000)

Competition 1.235 *** 0.596 *** 0.767 *** 0.515 ***

(0.000) (0.000) (0.000) (0.000)

Filiere 0.472 *** 0.246 *** 0.124 *** 0.071 ***

(0.000) (0.000) (0.000) (0.000)

Related variety 0.394 *** 0.322 *** -0.07 0.042 ***

(0.000) (0.000) (0.202) (0.000)

knowledge spillovers -0.036 0.004 0.047 0.195 ***

(0.581) (0.431) (0.341) (0.000)

Density 0.000 *** 0.000 *** 0.000 *** 0.000 ***

(0.000) (0.000) (0.000) (0.000)

Capacity 0.042 *** 0.079 *** 0.055 *** 0.047 ***

(0.000) (0.000) (0.000) (0.000)

Diversity 0.161 *** 0.106 *** 0.097 *** 0.104 ***

(0.000) (0.000) (0.000) (0.000)

Cultural infrastructure -0.103 ** -0.248 *** -0.126 *** -0.208 ***

(0.012) (0.000) (0.000) (0.000)

Political power 0.181 1.143 *** -0.061 0.08 ***

(0.514) (0.000) (0.746) (0.000)

Tolerance 0.223 *** 0.255 *** -0.151 *** -0.191 ***

(0.000) (0.000) (0.000) (0.000)

Quality of life 0.002 *** 0.000 *** 0.001 *** 0.000 ***

(0.007) (0.000) (0.001) (0.000)

Social life -0.13 *** -0.292 *** -0.065 ** -0.066 ***

(0.001) (0.000) (0.012) (0.000)

Education 0.252 *** 0.175 *** 0.126 *** 0.106 ***

(0.000) (0.000) (0.000) (0.000)

Patronage 0.111 ** 0.223 *** -0.003 0.026 ***

(0.012) (0.000) (0.912) (0.000)

R2 0.495 0.5667 0.5017 0.7169

Log likelihood -26718.79 -50402.04 -28132.11 -12564.19

AIC 77.91 146.96 82.04 36.68

BIC 45583.68 92937.12 48423.39 17339.78

OBSERVATIONS 686 686 686 686 686 686 686 686

28

determine the creative industry concentration in the three countries under study. On the one hand,

regarding the general determinants, it has been observed how competition, related variety and diversity

variables determine the creative industry clustering. On the other hand, specific creative forces such high

levels of education attainment of the population, good access to patronage and the existence of a tolerant

society seem to offer also a powerful explanation of creative industries clustering in Spain, Italy and

Portugal.

Policy implications of this research based on the findings are of significant importance for regional and

local policy makers in the three countries under study. It is important to understand that, the new EU

initiative called Europe 2020 strategy (smart, green and inclusive growth) aims at boosting the growth of

national economies and jobs by supporting a diversified, strong and competitive industrial base in Europe.

At the same time, several studies have recently provided sound evidence on the contribution of creative

industries to local and regional development in EU (De Miguel et al. (2012); Rausell et al. 2012; the

European Competitive Report 2010). Indeed, the European Competitive Report (2010) underlines that

those creative industries can be considered important innovators as well as important drivers of

innovation to other sectors of the economy. It is for this reason that regional policy makers need sound

evidence on the factors that might attract creative industries. Indicators of such factors can indeed be

integrated into concrete policy frameworks.

29

Annex: Correlations table between the dependent and the independent variables

Spain

Portugal

Creative

industries)

ln(creative

industries)

Knowledge

spillovers Filiere

Competitio

n Size Density Capacity Diversity

Related

Variety

Cultural

infrastructu

re

Political

power Tolerance Social life Education Patronage

Quality of

life

Creative industries) 1

ln(creative industries) 0.3505 1

Knowledge spillovers 0.0711 0.2388 1

Filiere 0.1269 0.309 -0.1197 1

Competition 0.0604 0.1956 0.425 0.0063 1

Size 0.0776 0.1086 0.7117 -0.0584 0.3447 1

Density 0.3429 0.3484 0.0501 0.2035 0.0616 0.0226 1

Capacity 0.1041 0.3275 0.0781 0.1606 0.1244 0.08 0.1094 1

Diversity 0.1961 0.5528 0.0104 0.1921 0.1101 0.0477 0.1457 0.2162 1

Related Variety 0.6997 0.4863 0.1374 0.1755 0.0809 0.1312 0.4304 0.1548 0.3182 1

Cultural infrastructure -0.0077 -0.0752 -0.0642 0.0141 0.0099 -0.0269 -0.0896 0.0843 0.0716 -0.0427 1

Political power 0.4247 0.3521 0.0533 0.1968 0.0623 0.0327 0.4301 0.0877 0.2133 0.5096 -0.029 1

Tolerance 0.0367 0.1679 0.0072 0.2331 0.011 0.0179 0.1384 0.3532 -0.0481 0.0135 0.0087 0.0103 1

Social life -0.0237 -0.0659 -0.0032 0.1616 0.0007 0.0189 -0.0527 0.1605 0.0243 -0.0477 0.4813 -0.0244 0.2081 1

Education 0.3002 0.6721 0.0726 0.4656 0.0885 0.0649 0.3086 0.3268 0.6208 0.4044 0.0832 0.4047 0.1043 0.1267 1

Patronage 0.0084 0.049 0.0132 0.1447 0.0775 -0.0005 -0.0252 0.0879 0.0862 0.0014 0.2682 0.0177 0.0947 0.6964 0.1647 1

Quality of life 0.9272 0.4942 0.1016 0.2095 0.0643 0.1043 0.3058 0.1347 0.3086 0.7619 -0.0154 0.4738 0.0525 -0.011 0.4417 0.0405 1

Creative

industries)

ln(creative

industries)

Knowledge

spillovers Filiere

Competitio

n Size Density Capacity Diversity

Related

Variety

Cultural

infrastructu

re

Political

power Tolerance Social life Education Patronage

Quality of

life

Creative industries) 1

ln(creative industries) 0.5159 1

Knowledge spillovers 0.0888 -0.0073 1

Filiere 0.0843 0.1511 -0.3491 1

Competition 0.0975 0.2662 0.4533 -0.1636 1

Size 0.138 0.2876 0.4168 -0.2023 0.7926 1

Density 0.7074 0.4883 0.0361 0.084 0.095 0.0677 1

Capacity 0.3438 0.3612 -0.0377 0.2723 -0.1058 -0.1094 0.499 1

Diversity 0.4251 0.7082 0.0143 0.1043 0.0127 0.0765 0.5078 0.476 1

Related Variety 0.9685 0.4841 0.1312 0.0365 0.1068 0.1362 0.7574 0.3237 0.4028 1

Cultural infrastructure -0.0566 -0.0807 0.1707 -0.1227 0.0961 0.0084 0.0016 0.2471 -0.1329 -0.0728 1

Political power 0.3378 0.2608 -0.1086 0.3618 0.0228 -0.0409 0.357 0.2606 0.2188 0.2688 -0.0361 1

Tolerance 0.1568 0.1753 0.0107 0.3001 -0.0705 -0.0535 0.103 0.4741 -0.032 0.1338 0.0671 0.1086 1

Social life -0.0312 -0.1569 0.3178 -0.1277 0.2519 0.0634 0.0219 0.1593 -0.1789 -0.0239 0.7876 -0.0185 0.139 1

Education 0.5346 0.5717 -0.09 0.4567 -0.1142 -0.0738 0.5793 0.6125 0.6149 0.5134 -0.1111 0.4921 0.3219 -0.1168 1

Patronage 0.0075 0.0359 0.2899 -0.1379 0.2876 0.0859 0.0535 0.1796 -0.0751 0.0085 0.6894 0.0115 0.1476 0.8887 -0.0715 1

Quality of life 0.9779 0.5747 0.0741 0.1397 0.064 0.113 0.7185 0.4004 0.4891 0.9397 -0.0459 0.3899 0.1968 -0.0235 0.6004 0.0227 1

30

Italy

Creative

industries)

ln(creative

industries)

Knowledge

spillovers Filiere

Competitio

n Size Density Capacity Diversity

Related

Variety

Cultural

infrastructu

re

Political

power Tolerance Social life Education Patronage

Quality of

life

Creative industries) 1

ln(creative industries) 0.3952 1

Knowledge spillovers 0.0009 -0.0013 1

Filiere 0.3009 0.588 -0.1605 1

Competition 0.073 0.3281 0.3354 0.1 1

Size 0.0424 0.0988 0.7145 -0.024 0.3793 1

Density 0.4155 0.494 0.0136 0.3875 0.0828 0.0196 1

Capacity 0.1947 0.5225 -0.0347 0.3653 0.3886 0.1002 0.1254 1

Diversity 0.1878 0.6189 -0.1338 0.5256 0.2293 0.0869 0.2208 0.5373 1

Related Variety 0.7117 0.5043 0.0029 0.3896 0.1246 0.0457 0.6534 0.2693 0.2759 1

Cultural infrastructure -0.0532 -0.1483 -0.0532 -0.0352 0.0724 -0.0049 -0.1732 0.3155 0.0643 -0.0801 1

Political power 0.4501 0.3426 -0.0165 0.396 0.0533 0.0226 0.317 0.1918 0.1694 0.5009 -0.0195 1

Tolerance 0.1184 0.2485 0.0232 0.2371 0.3129 0.1032 0.0326 0.7027 0.362 0.1405 0.2838 0.057 1

Social life -0.0367 -0.1335 -0.0395 -0.0583 0.1378 0.0054 -0.1407 0.3462 -0.0744 -0.0632 0.6455 0.0236 0.2747 1

Education 0.3112 0.5402 -0.0448 0.5238 0.0122 -0.021 0.3667 0.1967 0.3196 0.3383 -0.1498 0.3807 0.0095 -0.1373 1

Patronage 0.0058 0.0285 -0.075 0.0565 0.1517 -0.0127 -0.0515 0.3691 0.131 0.0008 0.5191 0.058 0.2534 0.6855 -0.025 1

Quality of life 0.8987 0.5754 -0.0308 0.4426 0.1318 0.0614 0.4302 0.3639 0.3794 0.7596 -0.0431 0.4613 0.2122 -0.0109 0.3874 0.0629 1

31

Bibliography

- Abramobsky, L; Griffith, R; Macartney, G. and Miller, H. (2008) “The location of innovative activity

in Europe”, The Institute for fiscal studies, WP08/10. http://discovery.ucl.ac.uk/14738/1/14738.pdf

- Acs, Z.J.; Varga A. (2002) Geography, endogenous growth, and innovation. International Regional

Science Review, 25 , 1

- Arauzo-Carod, J.M (2007) “Industrial location at a local level: Comments on the territorial level of

the analysis” Tjdschrift voor Economische en Sociale Geografie – 2008, Vol. 99, No, 2, pp. 193-208

- Arauzo-Carod, J-M.; Liviano-Solis, D. and Manjón-Antolín, M. (2010) “Empirical studies in

Industrial location: An assessment of their methods and results” Journal of Regional Science, vol 50,

no 3, pp. 685-711

- Arora, A.;Florida, R.; Gates, G. J. and Kamlet, M. (2000) " Human Capital, Quality of Place, and

Location". Working Paper, H.J. Heinz III School of Public Policy, Pittsburg: Carnegie Mellon

University

- Asheim. B. and Vang, J. (2005) "Talents and Innovative Regions: exploring the Importance of Face-

to-Face Communication and Buzz". Circle paper no. 2005/14

- Asheim, B.; Lars, C. and Vang,J. (2005) "FAce-to-FAce. Buzz and Knowledge Base: Socio-spatial

implications for learning and innovation policy". Circle paper no. 2005/14

- Asheim, B.; Boschma, R. and Cooke, P. (2007) "Constructing regional advantage: Platform policies

based on related variety and differentiated knowledge bases". Papers in Evolutionary Economic

Geography. Utrecht University. Urban & Regional research centre Ultrecht.

- Assmo, P. (2010) "Creative industries and regional economic development - the example of public

supported regional film centres in Sweden" in Charlie Karlsson,

- Audretsch, D. B. and Feldman, M. P. (1996) R&D Spillovers and the Geography of Innovation and

Production, The American Economic Review. vol 86 (3), 630-640

- Audretsch, D. B. (1998) "Agglomeration and the location of innovative activity", Oxford Review of

Economic Policy, Volume 14, No. 2.

- Audretch, D. B. (2003) “Innovation and Spatial Externalities,” International Regional Science

Review, 26(2), 167-174.

- Baumount, C and Boiteux-Orain, C. (2005) "Secteur culturel, metropolisation et centralites urbaines

le cas de l'Île-de-France". LEG, Laboratoire d'Economie et de Gestion, CNRS UMR 5118, Université

de Bourgogne in its series LEG - Document de travail - Economie with number 2005-06.

- Becattini, G. (2005) La oruga y la mariposa. Un caso ejemplar de desarrollo en la Italia de los

Distritos Industriales: Prato (1954-1993). Valladolid. Universidad de Valladolid, Secretariado de

Publicaciones e Intercambio Editorial.

- Berg, S-H and R. Hassink (2013) “Creative industries from an evolutionary perspective: A critical

literature review”, Papers in Evolutionary Economic Geography (PEEG).

- Bergman et al. (1991)

- Boix, Rafael y Vittorio Galletto (2006a): "Sistemas locales de trabajo y distritos industriales

marshallianos en España", Economía Industrial, nº 359, p.165-184. ISSN: 0422-2784.

- Rafael y Vittorio Galletto (2006 b): “El nuevo mapa de los distritos industriales de España y su

comparación con Italia y el Reino Unido”, Working Paper 06.04 Departament d'Economia Aplicada,

Universitat Autònoma de Barcelona // Comunicación presentada a la XXXII Reunión de Estudios

Regionales, Ourense 16-28 de Noviembre de 2006.

- Boix et al. (2011)

- Boix, R.; Lazzeretti,L.; Cappone, F. and Sanchez-Serra, D. (2012) " The geography of creative

industries in Europe: Comparing France, Great Britain, Italy and Spain" in Creative Industries and

32

Innovation in Europe. Concepts, Measures and Comparative Case Studies. Edited by Luciana

Lazzeretti. Routledge. Series: REgions and Cities

- Boix, R.; J. L. Hervás-Oliver, and B. De Miguel-Molina (2013) “I want creative neighbours”. Do

creative service industries spillovers cross regional boundaries? Working Papers in Applied

Economics WPAE-2013-15, Universitat de Valencia

- Bourdieu, P. (1960) Le capital social. Actes de la Recherche en Sciences Sociales 31, pp. 2-3

- Boschma, R. and S. Iammarino (2009) “Related Variety, Trade Linkages, and Regional Growth in

Italy”, Economic Geography, Vol. 85, Issue 3, pp. 289-311

- Boschma, R. and K. Frenken, (2009) "Some Notes on Institutions in Evolutionary Economic

Geography," Economic Geography, Clark University, vol. 85(2), pages 151-158, 04.

- Boschma, R. & K. Frenken, (2010) "The emerging empirics of evolutionary economic geography,"

Eindhoven Center for Innovation Studies (ECIS) working paper series 10-10, Eindhoven Center for

Innovation Studies (ECIS), revised Nov 2010.

- Bounken, R. (2009) "Creativity in Cross-Cultural Innovation Teams: Diversity and Its Implications

for Leadership", in Meusburger, P.; Funke, J., and Wunder, E. (eds.) Milieus of Creativity. An

interdisciplinary Approach to Spatiality of Creativity.

- Brandellero, A.; Calenge, R.; Halbert, L.; Waellishc, Y. (2006) Clusters, milieux d’innovation et

industries culturelles en Île-de-France. , Final Report, Université Paris-Est.

- Camagni (2008) “Towards a Concept of Territorial Capital” in : R. Capello, R. Camagni, B.

Chizzolini, U. Fratesi , Modelling regional scenarios for the enlarged Europe, Springer, Berlin, 2008

- Camagni, R. (2005) Economía Urbana, Ed. Antoni Bosch, Barcelona.

- Capello, R. (2004) Economia Regionale. Ed. Mulino, Bologna.

- Carlton, D.W., (1983) "The location and employment choices of new firms: an econometric model

with discrete and continuous endogeneous variables". Review of Economics and Statistics, 65: 54-70.

- Caves, R (2000) Creative Industries: Contracts Between Art and Commerce. Cambridge: Harvard

University Press, 2000

- Chinitz, B. (1961) “Contrasts in Agglomeration: New York and Pittsburgh”. The American

Economic Review, 51 (2), pp. 279-289

- Ciccone, A. and R. E. Hall (1996) “Productivity and the Density of Economic Activity”. The

American Economic Review, 86 (1), pp. 54-70.

- Clifton, N. and Cook, P. (2007) "The 'creative class' in the UK: An initial analysis". Regional

Industrial Research Report 46

- Cooke P. and Lazzeretti L. (eds), (2008), Creative cities, cultural clusters and local economic

development, Edward Elgar, Cheltenham, UK.

- Cooke, P., De Laurentis, C., Tödtling, F. and Trippl, M. (2007) Regional Knowledge Economies,

Cheltenham: Edward Elgar.

- DCMS (2001), Creative Industries Mapping Document 2001 (2 ed.), London, UK: Department of

Culture, Media and Sport, retrieved 2007-05-26

- De Miguel B., Hervás J.L., Boix R. and De Miguel M. (2012): "The importance of creative regions in

Europe: exploring industry agglomeration and the growth of European Regions" in Bas T.G. and

Zhao J. (eds), Comparing HIgh Technology Firms in Developed and Developing Countries: Cluster

Growth Initiatives, p.65-79. IGI Global Publisher, Hershey. ISBN: 9781466616462.

- De Propris, L., Chapain, C., Cooke, P., MacNeill S., and Mateos-Garcia, J., (2009) , Thegeography

and creativity , Report 27, NESTA, UK.

- DeVol, R.C. (1999) America’s High-Tech Economy Growth, Development, and Risks for

Metropolitan Areas. Milken Institute

- Driffield, N. and Menghinello, S. (2010) “Location patterns and determinants of MNT knowledge

intensive activities” Paper submitted to participants of the OECD workshop “Regional Innovation

Networks”, 7 june 2010.

33

- EIS (2008) Design, creativity and innovation. A scoreboard approach

- Ellison and Glaeser (1997, p.892) Ellison, G and E.L. Glaeser (1994/1997) “Geographic

Concentration in U.S. Manufacturing Industries: A Dartboard Approach”, The Journal of Political

Economy, 105 (5), pp.889-927.

- EU (2006) The Economic of culture in Europe. Available: http://ec.europa.eu/culture/key-

documents/doc873_en.htm

- Feldman 2000, p. 378-379; Feldman, M. P. (2000). “Location and Innovation: The New Economic

Geography of Innovation, Spillovers, and Agglomeration.” in G. Clark, M. Feldman and M. Gertler,

editors. The Oxford Handbook of Economic Geography. Oxford: Oxford University Press. pp. 373-

394.

- Flew, T. (2002) "Beyond ad hocery: defining the creative industries", paper presented to cultural

Sites, Cultural Theory, Cultural Policy, the second international conference on cultural policy

research, Te Papa, EWellington, New Zealand, 23-26 January 2002. Available:

www.library.auckland.ac.nz/subjects/bus/.../creative_industries.pdf (accessed 2012)

- Florida R, (2002) The Rise of the Creative Class, New York: Basic Books

- Florida R, (2005b) The flight of the creative class. New York, HarperCollins

- Florida 2008; Florida, R. (2008) Who's Your City?: How the Creative Economy Is Making Where

You Live the Most Important Decision of Your Life, Basic Books, Random House, ISBN 0-465-

00352-4

- Florida, R.; Mellander, C. and Qian, H. (2008) " Creative China? The University, Human Capital and

the Creative Class in Chinese Regional Development"

- Florida, R. (2005) Cities and the creative class, New York: Routledge

- Fritsch M. and Stützer M, 2008, "The Geography of Creative People in Germany," MPRA Paper

21965, University Library of Munich, Germany

- Galloway, S. and Dunlop, S. (2006) "Deconstructing the Concept of 'Creative Industries'" in Cultural

Industries: The British Experience in International Perspective, 2006, Humboldt University Berlin,

Edoc-Server, Available: http://edoc.hu-berlin.de. ISBN 978-3-86004-203-8

- Garnham, N. (1987) Concepts of culture - public policies and the cultural industries. Originally

published as a discussion paper by the Greater London Council in 1983, reprinted in Cultural

Studies, vol.1(1)

- Gertler, M. (2003) " Tacit knowledge and the economic geography of context, or The undefinable

tacitness of being (there)", Journal of Economic Geography, Volume 3, Issue 1, pp. 75-99

- Glaeser, E., Kallal, H., Scheikman, J. i Shleifer, A. (1992) Growth in Cities. Journal of Political

Economy, 100, pp. 1126-1152.

- Glaeser E W, 2000, "The new economics of urban and regional growth", in The Oxford Handbook of

Economic Geography, ed. Clark G., Gertler M. and Feldman M., Oxford University Press, Oxford,

83-98

- Globerman, S. (2001) “The location of higher value-added activities”. Industry Canada Research

Publications Program. Occasional Paper 27.

- Gordon, I. R., & McCann, P. (2000). Industrial clusters, complexes, agglomeration and/or social

networks? Urban Studies , 7 , 513 – 532.

- Gowers, A. (2006) Gowers Review of Intellectual Property. HM Treasury.

- Guimaraes, P., Figueiredo, O. and Woodward, D. (2003) “A Tractable Approach to the Firm

Location Decision Problem”, Review of Economics and Statistics, 85 (1), pp. 201-204

- Hansen, Stephen and Justin Van Fleet (2003) Traditional Knowledge and Intellectual Property: A

Handbook on Issues and Options for Traditional Knowledge Holders in Protecting their Intellectual

34

Property and Maintaining Biological Diversity. American Association for the Advancement of

Science (AAAS) Science and Human Rights Program, Washington, DC

- Hanson, G.H., (2000) “Firms, Workers, and the Geographic Concentration of Economic Activity,” in

Gordon Clark, Meric Gertler, and Maryann Feldman, eds.,

- Hoover (1937/1971) Location Theory and the Shoe and Leather Industries. Harvard University Press.

Cambridge, Massachusetts.

- Hoover, E.M. and Vernon, R. (1959) Anatomy of a Metropolis. Harvard University Press,

Cambridge, M.A.

- Howkins, J. (2001) The creative economy how people make money from ideas, England, The Penguin

press

- Jacobs, J. (1961/1971) Muerte y vida de las grandes ciudades. Ediciones Península.

- Jacobs, J. (1969) La economía de las ciudades. Ediciones Península.

- Karlsson, (2008) Handbook of Research On Cluster Theory, Edward Elgar Publishing

- KEA (2006) Economy of culture in Europe. Available: http://www.keanet.eu/en/ecoculturepage.html

- Kennedy, P. (2003) A guide to econometrics. Fifth edtion. Blackwell Publishing.

- Kourtit, K.; J. Möhlmann; P. Nijkamp and J. Rouwendal (2013) “The Spatial Distibution of Creative

Industries and Cultural Heritage in the Netherlands”, Tinbergen institute.

- Landry C. (2000) The creative city. A toolkit for urban innovators , Sterling, London.

- Lazzeretti, L., R. Boix and F. Capone (2008) Do creative Industries Cluster? Mapping Creative Local

Production Systems in Italy and Spain. Industry and Innovation, Vol. 15, No. 5, pp. 549-567

- Lazzeretti, L. R. Boix and F. Capone (2009) “Why do creative industries cluster? An analysis of the

determinants of clustering of creative industries”, IERMB Working Paper in Economics, n. 09.02,

April 2009

- Lazzeretti, L., R. Boix and F. Capone (2010) “On the concentration of creative industries in

specialized local production systems in Italy and Spain: Patterns and determinants”, EUNIP

Intenational Conference 2010, Reus, 9th

-11th

June2010

- Lazzeretti, L., Boix, R. and F. Capone (2011): “Do creative industries cluster? Mapping creative

local production systems in Italy and Spain”, en F. Belussi and S. Sedita Managing situated creativity

in cultural industries. Routledge. ISBN: 978-0-415-61355-2

- Lazzeretti L., Boix R. and Capone F. (2012): “Why do creative industries cluster?”, in Lazzeretti L.

(ed) Creative industries and innovation in Europe: Concepts, Measures and Comparative Case

Studies,, p.45-64. Routledge. ISBN 0415677408 / 9780415677400

- Lazzeroni M, (2010), High-tech activities, systems innovativeness and geographical concentration.

Insights into technical districts in Italy, European Urban and Regional Studies, 17 (1): 45-63

- Le Blanc, G. (2000) " Regional Specialization, Local Externalities and Clustering in Information

Technology Industries". CERNA, Centre d’économie industrielle Ecole Nationale Supérieure des

Mines de Paris. Presented in XII Villa Mondragone International Economic Seminar “Knowledge

Economy, Information Technologies and Growth” University Tor Vergata, Roma 27 juin 2000

- Lorenzen, M. and Frederiksen, L. (2008) "Why do cultural industries cluster?: Localization,

Urbanisation, Products and Projects". Published in Creative Cities, Cultural Clusters, and Local

Economic Development, Cook, P and Rossella Lazzeretti (2008) (eds.), Chelternham: Edward Elgar,

155-179

- Lucas, R. E. (1988) On the mechanics of economic development. Elsevier Science Publishers B.

V.(North-Holland).

- Malecki, E.J. (1994) Entrepreneurship in regional and local development. International Regional

Science Review 16: 119-153

35

- Malmberg, A. and P. Maskell, (2001) “The elusive concept of localization economies: towars a

knowledge-based theory of spatial clustering”, Environment and Planning A 34(3) 429 – 449

- Manjón-Antolin, M.C. and Arauzo-Carod, J.M. (2006) “Locations and Relocations: Modelling,

Determinants, and Interrelations” Working Paper 6-2007, Department of Economics URV, 2007

- Markusen, A., P. Hall and Glasmeier, A. (1986) High Tech America: The What, How, When and

Why of the Sunrise Industries. Boston, London, Allen and Unwin.

- Marshall, A. (1890/2009) Principles of economics: Unabridged eight edition, New York, Cosimo

publications. http://books.google.fr/books?hl=fr&lr=&id=IyZJXc_v-

TAC&oi=fnd&pg=PR30&dq=marshall+1890+Principles+of+Economic&ots=aVX0A-

Mu6s&sig=ESXPezk61hc8vWGmVMYslwOzVN8#v=onepage&q&f=false

- Maskell P. and Lorenzen M., (2004), ‘The cluster as market organization’, in Urban Studies, vol. 41,

no. 5-6, pp. 991-1009.

- Montuschi, L. (2001a) “Datos, información i conocimiento. De la sociedad de la información a la

sociedad del conocimiento”. UCEMA Documento de trabajo Nº192.

- Murphy, E. and Redmond, D. (2009) 'The role of 'hard' and 'soft' factors for accomodating creative

knowledge: Insights from Dublin's 'creative class'' In: Kirby, P. and Carmody, P (eds). The Legacy of

Ireland's Economic Expansion: Geographies of the Celtic Tiger. London and New York: Routledge.

- NESTA (2008) Beyond the creative industries. Mapping the creative economy in the United

Kingdom.

- Nonaka, I. and Takeuchi, H. (1995) The Knowledge-Creating Company: How Japanese companies

create the dynamics of innovation. Oxford University Press, Oxford.

- NORDEN (2006) Creative Directions – a Nordic framework for supporting the creative industries.

Editors Dominic Power and Johan Jansson, SIK The Swedish Institute for Food and Biotechnology.

- O'Connor , J. (1999) "The Definition of 'Cultural Industries'", online article at

www.pedrobendassolli.com/pesquisa/icc1.pdf

- O'Connor J. (2000) "The definition of 'Cultural Industries", The European Journal of Arts Education

Vol. 2 No. 3, February , pp. 15-27 Oct.

- O'Connor, J. (2007) The cultural and creative industries: a review of the literature. Creative

Partnership Series. Arts Council, England

- O'Connor, J. and Xin, G. (2006) "A new modernity?: The arrival of 'creative industries' in China",

International Journal of Cultural Studies; 9; 271; pp. 271-282

- OECD (1996) The knowledge-based economy. General distribution, OCDE/GD(96)102. Paris.

- OECD (2005) Culture and local development. Paris, France

- OECD (2008) The globalisation competition for talent. Paris, France

- Ohlin, B. (1933) Interregional and international Trade. Harvard Univerity Press. Cambridge.

- Overman, H.G. and Puga, D. (2010) “Labour Pooling as a Source of Agglomeration. An Empirical

Investigation” in Agglomeration Economies. Edited by Edward L. Glaeser. The University of

Chicago Press. ISBN: 0-226-29789-6

- Paci, R. and S.Usai (2005) “Agglomeration economies and growth in Italian local labour systems

1991-2001”. European Regional Science Association, ERSA conference papers 01/2005

http://dev3.cepr.org/meets/wkcn/2/2357/papers/paci.pdf

- Polanyi, M. (1966/1997) The tacit dimension. Garden City, New York.

- Porter, M.E. (1990) "The Competitive Advantage of Nations", Free Press, New York, 1990

- Porter, M. (2000) "Locations, Clusters, Company Strategy" in Par Gordon L. Clark,Maryann P.

Feldman,Meric S. Gertler (2000) The Oxford handbook of economic geography. Oxford University

Press

36

- Potts, J, D. (2009) “Why creative industries matter to economic evolution” Economics of Innovation

and New Technology, 18 (7-8). pp. 663-673

- Power and Nielssen (2010) Priority Sector Report_Creative and Cultural Industries. Deliverable D9-1

- Prager, J-C. and Thisse, J-F. (2009) " Les enjeux géographiques du développement économique"

Agence Française de Développement Département de la Recherche

- Pratt, A. (1997) "The cultural industries production system: a case study of employment change in

Britain, 1984-91" Environment and Planning, vol. 29, 11, pp. 1953-74

- Pratt, A. (2004) "Mapping the cultural industries: regionalization; the example of south-east

England"in Power, D. and Scott, A. J. Cultural industries and the production of culture, London,

UK; Routledge, 2004, pp.19-36

- Pro inno (2007) Benchmarking national and regional support services for SMEs in intellectual and

industrial property. KMU FORSCHUNG AUSTRIA Austrian Institute for SME Research

- Raussell P (coord.), Abeledo R, Boix R, De Miguel B, Hervás JL, Marcos F, Pérez D, Pérez P,

Sorribes J, Vila L (2012): Culture as a factor of economic and social innovation (English) / La cultura

como factor de innovación económica y social (Spanish). SOSTENUTO final report, University of

Valencia - IIDL, Valencia.

- Reardon, M. (2009) “Tolerance, Amenities and Economic Systems: Regional Development of the

Knowledge Economy”, Nordregio Electronic Working Paper 2009: 2

- Saxenian, A. (1994) Regional adventaje: culture and competition in Silicon Valley and Route 128.

Harvard Univesity Press.

- Scott, A. J. (2000a) "The cultural economy of Paris". International Journal of Urban and Regional

Research. Volume 24, Issue

- Scott, A, J. (2005) On Hollywood. The Place, the Industry, Princeton: Princeton

- Scott, A. J. (2006) "Creative cities: conceptual issues and policy questions". Journal of Urban

Affairs. Volume 28, Number 1, pp. 1-17

- Selada, C., I. Vihena da Cunha and E. Tomás (2010) “Creative Clusters in Low Density Urban

Areas: A Case Study Approach”. Presented in ERSA conference 2010.

- Sheih and Mai (1997).

- Sivitanidou, R. (1999) "The Location of Knowledge-Based Activities: The Case of Computer

Software"

- Suarez-Villa, L. and Walrod, W. (1997) “Operational Strategy, R&D, and Intra-metropolitan

Clustering in a Polycentric Structure: The Advanced Electronics Industries of the Los Angeles

Basin.” Urban Studies, vol. 34, no. 9, 1997, pp. 1343-1380.

- Towse, R. (2003) "Cultural industries" A Handbook of Cultural Economics, Ed. Ruth Towse.

Chelternham: Edward Elgar, 2003, pp. 170-182

- Tschang, T. F. and Vang, J. (2008) "Explaining the spatial organization of crative industries: the case

of the U.S. Videogames industry". Paper to be presented at the 25th Celebration Conference 2008 on

Entrepreneurship and innovation - organizations, institutions, systems and regions. Copenhagen,

CBS, Denmark, June 17-20, 2008

- Turok, I. (2003) "Cities, Clusters and Creative Industries: The Case of Film and Television in

Scotland". European Planning Studies, vol. 11 (5), 549-565

- UK Department of Culture, Media and Sport (2001, p.5

- UNESCO (2005) Convention on the Protection and Promotion of the Diversity of Cultural

Expressions 2005. Available : http://portal.unesco.org/en/ev.php-

URL_ID=31038&URL_DO=DO_TOPIC&URL_SECTION=201.html

- UNCTAD (2008) Creative Economy. The challenges of Assessing the Creative Economy: Towards

Informed Policy-making. United Nations. ISBN 978-0-9816619-0-2

37

- UNCTAD (2010) Creative Economy. Report 2010. A Feasible Development Option. United Nations.

ISBN 978-0-9816619-0-2

- Van Oort, F.; Weterings, A. and Verlinde, H. (2003) "Residential amenities of knowledge workers

and the location of ICT-firms in the Netherlands". Tijdschrift voor Economische en Sociale

Geografie – 2003, Vol. 94, No. 4, pp. 516–523.

- Vang, J. (2005) " Rethinking the Spatial Organization of Creative Industries", CIRCLE Electronic

Working Paper Series, Paper no. 2005/10

- Vang, J. (2007) "The spatial organization of the news industry: questioning assumptions about

knowledge externalities for clustering of creative industries" Innovation: Management, Policy and

Practice Vol. 9 (1), pp. 14-27

- Viladecans, E. (2002) "The growth of cities: Does agglomeration matter?" Document de treball

2002/3. IEB Institut d'Economia de Barcelona. Centre de Recerca en Federalisme Fiscal i Economia

Regional.

- von Hippel, E. (1994). "Sticky information and the locus of problem solving: Implications for

Innovation. Published in Management Science 40, no. 4, pp. 429-439

- Weber (1929/1968) Theory of the location of industries. The University of Chicago Press. Chicago

and London.