Relatedness put in place On the effects of proximity on firm performance

Lisa Östbring

Department of Geography and Economic History

Umeå University, Sweden

GERUM 2015:1

GERUM – Kulturgeografi 2015:1

Institutionen för geografi och ekonomisk historia

Department of Geography and Economic History

Umeå University

SE – 901 87 Umeå

Sverige/Sweden

Tel: +46 90 786 5000

Fax: +46 90 786 6359

http://www.geoekhist.umu.se

E-mail: lisa.ostbring@umu.se

This work is protected by the Swedish Copyright Legislation (Act 1960:729)

ISBN: 978-91-7601-243-7

ISSN: 1402-5205

Cover image: © Higyou, Dreamstime.com

(http://www.dreamstime.com/stock-image-cartoon-crowd-jigsaw-field-

image12232251)

Electronic version available at http://umu.diva-portal.org/

Printed by: Print & Media

Umeå, Sweden 2015

i

Acknowledgements In the process of writing a doctoral thesis one navigates through countless

spectra of emotions, which are more or less familiar to oneself in the

beginning of the process. There are great expectations and hope, but also the

occasional feeling of despair. There is great freedom and flexibility, but also

hard work and pressure. There is the joy of having efforts confirmed and

appreciated through acceptance to a journal, but there is also the feeling of

belittling that accompanies rejection. There is the fear of public speaking and

doubt in one’s ability to convey information the first time one enters into a

classroom as a teacher instead of as a student. There is also the wonderful

sensation of being an authority in a subject matter and having the

responsibility and ability to encourage, guide, engage and motivate students

into learning slightly more than they thought possible.

All in all, my time as a PhD-student has been a period of extensive

personal and professional growth. I have met numerous extraordinary

people during this time that have contributed to this thesis in several indirect

and direct ways. I am eternally grateful to all of you and some of you are

further mentioned below.

Three people that have contributed indirectly but substantially to the

completion of this thesis are my children, Sigvard and Majken, and my

husband Erik. You guys give me perspective and you remind me that there is

a wonderful life outside of work and you challenge me to evaluate and

articulate my values. Sigvard and Majken, your curious questions motivate

me to continuously work on my communications skills. Your appreciation for

little things in everyday life inspires me to strive for mental and emotional

presence in everything that I do. Thank you Erik, for being your magnificent

self. Your strong belief in me and my abilities has remained unchanged

throughout this long process. You encourage me to continue when I am in

doubt and you are always the first one to celebrate my achievements. You are

proud when I say “oh, that’s nothing!” and you find ways to motivate me like,

“I’ll grow my moustache back out when you reach that stage” or “We’ll drink

champagne once you have finished that”.

I owe great thanks to my supervisors Urban Lindgren and Rikard

Eriksson. Urban, you have this ability to always know what I need. Be it an

encouraging pep-talk, discussing ideas, or just leaving something to be

resolved by me in due time. You have had confidence in me from the

beginning and you have trusted my ability to come up with my own ideas.

You have this egalitarian perspective that we can both learn from each other,

which was initially terrifying! I did not feel like an authority on any matters

discussed. But I have learnt and now I really think that we can complement

each other’s knowledge. Thank you Rikard, I am in awe of your level of

ambition, your knowledge and thoughtfulness, your efficiency and accuracy,

ii

and your willingness to help in any ways imaginable. Your close readings of

my texts and your precise questions have motivated me to constantly do

better. The pair of you make a wonderful, inspiring, and dynamic supervising

duo that I am fortunate to have worked with.

I am grateful to my two Erikas; Erika Knobblock and Erika Sandow

whom, during different periods of my PhD-project, took me in during my

numerous stays in Umeå. Your hospitality and generosity know no

boundaries. Staying with you guys have enriched my social life

tremendously and you have come to be two of my closest friends. In the later

part of the project, Katarina Haugen and Olle Stjernström took over some of

the responsibility of housing me. You selflessly invited me to your home,

disregarding any inconvenience it may entail. A special thanks to Katarina

for our talks of big and small things. You too, have come to be a close and

valued friend. Thank you also, to Johanna Liljenfeldt for so generously

letting me stay with you, often with short notice! There are more of you that

have offered me a place to stay, thank you to all of you! This project would

not have been possible, had it not been for the generosity and kindness of my

colleagues and friends.

During my time as a PhD student there have been several people that have

helped me with practical issues related to my work. I am grateful to Erik

Bäckström, Lotta Brännlund, Ylva Linghult and Fredrik Gärling for your

invaluable help in; working with the database, administrative concerns, and

dealing with students. Your help is highly appreciated and you deliver it with

such ease and always with a smile. Thank you to Magnus Strömgren for

always finding the time to help out with answering unanticipated questions

from the students related to the GIS- and statistics software in the computer

lab. Great thanks to Martin Henning, Einar Holm and Olle Stjernström for

commenting on an earlier draft of the thesis. Your insightful comments

contributed to significantly improving the quality of the thesis.

Special thanks to Lars Larsson and Kajsa Åberg for always making me

smile and think. You do this by being funny and serious at the same time.

Also, thank you to Olle Olsson for years of teaching together and for your

ability to be your awesome self, and to Mats Borrie for our little talks in the

corridor about this and that. Also, a great thank you to Einar Holm and

Kerstin Westin for thinking, early on, that I could contribute with something

in academia. And thank you Einar, for introducing me to Sture Öberg. Apart

from showing me the width and depth of human geography, Sture managed

to do something that no one else has; give me advice on life and work that I

listened to. I have pondered over things you said in the passing and it has

been a great help in my professional and personal life.

Thank you to my parents-in-law, Agneta and Rolf, for so generously

helping with the logistics of my travels. You have made everything run so

much smoother than it otherwise would have! Thank you also to Linda

iii

Thorell for regularly having lunch with me and by so doing giving me a social

life during long periods of working alone, from home. Finally I want to thank

my parents, Kaj and Lena, for always believing in me and giving me the

impression that you truly think I can accomplish anything I set out to do.

And thank you dad for helping out with the kids during my numerous visits

to Umeå.

Lisa Östbring

Umeå, March 2015

iv

v

Table of Contents

1. Introduction 1 1.1. Contribution and aim 4 1.2. Outline of the thesis 6

2. Theoretical research context 7 2.1. Setting the scene: From agglomeration externalities to proximity 7

2.2. An evolutionary framework 9

2.3. Knowledge in the learning economy 10

2.4. Sector and firm as structures of knowledge 14

2.5. Proximity and relatedness 16

3. Methodological comments 22 3.1. The data 23 3.2. Measuring knowledge and ability 25 3.3. Estimating proximity in the labour force 26 3.4. Measurement of innovativeness 29

4. Paper summaries 31 4.1. Paper I: Labour mobility and plant performance: On the (dis)similarity

between labour- and capital-intensive sectors for knowledge diffusion and

productivity 31 4.2. Paper II: Labor mobility and organizational proximity: Routines as

supporting mechanisms for variety, skill integration and productivity 32 4.3. Paper III: Relatedness through experience: On the importance of

collected worker experiences for plant performance 35 5. Concluding discussion: Relatedness put in place 37

5.1. Findings 37 5.2. Discussion 42 5.3. Directions for future research 45

6. Sammanfattning (Summary in Swedish) 46 7. References 54

vi

Appendices

Paper I: Östbring, L., and Lindgren, U. (2013): Labour mobility and plant

performance: On the (dis)similarity between labour- and capital-

intensive sectors for knowledge diffusion and productivity.

Geografiska Annaler: Series B, Human Geography 95:287-305.

Paper II: Östbring, L.; Eriksson, R.; and Lindgren, U. (submitted): Labor

mobility and organizational proximity: Routines as supporting

mechanisms for variety, skill integration and productivity.

Paper III: Östbring, L.; Eriksson, R.; and Lindgren, U. Relatedness through

experience: On the importance of collected worker experiences for

plant performance.

1

1. Introduction This thesis considers the effects of learning on firm performance. It

specifically examines how various forms of proximity between agents

(individuals) influence the competitiveness of firms. The emphasis is on the

role played by the individuals, their knowledge and their collective ability to

learn and generate novelty within the firm. The thesis rests on the assumption

that learning, knowledge and innovation are central not only to firm

competitiveness, but also to the discipline of economic geography.

During recent decades, knowledge, learning and innovation have gained

increased attention as the main drivers of the economy. Today, it is commonly

argued that the importance of knowledge for sustained firm competitiveness

supersedes that of capital. Early contributions by economists highlighted the

role of human capital and technological change as promoters of economic

growth (Arrow 1962; Romer 1990; Solow 1956). In their studies of an

increasingly globalized world, economic geographers have developed this idea

further. Knowledge and human capital are now often considered the most

localized and most important of all production factors. The accessibility of

information and goods, for both consumers and producers, has changed

patterns of production and consumption and made place “slippery” in many

respects (Dicken 2007; Markusen 1996). The locational advantage of firms is

no longer mainly determined by proximity to physical input factors or markets

or geographically differentiated wages (von Thünen 1966; Weber 1929).

Instead, it is increasingly access to knowledge that has become crucial to firm

survival and competitiveness. The escalated competition brought about by

globalization has increased the rate and frequency of processes of creative

destruction (Schumpeter 1942), in which only the most innovative firms in an

industry survive. Therefore the success of a firm is highly dependent on its

available knowledge resources. It is through these resources that external

knowledge is absorbed and internal knowledge combined and applied in new

ways to generate novelty.

The prominent role of knowledge in the production of goods and services

has given rise to concepts like the knowledge society (Bell 1973), the learning

economy (Lundvall and Johnson 1994) and the knowledge-based economy

(David and Foray 1995). Two of these concepts, the learning economy and the

knowledge-based economy, are often used synonymously, but when they were

first introduced their meanings differed. David and Foray (1995) described the

knowledge-based economy as relying on science and technological knowledge,

mainly in science-driven high-tech sectors. Lundvall and Johnson (1994)

described a learning economy in which all sectors seemed to rely on learning

more than on any other input factor for their development and

competitiveness. It is the perspective of a learning economy in which all

industries can generate economically valuable knowledge that is adopted in

2

the present thesis. The success of a firm is no longer determined by the

distance to physical input factors, which through globalizing processes have

become increasingly ubiquitous (Maskell and Malmberg 1999b), but rather by

the accessibility of relevant knowledge. Even though information accessibility

is at an all-time high, knowledge is more specialized and localized than ever

(Howells 2012; Maskell and Malmberg 1999a). Every place in the space

economy fosters its own unique combination of formal and informal

institutions and knowledge, which to a large extent is the result of past

economic, political and social structures (Massey 1984). The socialization of

knowledge makes it localized because social depth is not easily transferred

over the great physical distances that globalization has brought about. In

economic geography (Morgan 2004) and in theories of the firm (Nonaka et al.

2000), it is argue that social depth is crucial in transforming information into

knowledge. This has motivated an upsurge of theoretical and empirical

contributions on non-physical forms of proximity between agents that could

alleviate the need for physical proximity in the generation of trust and

knowledge (Almeida et al. 2002; Boschma 2005; Boschma et al. 2009;

Grabher 2002a; Lam 1997; Torre and Gilly 2000; Torre and Rallet 2005). It

is in this body of literature that the thesis is embedded.

Traditionally, economic geographers have studied the relationship

between geographical agglomerations and knowledge production. The

primary focus has been on the role of geography in facilitating external

relations between firms. For example, localised learning refers to interactive

learning processes that are stimulated by localised capabilities and the ability

to interact easily with other actors in one’s physical proximity (Malmberg and

Maskell 2002; Maskell and Malmberg 1999a). Similar perspectives on

knowledge creation through external relations are emphasized in theories of

learning regions (Morgan 1997), innovative milieus (Maillat and Lecoq 1992),

industrial districts (Piore and Sabel 1984) and clusters (Porter 1990). All these

different perspectives emphasise the economically desirable characteristics of

co-location, such as competitiveness and high growth. Yet another strand of

literature focused on the generation of knowledge is that concerned with

national and regional innovation systems (Asheim and Gertler 2005; Asheim

et al. 2011; Cooke 2001; Lundvall 1992). These perspectives include

institutions and relationships between various forms of actors within a

geographically delimited territory in their analyses of innovation processes. A

commonality between all these theories is the focus on knowledge

externalities as important sources of novelty. The role of external knowledge

for firms is neither denied nor disregarded, but it is not the focus of the thesis.

Firms cannot rely exclusively on external knowledge for their innovativeness,

nor can they absorb all available knowledge. It is the in-house composition of

knowledge that determines the absorptive capacity of the firm (Cohen and

Levinthal 1990) and the combinative capabilities (Kogut and Zander 1992).

3

Whereas absorptive capacity represents the ability to recognize and absorb

external knowledge, the combinative capabilities of the firm are determined

by employees’ ability to learn from each other and combine their knowledge

in new ways. It is these internal abilities of the firm to learn and innovate that

are under scrutiny in the thesis. This is because the economy is ultimately

made up of individuals in different networks and structures and the firm is the

primary organizing unit of interactive learning processes. Penrose (1959)

acknowledged that the distinction between firm and market is administration,

which occurs in firms but not in markets. Furthermore, Boschma (2004)

discusses and questions whether and how regions can be competitive, as

learning and innovation occur in or between firms that are constituted by

individuals. Brenner and Broekel (2011) examine to what extent the

innovation performance of regions can be measured. They present a number

of indicators of how the region, given that it is a coherent unit, may stimulate

or contribute to firm innovativeness, thus acknowledging that innovation

processes take place within and between firms by combining individuals’

knowledge in new ways.

In the current economic geography literature, and in economics and

innovation studies, the beneficial role of relatedness for economic growth is

highlighted for regions, firms and plants (Boschma et al. 2014; Eriksson and

Lindgren 2009; Neffke et al. 2008). The relatedness, whether it be co-located

firms in related industries (Boschma and Iammarino 2009; Neffke et al.

2012), firms diversifying into activities that are related to their main activity

(Tanriverdi and Venkatraman 2005) or a labour pool of related competences

(Boschma et al. 2009), presumably offers the greatest potential for successful

innovation. The reason is that added knowledge and technology that are

related to existing knowledge and technology in a plant are similar enough to

be absorbed easily, while at the same time different enough to generate ideas

for novel applications. A firm may have all the physical capital to produce a

related good that renders the cost of diversification low. Similarly, a firm in a

locality with a high number of related industries and firms reaps efficiency

gains (externalities) both from knowledge spill overs and from supporting

structures aimed at promoting the regional knowledge in a specific field (see,

e.g., the regional innovation systems literature; Cooke 2001).

In theories of the role of relatedness in the labour pool, whether it concerns

the in-house competences in a firm or labour mobility studies, the link

between relatedness and competitiveness is less clear cut than in theories of

the benefits of related agglomeration or firms expanding into related

activities. Cognitive relatedness in the labour pool facilitates effective

individual and group learning, i.e., knowledge growth. In the case of labour

mobility, it presumably also ensures effective integration of new employees in

firm activities. The idea that relatedness represents a cognitive distance,

which is neither too long nor too short, stems from the literature on proximity

4

dimensions (Boschma 2005; Kirat and Lung 1999; Torre and Gilly 2000;

Torre and Rallet 2005). The literature on (mainly) non-physical proximity

dimensions states that there should be cognitive relatedness between agents

for effective learning to occur. But there are also other dimensions (social,

organizational, institutional and geographical) that potentially can reduce the

demand for cognitive relatedness through their socializing functions. Short

distances in these dimensions reduce uncertainty and promote trust and

understanding between individuals. It is suggested in the present thesis that

individuals can come to be associated in such a manner that they can learn

from each other, and advance firm knowledge, despite a long distance between

(heterogeneity of) their formal knowledge.

1.1 Contribution and aim As discussed in the introductory section, knowledge is considered the most

important resource in the production of goods and services in the learning

economy. There is nothing controversial in claiming that processes of learning

and innovation are essential to firm competitiveness in all sectors. For

example, Hipp et al. (2000), Miles (2005) and Tether et al. (2001) have

studied learning and innovation processes in service industries, whereas Acha

and Brusoni (2003) and Tunzelmann and Acha (2005) have studied

innovation processes in low- and medium-tech industries. It is through

learning that knowledge is obtained and applied to potentially generate

novelty. But whereas the role of knowledge production and spill over effects

has been thoroughly discussed in economic geography (Heimeriks and

Boschma 2013; Neffke et al. 2011), few studies have dealt with the ability of

firms to absorb existing knowledge (Howells 2002). In the present thesis, the

primary study object is the plant and its employees, i.e., the labour force. In

the micro-perspective, the focus is on the interactive learning processes and

absorptive capacities of firms and plants. It is through the employees that the

firm consumes and absorbs knowledge, and by examining the knowledge

variety in plants, the present thesis hopes to contribute to our understanding

of the role of the labour force in firm competitiveness.

Apart from absorbing external knowledge through monitoring and co-

operations with other actors, firms can also acquire knowledge by hiring new

individuals. In such instances, one or a few individuals are brought into the

firm, which constitutes a specific socio-technical context. In this context, the

new employees are expected to communicate their knowledge, whereas

existing employees are expected to absorb, learn and apply the new

knowledge. The effect that labour force knowledge has on competitiveness and

firm performance is primarily determined by the collective ability of

employees to learn and apply the new knowledge. Therefore, analyses of the

relational character of knowledge and learning should take multiple proximity

dimensions and thus types of relations into account. Torre and Gilly (2000)

5

suggest that despite the frequent discussions of the beneficial role of proximity

in economic geography, the exact workings of proximity remain a black box.

This is a black box in urgent need of opening; if the concept is to be more than

a theoretical notion, it needs to be analytically relevant and empirically

applicable. Both Boschma (2005) and Torre and Gilly (2000) discuss the

probable interconnectedness and overlap effects between different forms of

physical and non-physical proximity. Nevertheless, most existing empirical

work on proximity dimensions and knowledge flows (i.e., labour flows) within

and between firms analyse one form of non-physical proximity (see, e.g.,

Boschma et al. 2009; Timmermans and Boschma 2013; Eriksson 2011). These

studies have shown slightly diverging results, which may be attributed to the

interconnectedness of multiple forms of proximity. The potentially variable

impact of formal knowledge variety on plant performance, due to differences

in other types of proximity, is the primary focus of the present thesis.

Furthermore, the thesis also considers the in-house composition of

knowledge, not only the relationship between inflowing knowledge (labour)

and existing knowledge. By so doing, the relative importance of high levels of

human capital can be contrasted with the influence of knowledge relatedness

(variety). Boschma et al. (2009) found high levels of knowledge relatedness to

be more important to firm performance than having high levels of human

capital. But without the integration of human capital and knowledge variety

into the same analytical model, the relationship between the two forms of

knowledge and their unique and combined influence on firm performance

cannot be properly addressed.

Our understanding of the shifting role of the labour force, from physical

labour to knowledge resources, is still limited. Therefore, starting from the

literature on proximity dimensions and relatedness, the aim of the present

thesis is to examine how different forms of proximity in the labour force

influence the competitiveness of firms in the learning economy. The point of

departure is the understanding that the proximity dimensions afford trust and

reduce uncertainty between individuals. These are aspects that promote

interactive learning processes.

There are multiple claims that processes of knowledge production and

learning differ in different sectors. Pavitt (1984) introduced a taxonomy of

drivers of technical change in different sectors that has been extensively used

ever since. Furthermore, Strambach (2013) claims that knowledge and

innovative behaviour accumulate at the level of the industry, the firm, and the

region, whereas Malerba (2006) points to the differentiated innovation

processes in different sectors. But most empirical studies of labour force

knowledge as a promoter of growth have either studied the entire economy

(e.g., Boschma et al. 2009) or just one sector (e.g., Bienkowska 2007). This

leaves either very broad generalizations about the workings of knowledge

throughout the economy, or thorough descriptions of the particular. In the

6

thesis, the different conditions under which the labour force interacts and

applies its knowledge are examined. Each plant in the economy comprises a

unique context that is reproduced by the participants and their knowledge and

abilities, and by cognitive and physical structures within the plant

(Nooteboom 2009). By scrutinizing the interplay between local micro-

diversity (i.e., the skills and knowledge of the local labour force) and

overarching structures (e.g., firm routines or technological regimes), an

integration of micro- and meso-level mechanisms is achieved, which allows

for a rich analysis of the role of knowledge in promoting competitiveness for

firms and plants. Through this integration of the role of the labour force and

its knowledge, on the one hand, and structures in the form of firm routines

and production technologies, on the other, it is acknowledged that both parts

affect the potential for innovation. It is shown that labour force composition

and knowledge are essential parts of firm competitiveness, but only if the

surrounding structures offer ways of applying knowledge that is circulating.

In the thesis, knowledge is analysed at the level of the plant and combined

with factors that are believed to affect the innovative power of learning in

firms. These factors are embedded in the local context and industry setup.

Moreover, the labour force is formed by real-life individuals that react

differently to the same situations due to their different abilities, knowledge

and experiences. Therefore, the efficacy of labour force knowledge in

enhancing firm performance is not only contingent on the sum of the

employees’ formal knowledge. Learning processes are social processes and

similarities in past work experiences and institutional environment influence

the ease and success of interactive learning processes. Factors that are

assumed to influence the innovativeness of firms are explored in three

empirical studies. Each of them pursues a different aspect of learning and

knowledge application in firms. The three papers have different foci; the

effects of cognitive and geographical proximity on learning in plants in

different sectors, the effects of cognitive, organizational and geographical

proximity on learning at the plant level and the effects of various forms of

cognitive proximity on learning in plants. At the same time, they share the

umbrella theme of the labour force as a (mainly localized) knowledge resource

and potential promoter of economic growth.

1.2 Outline of the thesis The thesis comprises three empirical papers, a theoretical framing and

discussion and remarks on methodological considerations and delimitations.

In this introductory section, Section 1, the overarching aim of the thesis and

the three papers is introduced. The following section, Section 2, places the

empirical papers (referred to as Paper I to III) in the theoretical context to

which the thesis aims to contribute. By taking an evolutionary perspective on

knowledge production and reproduction, the context-dependent nature of

7

learning and competitiveness is highlighted. The role of the individuals and

the potential benefits of being embedded (Granovetter 1985) in different

contexts for learning processes are discussed. There is also a short discussion

of the presumed hindering or enabling meso-level structures associated with

learning and knowledge application in specific sectors. Furthermore, the

underlying theoretical implications of the proximity dimensions framework

(Boschma 2005; Torre and Gilly 2000) are considered. It is argues that the

relational (non-physical) aspects of the proximity dimensions are of greater

importance to learning processes on the micro-level of interpersonal

knowledge exchange. In Section 3, methodological considerations are

discussed, mainly those concerned with operationalization of theoretical

concepts. Section 4 offers summaries of the empirical papers and in Section 5

the main findings and conclusions of the papers are discussed. In this section

there is also an introduction of potential future research based on the findings

in the thesis. A short Swedish summary is provided in Section 6.

2. Theoretical research context This section presents a general theoretical background for the thesis. It does

not reflect the theoretical frameworks for the different papers, which can be

found in the initial part of each paper. Instead it provides a description of the

theoretical context to which it aims to contribute. It starts with a short

overview of the changing role of proximity in theories in economic geography.

It then moves on to specifically present the evolutionary perspective on

growth, and the role of physical and non-physical proximity in interactive

learning and knowledge production for firms. The point of departure is the

understanding that contemporary capitalist society is substantially shaped by

knowledge and skills. These properties of society reinforce the role of

individuals in theories of economic growth. It also points to the importance of

micro-level analysis (the level of the firm or plant) as complements to meso-

or macro-level analyses of sectors, networks, regions and nations in

understanding the uneven spatial development of successful economic

activities.

2.1. Setting the scene: From agglomeration externalities to proximity

Already in 1890, Marshall (1890) acknowledged that there was something

“in the air” in agglomerations that seemed to stimulate the competitiveness of

firms. Proximity is thus not a new idea in economic geography, but for a long

time it was mainly physical proximity to numerous external auxiliary

functions that was thought to be important to firm competitiveness. The

benefits of high concentrations of agents and firms have been repeatedly

highlighted in the agglomerations literature. Proximity to competitors, sub-

8

contractors, clients, large markets and a specialized labour pool have all been

considered locational advantages for firms, dating back to von Thünen (1966

(English version of the 1826 publication Der Isolierte Staat)) and Marshall

(1890) and their predecessors. After the Second World War, Fordist mass

production was gradually replaced with leaner forms of production. This was

mainly a result of technological advancements occurring at an increasingly

rapid pace, which allowed for consumer influence and custom-made products.

On a global scale, the technological advancements meant substantial

improvements of, initially, the physical infrastructure and, later on, the digital

infrastructure (Dicken 2007), which gave the impression that ‘distance was

dead’ (Cairncross 1997). The role of science and technology in the

advancements of the global networks that emerged steered the foci of

economic geography research toward specific geographical agglomerations

and clusters, namely those of highly competitive and successful high-tech

industries (Cooke 2004; Frenkel 2001; Klepper 2010; Saxenian 1994;

Sternberg 1996).

With the focus on high-tech regions came an inflated interest in knowledge

and people as resources, rather than just physical labour. In the literature,

there was increased mention of the embodiment of knowledge. With such a

socialized competence view came reflections on how human capital,

knowledge workers and talents formed the local environment (Florida 2002;

Lucas 1988; Romer 1990; Scott 2004). This was described as occurring not

only indirectly, through their place of work, but also through their lifestyle

preferences. The role of individuals in shaping the local environment was put

forward by economic geographers as being twofold: partly as being part of the

labour force and as such –, small units of cognitive abilities – collectively

shaping the innovativeness of firms, and partly as being cultural and social

beings that were part of communities and networks that reproduced not only

knowledge, but also values and norms (Bathelt and Glückler 2005; 2013;

Farole et al. 2010; Storper and Scott 2009; Storper and Venables 2004).

Silicon Valley is probably one of the most researched regions with regard to

the relationship between technological innovativeness and social and

institutional structures (see, e.g., Saxenian 1994; Kenney and von Burg 1999).

The recognition of workers as embodying knowledge and knowledge

production as occurring in a cumulative fashion in a specific socio-technical

context (Strambach and Klement 2012) offered further insights into the

variable outcomes of seemingly similar knowledge inputs in economic

processes in different spatial contexts (Gertler 2003; Leiponen and Drejer

2007).

The benefits of geographical proximity, as discussed in the agglomerations

literature, came to concern the structures and networks, in a region, that

supported or sustained high levels of technical innovation rates. Institutional

and social perspectives on differentiated regional growth introduced non-

9

physical forms of proximity between actors. As mentioned in the introduction,

geographical proximity – or co-location – facilitates and stimulates

interaction between actors (Knoben 2009). In theories of the learning region,

geographical proximity facilitates collaborations based on trust between

actors (Morgan 1997), and in the industrial districts literature the role of

collaborations between small- and medium-sized firms is stressed (Piore and

Sabel 1984). The importance of geographical proximity in the innovative

milieu literature is based on the facilitated relations between knowledge

producing actors (Maillat and Lecoq 1992). The role of geographical proximity

in cluster theory highlights proximity to competitors and partners (Porter

1990). Eriksson (2011) found an inverse relationship between the positive

effects of related variety and distance of origin of labour inflow, indicating the

importance of place-specific institutions in understanding variety in

production technology. Similarly, Rigby and Essletzbichler (2006) and

Haltiwanger et al. (1999) found great spatial differences in production

techniques within the same manufacturing industries. The difficulty in

imitating and learning from other firms in the same sector but from different

regions preserves the heterogeneity between firms and demonstrates a place

dependence of knowledge that is due to specific social and institutional

characteristics.

There is a voluminous body of literature on place-specific institutions and

firms as structures that affect, and are affected by, the behaviour of micro-

agents. The reciprocal relationship allows a collective past to influence the

perceived probability of success associated with different actions (see, e.g.,

Amin and Cohendet 2005; Bathelt and Glückler 2013; Farole et al. 2010;

Maskell 2001). These structures are integrated into individuals’ cognitive

abilities, partly as past experiences and partly as current practices.

2.2. An evolutionary framework The seminal work of Nelson and Winter (1982), on the evolutionary

perspective, has had a major impact on the promotion of evolutionary theory

in economics. According to Nelson (1995), evolutionary theory in any

discipline discusses the same three key principles: they explain the temporal

development of something, they consider random elements that generate

variation, and they emphasize the existence of inertial forces that provide

continuity among entities. In evolutionary economics, the three main

mechanisms that influence the outcome of economic activities are: the

routines, the search for new and better routines and the selection

environment. Evolutionary economics is quite different from neoclassical

economics in its rejection of the economy as a closed system. Furthermore, it

argues that technological change should be the primary focus of economic

analysis. There is great emphasis on novelty and change in evolutionary

economics, which is in contrast to traditional equilibrium theory. In

10

evolutionary economics, decreasing returns due to price competition is a

secondary phenomenon in comparison to innovation processes and the search

for novelty and Schumpeterian rents (Boschma and Frenken 2006).

Following theories of evolutionary economics, an evolutionary economic

geography framework has recently been developed (Boschma and Frenken

2006; 2009; Boschma and Martin 2007; 2010). The concept of evolution was

not altogether absent in geography and economic geography prior to this work

(see, e.g., Martin 2000; Massey 1984; Rigby and Essletzbichler 1997), but the

intention of the framework is a clearly defined theory of economic

development in economic geography. The ambition with the framework is a

theory complemented with research methodology and a research agenda. In

evolutionary economic geography, firm routines and their development over

time are core attributes in the competitiveness of firms. The overarching aim

of evolutionary economic geography is thus to analyse the changing spatial

distribution of routines and the resulting impact on the distribution of

economic activity.

Within an evolutionary framework, analyses of economic processes can be

carried out on multiple levels (Boschma and Frenken 2006). At the macro-

level, the development of sectors and networks can be analysed within spatial

systems, such as cities, regions or nations, and their structural and dynamic

prerequisites compared in a global system. At the meso-level, evolutionary

economic geography highlights the spatial development of sectors and

networks and the analysis of relations of competition and complementarity.

Finally, at the micro-level, the firm and its routines are under scrutiny. The

development of the firm and the effects of variety on the competitiveness of

the firm are highlighted as important research topics. At all levels of analysis,

the fundamental assumption is that structures (institutions, routines, etc.)

lead to the accumulation of knowledge and capabilities, in firms, sectors,

networks and regions, which can be steered in different directions through a

combinatorial dynamic driven by the interactions of agents (Strambach and

Klement 2012; Strambach 2013). The present thesis focuses exclusively on the

micro-level of analysis and learning processes that occur as the result of

internal variety, at the level of the plant. The reproduction and adaption of

firm routines occur within the firm and its plants as responses to market

changes. The accumulated knowledge in a plant determines the efficacy of

knowledge absorption and adaption. Therefore, studying the in-house

composition of knowledge and abilities is a crucial step in understanding the

variable ability of firms to adapt and consequently remain competitive.

2.3. Knowledge in the learning economy The territorial accumulation of knowledge occurs on multiple scales and

has many names, for example varieties of capitalism (Hall and Soskice 2001),

regional innovation systems (Cooke 2001) and localized learning (Maskell and

11

Malmberg 1999). The spatial accumulation of knowledge in structures and

people suggests that there is a place-specific characteristic of knowledge and

how it is produced, reproduced and ultimately applied (Lam 1997; Rigby and

Essletzbichler 2006; Storper 1997). Knowledge accumulates in regions due to

the interactions of social and technological networks that have developed over

time (Leiponen and Drejer 2007; Rigby and Essletzbichler 1997). Knowledge

can thus be seen as a structuring component throughout society. On the

individual level, people make increasingly reflexive decisions about their

everyday life and practices based on their knowledge (Giddens 1990). On the

systems level, knowledge has come to be an essential part in explaining how

the economy works (Bell 1973; Lundvall and Johnson 1994). But firm

routines, the industry and its technological regime and the local labour market

all influence the choices and outcomes of knowledge applications. The path-

dependency of knowledge generates non-physical proximity between

individuals belonging to the same firm, industry and labour market. Thus,

knowledge does not only have to be technologically relevant to be conversed

and applied within the firm. It also has to be understood and communicated

throughout the specific firm in a specific region. This thesis scrutinizes the

micro-diversity of labour force knowledge and links it to macro-structures in

an attempt to better understand the contingent nature of learning and

innovation.

The concept of the learning economy (Lundvall and Johnson 1994) did not

only introduce knowledge and processes of learning as key factors in

innovation and firm competitiveness. It also pointed to the role of knowledge

in all sectors, not only in the traditionally R&D-intensive sectors, but also for

firms in, for example, the service industry. Barras (1986; 1990) argues that the

rise of ICT (Information and Communication Technology) led to an industrial

revolution in the service sector. The use of ICT allowed firms in the service

sector to improve their efficiency and quality and develop new services at a

faster pace than had previously been possible. In the learning economy, the

acquisition, creation and implementation of knowledge are the key factors in

explaining competitive advantage, as firms must continuously innovate to not

lose market shares (Nonaka 1994; Lawson and Lorenz 1999). Under such

circumstances, it is not only the learning that is important, but also the ability

to apply knowledge to productive practice.

The embodiment of knowledge in people makes it a unique resource.

Because people are embedded in local contexts, their knowledge is not only

personal, but also localized and contextual. Therefore, the effects of sharing

bits of knowledge vary between interactions due to different interpretations

(Nonaka et al. 2000). These interpretations are based on personal experiences

(Polanyi 1962; March 2010), understanding of the firm routines (Nooteboom

2009) and technological knowledge of the specific industry (Pavitt 1998). But

innovations within firms are not just the result of the combined abilities of all

12

employees to absorb, converse and apply knowledge. There are also external

factors that always, but variably, affect the formation of a context and

influence the propensity for knowledge application and innovation. These

factors are the industry in which an activity is carried out (see, e.g., Pavitt

[1984] on the determinants for sectoral technological trajectories) and the

firm in which the activity takes place (see, e.g., Dosi et al. [2008] on firm

routines as a selection mechanism for knowledge absorption). Whereas place-

specific institutions (i.e., Gertler 2003; Maskell and Malmberg 1999a)

contribute to the formation and production of knowledge, there are factors

associated with firm and industry that do not only steer the knowledge view

within the firm, but also affect the firm’s capacity to implement the ideas of

individual employees. Individuals, firms, sectors and regions are all shaped by

past events and experiences. The configuration and development of economic

activities in specific places are evolutionary and path-dependent on multiple

levels.

In contemporary economic geography, descriptions of knowledge have

come to include relational and social aspects. For instance, Howells (2002)

describes knowledge as a socially constructed process, whereas Amin and

Cohendet (2005) define it as a process and practice rather than a possession.

Lam (2000) considers knowledge to exist between individuals rather than

within them. Thus, knowledge is a resource in itself, but it also impacts how

the firm uses other resources. Knowledge is therefore a major differentiating

factor for firm competitiveness, because the ability of firms to procure,

converse and apply knowledge varies. In addition to that, knowledge is also a

volatile resource due to its embodiment in people. If a person decides to

change jobs, the firm risks losing crucial bits of knowledge (Howells 2012).

Due to the dependence on knowledge, and implicitly on the labour force or

even single individuals, throughout the capitalist economy, the importance of

micro-level analysis of labour force characteristics and collective learning

processes cannot be stressed enough. The present thesis aspires to uncover

some of the complexity associated with learning, innovation and economic

growth within firms.

Although knowledge and learning have only recently begun to attract the

attention of economic geographers, the economy has always depended on

knowledge (Hudson 2009). The complexity of knowledge and learning, and

the later association with innovation, has generated a vast body of literature,

ranging from psychology and cognitive science (Vygotsky 1962) to sociology,

organizational studies (Cohen and Levinthal 1990; Lam 2000) and economics

(Harrison et al. 2001; Penrose 1959). Whereas philosophy largely views

knowledge and learning as internal processes within the individual, sociology

sees learning as an interactive process between two or more individuals. In

economics, Penrose (1959) regarded the varying ability of firms to

administrate similar production factors to be the main differentiating factor

13

in firm growth. The notions introduced by Penrose have been further

developed in economics, and knowledge has come to be regarded as the main

resource or capability of the firm in the knowledge-based theory of the firm

(Grant 1996a).

Howells (2012) goes back to reinvestigate Michael Polanyi’s work on tacit

knowledge and knowledge flows to determine the appropriateness of the

rather flexible use of the concept in economic geography. He theorizes that

knowledge cannot flow freely, only information can. Instead, the embodiment

of knowledge in people requires human interaction for learning to occur. As

mentioned in the introduction, information becomes knowledge only when it

is meaningful to the receiver (Nonaka et al. 2000). Meaning is created from

associating information with what one already knows and has experienced

(March 2010; Polanyi 1962). If past experiences help individuals interpret

information and situations, the implication is that processes of understanding

are tacit. Therefore, processes of knowledge creation and learning are always,

but varyingly, tacit and thus socially embedded (Morgan 2004) and

cumulative (Strambach 2013). In his early work, Polanyi (1962; 1966) put

forth the idea that knowledge exists in a wide spectrum (not a dichotomy)

ranging from wholly tacit to wholly explicit knowledge. But because the

understanding and application of knowledge are based on personal and

collective experience, which affects perception and anticipation, there is no

such thing as completely explicit knowledge – it is always subjective and

contextual.

In economic geography, Bathelt and Glückler (2003) deal with this

subjectivity in terms of a relational perspective on knowledge, in particular,

and resources, in general: Knowledge is in the eye of the beholder and the

socio-technical context determines how knowledge advancements are best

achieved. March (2010) also acknowledges the subjectivity of knowledge by

discussing employees’ experiences as useful but imperfect teachers in dealing

with situations at the current place of work. The success of communication

between two individuals, thus, depends as much on past experiences and

awareness of practices associated with a specific industry and firm as it does

on their formal knowledge. The contextual and relational aspect of knowledge

was introduced in economics and the resource-based view of the firm in the

late fifties by Penrose (1959). She claimed that the main differentiating factor

for firms with apparent similarities in input factors was their capability to

combine existing resources in different ways. A few decades later, the same

phenomenon but on a different scale – the local environment or the region –

was picked up by economic geographers. Theories of localised capabilities

(Maskell and Malmberg 1999a) and learning regions (Morgan 1997) are a few

contributions that consider the socially embedded and relational character on

knowledge economic geography. The socializing and structuring of

knowledge occurs within firms as well as within regions. It is therefore

14

reasonable to assume that knowledge and the effects of learning in firms and

regions cannot be understood in isolation, but rather in combination with

socializing structures, such as technological trajectories and formal and

informal networks. Individuals in a certain place produce and reproduce

capabilities through their formal competences and interpretations of values,

norms and goals.

2.4. Sector and firm as structures of knowledge One form of macro-structure is the technology used within a sector or

specific industry. The possibility that different production technologies offer

structurally different opportunities for learning and innovation has been

somewhat overlooked in studies on knowledge dissemination within and

between firms. Pavitt (1998) stated that the specificity of input factors in

various sectors renders them differently equipped to integrate inflows of skills

and knowledge into plant routines. The reason is differences in technological

regimes and routines (Dosi 1982; Nelson and Winter 1982), which generate

unequal absorptive capacities (Cohen and Levinthal 1990; Nooteboom et al.

2007) and partly explain asymmetric industrial development (Patel and Pavitt

1997). Because the technological trajectories of firms vary, their propensity to

act on and respond to new knowledge differs with regard to how well the

information is understood and can be used within the plant. According to

Pavitt (1984), learning processes are sector-specific and dependent on the

firm’s principal activity and the knowledge inputs.

Knowledge has a cumulative aspect to it, which can be explained by strong

firm routines and rigid structures, but cumulativeness may also arise from a

dependence on machinery and physical capital in other forms (Dosi 1982;

Patel and Pavitt 1997; Pavitt 1998). Thus, firms that rely heavily on their

physical capital may exhibit high levels of path-dependence partly because of

knowledge accumulation in physical capital. This kind of knowledge

accumulation lowers the probability of innovation through labour (i.e.,

knowledge) inflow. It also impedes the ability to absorb and utilize external

knowledge. Lagerholm (2007) showed that within mature industries (i.e.,

manufacturing in the Swedish context), process innovation is mainly the

result of supplier innovation rather than in-house innovation. Thus,

substantial knowledge advancements are brought to the plant via investments

in new machinery rather than through knowledge contributions by in-house

or in-flowing staff. In labour-intensive sectors, on the other hand, both

process and product innovations are mainly the result of in-house knowledge

applications, and the sheer number of innovative opportunities is thus greater

than in capital-intensive sectors. Therefore, the variety (relatedness) of

knowledge in industries that are dependent on their labour rather than their

15

physical capital will be characterised by a larger number of potential

applications of that knowledge.

Within firms, knowledge is administrated and structured in accordance

with the knowledge view and firm goals (Nonaka et al. 2000). The firm can,

thus, be perceived as an administrative unit (Penrose 1959) of individuals that

creates cohesion through its routines. It also influences the knowledge view of

its employees through these structures. When Nelson and Winter (1982)

originally presented their evolutionary theory of the firm, they perceived

routines as the capabilities of the firm and as the genes of the firm. This

implied that routines could both have gene-like generative qualities and be

manifested as firm behaviour. The ambiguity of the concept has persisted and

the interpretations and use of the concept are manifold, but there is consensus

on the context-dependence and regularity aspects of routines (Cohen et al.

1996; Dosi et al. 2008). Context-dependence and regularity suggest that

routines create a form of proximity (intra-organizational proximity) between

agents within the same firm and subsequently steer their view on knowledge

usability and applicability. This will be further discussed in the section on

proximity.

Within the conceptual evolutionary economics literature, routines are

often conceived of as the latent genes of the firm, synonymous with macro-

level social and cognitive maps of how to interpret problems and context,

manifested through firm behaviour. Whereas Pentland (1995) and Pentland

and Reuter (1994) liken routines to grammar, where a routine is not one

pattern but a set of possible patterns, Miner et al. (2008) consider

organizational routines to be the foundation of organizational learning

because routines function as organizational memory. These perspectives

agree on the role of routines as the rules, social and physical structures and

cognitive scripts and abilities of organizational members. These

organizational restraints create path dependence on the firm level, in both the

knowledge view and attitudes toward external knowledge. Routines at the

micro-level are often concerned with the individual’s explicit task and position

within a routine process and the performance of that task (Becker 2004). Most

empirical research has been carried out as case studies at the micro-level (e.g.,

Becker and Zirpoli 2008; Feldman 2000, 2003; Weick and Roberts 1993),

where specified routines have constituted tangible and measurable processes

from which inference has been drawn about their interconnectedness, their

adaptable and/or stable nature and the role of individual agency.

The dichotomy between micro- and macro-views of routines is misleading

due to the interdependent relationship between the two. Routines take place

in a context that makes explicit routines partly tacit through individuals’

interpretation of “how things are done” within in firm. In the everyday

workings of routines, individual members use their understanding of the

macro-level routine to adjust their performance at the micro-level in response

16

to circumstances and outcomes (Feldman 2000, 2003). In order for a routine

to be efficient, there need to be strong and continuous connections and

interactions between the different agents partaking in the process (Feldman

and Rafaeli 2002). The connections should preferably generate extensive

knowledge of the links between distributed activities within the organization

(Weick and Roberts 1993). This can be expressed as a high level of contextual

understanding, which presumably leads to extensive knowledge about the

individual’s and others’ role in the firm. Furthermore, organizational

members are aware of latent structures and “how things are done” in the firm,

which potentially allows rapid application of internal ideas.

2.5. Proximity and relatedness In the present thesis, the analysis of learning and innovation processes

emphasizes the relations between individuals rather than using a full systems

approach. There is no doubt that the relations between individuals are

embedded in greater systems and that external knowledge input is necessary

for firms to sustain competitiveness (see, e.g., Bathelt et al. [2004], Grabher

[2002b; 2004] and Maskell et al. [2006] on the importance of inter-

organizational projects for firm innovativeness). However, by only focusing

on relational aspects of individual attributes in the labour force and thus

abstracting from a number of influences, an increased analytical clarity is

achieved. MacKinnon et al. (2002) argue that in order for learning to be a

relevant factor in explaining the spatial distribution of activities, the role of

knowledge needs to be carefully grounded empirically. By focusing on

individual knowledge and interactive learning within plants, the thesis

attempts to isolate the effect that learning has on plant performance. This

section presents the proximity dimensions framework. It also discusses the

role the proximity dimensions for interpersonal learning play on the micro-

level of the plant.

The Russian developmental psychologist Vygotsky (1962) claimed that

learning occurs through interaction between internal cognitive processes and

the context. He introduced the idea that a person can gain a certain amount of

knowledge on her own through internal processes, but that through

interaction with others (helpers or instructors) there is potential for further

knowledge advancements. He called this potential further learning that could

be achieved through interactions the Zone of Proximal Development (ZPD).

In order to promote someone else’s learning, there must be pieces of shared

interests or mutual focal points from which to depart, otherwise

communication will fail. In reference to recent contributions within economic

geography, Vygotsky’s theory relates to the growing body of literature on

cognitive proximity and other forms of non-physical (relational) proximity.

When the distance between agents, in one or several of these non-physical

proximity dimensions, is just right, related –, learning is facilitated and

17

innovation more likely to occur than if there are long distances between the

dimensions (Boschma 2005; Harrison et al. 2001; Song et al. 2003; Torre and

Rallet 2005). Referring back to Vygotsky (1962), long distances in all

proximity dimensions cause a void between individuals that complicates

communication and learning. On the regional level, a short distance in

multiple proximity dimensions between firms may lead to a regional lock-in.

This is a scenario in which understanding and agreement between agents are

high, but equally high is the risk of myopia (Maskell and Malmberg 2007) and

an inability to identify new relevant technological knowledge (Boschma 2005;

Visser and Boschma 2004). The present thesis considers that the same

phenomenon is likely to occur on the interpersonal level of the plant.

Theories of non-physical types of proximity between firms were originally

introduced by the French Proximity Dynamics Group (Kirat and Lung 1999;

Torre and Gilly 2000; Torre and Rallet 2005). They distinguish between two

dimensions of proximity: the physical dimension, which is geographical

proximity, and the non-physical dimension, which is organizational

proximity. Their theory illustrates how interactive knowledge creation and

innovation processes are shaped by a relationship between relational factors

and physical distance. It is further argued that the relevance of the different

proximities changes in different stages of the innovation process. In their

framework, geographical proximity refers to physical distance in a twofold

manner: first: firstly, as perceived by individuals and thus partly socially

constructed, and second, as a separation of activities and relations in space.

The organizational dimension, on the other hand, refers to a relational

distance that affects the ability to coordinate actions. This dimension is

constructed in accordance with two related logics. First, similarity in firm

routines, which are partly built on shared beliefs and high degrees of tacit

knowledge, offer the same reference space and shared understanding of

structures and knowledge view. This is a perspective also taken by Blanc and

Sierra (1999). Second, firms that are close in organizational terms belong to

the same networks and have interactions of various forms. In the present

thesis, where labour force knowledge and relatedness between individuals are

under scrutiny, the former definition of organizational proximity is adopted.

The logic here is that individuals will benefit from having a shared

understanding of structures and knowledge view in their knowledge

interactions. A short organizational distance between individuals ensures

overlapping, or ideally identical, reference spaces. This proximity, or

familiarity, represents an institutional micro-context that will reduce

uncertainty and promote trust between individuals. Trust is a key factor in the

transfer of highly tacit knowledge and thus in interactive learning.

Multiple proximity models have been developed (e.g., Blanc and Sierra

1999; Kirat and Lung 1999; Zeller 2004) since the initial work of the Proximity

Dynamics Group. The most detailed account of the analytical relationship

18

between proximity and innovation is provided by Boschma (2005). In his

model, Boschma presents five proximity dimensions: geographical,

organizational, cognitive, institutional and social proximity. Geographical

proximity is defined as both absolute and relative distance, and the impact on,

and relationship with, the other proximity dimensions is underlined.

Boschma’s (2005) definition of organizational proximity differs from that of

Torre and Gilly (2000) in its focus on control of relations and hierarchy. In his

model, organizational proximity refers to the coordination and control of

intra- and inter-organizational relations. Organizational proximity is achieved

when firms are connected through coordinated channels and activities, and

this proximity reduces uncertainty and the risk of opportunism. This

description of organizational proximity is better suited to examination of

inter-firm learning processes and innovation. This is because it is associated

with the accessibility and power relations between firms that determine how

organizational networks are shaped.

The cognitive proximity dimension, as presented by Boschma (2005),

deals with differences and similarities in the capabilities of economic agents.

Antonelli (2000) and Howells (2012) have argued that knowledge is

distributed among different firms and industries. Firms learn and create

knowledge by combining existing pieces of complementary knowledge in new

ways (Nooteboom 2000). In order for firms to absorb external knowledge

(Cohen and Levinthal 1990) and combine it with internal knowledge, the

external knowledge must be similar enough – related – to the existing

knowledge to be recognized as valuable (Boschma and Lambooy 1999). The

cognitive distance should not be too long, nor should it be too short, because

short distances do not offer any learning potential. The institutional

proximity dimension refers to the degree of similarity in values, norms, rules

and laws between economic agents. It is, thus, regarded as being dependent

on both informal and formal structures. Boschma (2005) discusses the close

and intertwined relationship between institutional proximity and

geographical proximity. In the present thesis, they are regarded as

representations of the same phenomenon, namely similarity in norms, values

and understanding of formal and informal rules. Because the labour force and

labour flows are examined, the physical distance between firms influences the

probability that a person will change employer from one firm to another. But

the distance also represents an institutional similarity, or difference, that

affects the understanding of norms and practices among individuals and may

therefore influence the level of trust between individuals. As mentioned

previously, trust is thought to be necessary for the transfer of highly tacit

knowledge. Lastly, the social proximity dimension, as presented by Boschma

(2005), is associated with social ties between individuals resulting from

friendship and kinship. Thus, the social dimension is the only dimension that,

in its original form, specifically focuses on the micro-level of inter-individual

19

relationships. This proximity dimension is not further examined in the thesis,

but the other dimensions are considered on the micro-level of interpersonal

distances within the plant. The rationale for the micro-perspective is that, in

order to understand how firms learn, we need to examine the knowledge that

constitutes the firm and its absorptive capacity. Following Torre and Gilly

(2000) and MacKinnon et al. (2002), the thesis attempts to empirically

ground the concept of knowledge by further opening the black box of

proximity relations in relation to knowledge transfer and learning. Each proximity dimension offers a way to construct shared values between

individuals and potentially also a shared basis for communication and

learning. In the proximity dimensions framework suggested by Blanc and

Sierra (1999), they identify a dimension that they call relational proximity:

“Relational proximity accounts for the existence of non-economic

relationships among individuals in the form of a common working ethos, a

common language and culture, good mutual knowledge, mutual trust,

mutually respected norms of behaviour” (Blanc and Sierra 1999;197). In the

present thesis, all of the above-discussed proximity dimensions fit under the

description of relational proximity, suggested by Blanc and Sierra (1999). This

relational characteristic implies that what two individuals lack in mutual

formal knowledge may be compensated for by mutual norms and codes of

conduct or a background in similar organizational contexts. Whereas

Boschma (2005) argues that cognitive relatedness is a necessary but not

necessarily sufficient prerequisite for learning to occur between firms, it is

argued here that, on the interpersonal level, relatedness in any one of the

proximity dimensions may be sufficient for learning to occur. The reason for

this goes back to Vygotsky’s ZPD and the need for a connection between

individuals in order for them to communicate efficiently, begin to trust one

another and subsequently learn from each other. The temporal aspect is also

crucial to the distinction between intra- and inter-firm learning. Within the

firm or plant, individuals remain for extended periods of time and inflows of

labour are regarded as, more or less, permanent additions to the labour force.

In inter-firm learning, numerous relations are temporary. Therefore, there is

a greater need for cognitive relatedness, which allows for efficient

communication of the unique knowledge that each project participant can

contribute. In the case of mergers and acquisitions, it is unlikely that two firms

with completely different core competences will merge. In such cases,

similarity or relatedness in the organizational or institutional dimension can

facilitate or impede learning processes (Boschma 2005) initiated by

relatedness in the cognitive dimension.

Regardless of the level of analysis, there is still great uncertainty about

whether or not the proximity dimensions are of equal importance to learning

and innovation. Moreover, few previous empirical studies have examined to

what extent proximity in one dimension can replace, or compensate for, a long

20

distance in another dimension. The theoretical literature on proximity

dimensions and relatedness between agents offers plausible interpretations of

the relational characteristics of knowledge. But many of the suggestions made

in the literature have yet to be empirically tested and grounded. The effects of

cognitive and geographical proximity on learning and innovation have been

the most theorized and empirically tested on both the interpersonal and

interfere level. On the interpersonal level, Boschma et al. (2009) found that

cognitively related in-house competences as well as cognitively related labour

inflow positively impacted plant performance. They further identified

interdependence between the cognitive and the geographical/institutional

dimension. Cognitively unrelated inflow that originated from the same local

labour market positively influenced plant performance. This is interpreted as

indicating that cognitive relatedness is not always a necessary condition for

interpersonal learning to occur. It is rather the combined influences of

multiple types of proximity/distance between individuals that determine the

learning potential. On the other hand, Neffke and Henning (2013) created a

measure of relatedness between industries, based on labour flows between

them. Their results indicate that firms are more likely to diversify into

activities to which they have pre-existing ties through their labour force than

into activities that diverge substantially from their core activities.

Furthermore, Neffke et al. (2011) showed that evolution of the regional

industrial setup is strongly path-dependent. They demonstrated that

industries that are related to pre-existing industries are more likely to enter a

region than those that are not.

In studies on labour mobility, the cognitive dimension is commonly

measured as relational differences in formal education, sector belonging,

occupation similarity or intensity of previous knowledge exchange between

actors (Boschma et al. 2009; Neffke and Henning 2013; Timmermans and

Boschma 2013). There is coherence in the view that, for fruitful

communication to occur, knowledge and skill should be neither too similar

nor too different. Therefore, bringing together people with related skills and

knowledge is the most efficient way to promote learning. In such instances,

people are able to understand one another while simultaneously drawing on

each other’s knowledge to advance their own knowledge, as suggested by

Vygotsky (1962). But the indicators for cognitive relatedness do not consider

the plethora of factors that influence individuals’ cognitive abilities. Merely a

fraction of an individual’s skills and abilities are measured with each of the

different techniques of estimating cognitive relatedness. This is mainly due to

methodological and empirical restraints and limitations and leads to sub-

divisions of cognitive relatedness. This is fine, as there is no way to include all

factors that influence individual ability, mainly because of the subjectivity of

knowledge and learning and partly because there are no registers and micro-

data that include such levels of detail. But a greater awareness of the

21

multifaceted ways of achieving cognitive relatedness, in particular, and

relatedness in other dimensions, in general, is important.

As in most empirical work, Paper I and II in the present thesis included

and analysed, analyse one aspect of cognitive relatedness. Because different

studies use different definitions of cognitive relatedness, there is a risk of over-

or under-estimating the effects of relatedness on firm/plant performance.

This is due to the exclusion of other factors that also influence the abilities and

skills of the agents under study. One such other factor is organizational

proximity that arises from similarity in firm routines. With regard to the

notion that firm routines are fundamental in creating organizational

proximity, there are several researchers who claim that knowledge within a

firm is distributed rather than dispersed, meaning that everyone knows some

information, but no one knows all of it (Becker 2004; Cohen and Bacdayan

1994; Feldman and Rafaeli 2002; Nonaka et al 2000; Lam 2000; Weick and

Roberts 1993). This unique context creates isolating mechanisms that hinder

not only external imitation of firm capabilities, but also internal articulation

of what these capabilities are exactly. The tacit nature of firm routines

generates intra-organizational proximity and possibly also a social proximity

that together create trust between individuals. This trust potentially allows

them to discuss ideas over wider cognitive spectra than would otherwise have

been possible.

Due to the relational and contextual nature of knowledge and the varying

degree of tacitness, knowledge is difficult to transfer and develop. Labour

mobility is thus the most efficient channel for dissemination of highly tacit

knowledge, but presumably also for dissemination of less tacit knowledge

(Szulanski 1996). Many have studied the transfer of embodied knowledge by

examining the effects of labour mobility on performance and innovation (see,

e.g., Agrawal et al. 2006; Almeida and Kogut 1999; Boschma et al. 2009;

Breschi and Lissoni 2009; Maliranta et al. 2009; Song et al. 2003;

Timmermans and Boschma 2013). Whereas Almeida and Kogut (1999)

showed that the mobility of engineers influenced local transfers of knowledge,

Power and Lundmark (2004) found labour mobility among key persons to be

the driving force behind the upgrading of skills within ICT clusters. Breschi

and Lissoni (2009) built on the results of Almeida and Kogut (1999) and put

forward the idea that labour mobility is the main conductor of local knowledge

transfer, superior to formal networks and social ties. Lacetera et al. (2004)

were able to quantify the effect that inflowing high-ability individuals can have

on their colleagues, thus showing that single individuals can impact the

knowledge of the entire workforce.

When knowledge is put in a new context, there is potential for completely

new applications of that knowledge, due to the new employee’s interpretation

of what is needed as well as the “old” employees’ understanding of the new

knowledge. Strambach and Klement (2011) argue that there is a combinatorial

22

knowledge dynamic at the micro-level of the firm. This works, in combination

with the cumulative knowledge dynamics inherent in the socio-technical

context of the region, to generate novelty (innovation). By bringing a specific

skill-set from one context to another, the individual potentially contributes to

the generation of new combinations of existing knowledge as well as to the

creation of entirely new knowledge (Asheim and Coenen 2005; Nonaka et al.

2000; Pinch and Henry 1999). But there is still uncertainty, not only about the

learning capability of the employees, but also about the external

circumstances (i.e., factors external to the individuals partaking in the

knowledge exchange) under which knowledge is most likely to be successfully

applied within a plant. The present thesis incorporates theories of sector- and

firm-specific structures into the description of the relationship between

physical and non-physical proximity and interactive learning in firms.

3. Methodological comments The thesis is based on quantitative modelling of socio-economic processes

within and between plants. Methodology is closely related to the philosophy

of science, as different methods are used to capture different aspects of the

social world. The scientific methods employed in the thesis are shaped by

critical realism (Bhaskar 1978). The collective learning processes associated

with the knowledge composition within a plant are considered to be

contextual and cumulative and manifested in changes in plant productivity.

The view on learning as a process and practice driven by the variety of

accessible knowledge fits well into an evolutionary perspective on the

accumulation of knowledge and the growth of the economy, as discussed

earlier in the thesis. The evolutionary perspective has many common traits

with a critical realist ontology. Castellacci (2006) puts forth a view on reality

as complex, differentiated and structured, as a system that is open and

continuously changing and transforming, and as a world characterized by

great uncertainty.

In critical realism, the world exists independent of human awareness

(Sayer 1981). Furthermore, the world exists in three domains: the real, the

actual and the empirical (Collier 1994). The real domain contains all physical

objects, but also mechanisms and structures associated with the objects and

combinations of objects. Thus, it acknowledges the existence of non-physical

objects and processes in society. In the actual domain, events occur

irrespective of whether humans register them or not. In much social science

production, mechanisms in the real world are used to explain the development

of events in the actual world. Finally, the empirical domain represents the

events that are experienced by humans (Collier 1994).

Social science, in general, and evolutionary economic geography, in

particular, study phenomena in the empirical domain (the spatial distribution

23

of economic activities), the ambition being to understand the mechanisms

that underlie these phenomena (institutional settings, social networks,

embodied knowledge, etc.). Similar to the evolutionary perspective on

economic growth, critical realism regards science as constantly evolving.

Theories of the deeper domains are produced and reproduced based on their

plausibility in the empirical world. There is likely not one absolute truth in

social science, but there are explanations that are more plausible than others,

based on observations in the empirical world. Thus, constantly checking for

sources of error in the empirical setup and building on previous research are

essential elements in validating the theoretical claims made about structures

and mechanisms in the real domain (Collier 1994). Scientific explanations of

events offer theoretical descriptions of the mechanisms that caused them. By

empirically testing knowledge relations and innovations in firms, fragments

of the real world can be captured that, in turn, can inform us about probable

causal mechanisms. These causal mechanisms, found in both humans and

structures, can improve our understanding of pieces of the real world. By

combining insights from numerous theoretical descriptions, a synthesized

understanding of the functioning of the real world may be achieved (Lawson

1997).

3.1. The data The empirical analyses in the three papers are based on data from the

ASTRID database at the Department of Geography and Economic History at

Umeå University. The database integrates several administrative registers

maintained by Statistics Sweden (SCB) and offers detailed, geo-referenced,

longitudinally linked information about all individuals, plants and firms in

Sweden. By integrating registers, it is possible to connect attributes to people,

and people to plants and firms with attributes of their own. The longitudinal

aspect of the database enables analyses of changes in work force composition

within individual plants, as well as recording of individuals’ past work

experiences with regard to plant, industry, and firm. Because it includes

recorded coordinates for all buildings, residences and plants, the database has

a high spatial resolution. This enables analyses of mobility patterns and

agglomeration activities. The data in ASTRID are measured and encoded on a

yearly basis, and the values of the variables in the database reflect the situation

at a specific point in time. Thus, it is not possible to see whether a person has

had multiple employments in a year. But by comparing personal wage and

social benefits, it is possible to determine whether a person has been employed

and unemployed during the same year. The analytical unit in all the empirical

papers is the plant. A plant is traditionally associated with manufacturing, but

in the database and in this thesis it a workplace. This workplace can be of any

size and belong to any industry. The plant represents a physical building or

part of a building that is administrated by a firm. A firm can have multiple

24

plants, which all have their unique plant identification number (CFAR), but

share the same organizational identification number (OrgID). Individuals are

associated with plants, not with divisions or subunits within a plant.

Therefore, it is not possible to determine whether a person has changed jobs

or positions within a plant, but intra-firm mobility between plants can be

detected. Mobility is captured as a change in plant identification number and

change in plant coordinates for individuals in the database, and consequently

internal mobility (intra-plant mobility) is concealed in the data, not just in the

analyses conducted in the thesis but in the structure of the database in general.

The intra-firm mobility between plants is recorded as a change in plant

identification number within the same firm identification number.

The Swedish labour market is characterized by fairly high job mobility

rates, which correspond to those in several other north European countries,

e.g., Denmark, Finland and the Netherlands (EUROFOUND 2006). Between

1988 and 2012, 68 percent of all employees who ended their employment did

so to change jobs, not due to retirement or layoffs triggering unemployment

(Andersson et al. 2014). Thus, job change is the dominant type of labour flow

on the Swedish labour market. Furthermore, job mobility is more prevalent in

some segments of the labour market than in others. Highly educated

individuals show higher job change rates than do the less educated. Men

change jobs more often than women do, younger individuals more often than

older people and highly paid individuals more often than those in the lower

wage segments (Jansson 1997; Widerstedt 1998). In the analyses of job

mobility (Paper I and II), only mobility related to job change is considered.

Thus, transition from unemployment to work is excluded from the analyses,

whereas change of workplace is included.

The empirical analyses are based on data covering the period between 2000

and 2010. Paper I uses data from the period 2002-2005, whereas the analysis

in Paper II is based on data from 2003-2008. Paper III examines data from

the period 2000-2010. Thus, the results in Paper II and III could,

theoretically, be affected by the financial crisis that struck the labour market

in 2008 and 2009 (Öhman 2010). But the Swedish labour markets and

financial markets were relatively mildly impacted by the global crisis. Between

the middle of 2008 and the end of 2009, 120,000 people lost their jobs. This

can be compared to the crisis in the early 1990s during which 600,000 people

lost their jobs (Öhman 2010). During troughs, job mobility decreases

(Andersson and Tegsjö 2006), but in Paper II all job changes were recorded

between 2003 and 2006, and only plant performance was recorded up until

2008. In Paper III the empirical analysis is based on the in-house composition

of knowledge in knowledge-intensive business services (KIBS). Thus the flows

are implicitly considered in the yearly estimations of in-house knowledge

variety. The performance of the plants may have been negatively impacted by

the recession, which may have led to an underestimation of the impact of

25

different forms of knowledge variety on performance. But the empirical setups

in the two papers have mitigated the risk of any severe underestimation. In

Paper II a longitudinal panel was estimated, and in Paper III pooled OLS

regressions with observations over a ten-year period and yearly dummies to

control for time-specific heterogeneity were calculated.

In Papers I and II the focus is on labour mobility and its impact on plant

performance. The models used in these papers do not control for the exit of

labour from the plants, only the entry. Thus, in some cases the inflow of labour

replaces exiting employees, whereas in other cases the inflowing labour is a

new addition to an expanding plant. There are a few central concepts in the

papers that are discussed in greater detail below in an attempt to visualize the

transformation of theoretical concepts into quantitative variables.

3.1 Measuring knowledge and ability This section deals with the concept of knowledge and ability and to what

extent knowledge can be assessed through quantitative methods. The central

role of knowledge in the learning economy has been emphasized throughout

the thesis. The European Commission states that:”…the key aspect of human

capital has to do with the knowledge and skills embodied in people and

accumulated through schooling, training and experience that are useful in

the production of goods, services and further knowledge” (De la Fuente and

Ciccone, European Commission 2002, p7). Formal measures, such as

academic degrees and occupation titles, offer guiding information about the

theoretical and practical scope of parts of an individual’s knowledge. Relating

back to the critical realist ontology, an individual’s formal knowledge

(education, occupation or industry experience) is a signal (Spence 1973) in the

empirical domain, which allows the researcher to make informed assumptions

about a portion of the individual’s ability. Arrow (1973) introduced the idea of

the educational system as a filter through which the most able students pass,

whereas the less capable individuals are caught in the filter and prevented

from further knowledge advancements. In Spence’s (1973) article on job

market signalling, academic degrees are signals to employers about the ability

of the individual to perform. Over the years, the individuals change and adapt

their signals with increased experience and additional work titles. Thus,

formal measures of knowledge are conscious efforts on the part of the labour

force to signal not only ability, but also acquired knowledge to employers.

Factors that are believed to socialize knowledge and affect learning are, to

some extent, accounted for through the integration of proximity dimensions

other than the cognitive dimension into the analytical models. But the main

focus of the papers is on the relationship between formal knowledge variety

and plant performance. Educational background is used to estimate

relatedness in knowledge between individuals in Paper III. Educational

background and industry experience are used as measures of knowledge in

26

Paper I and occupational title is used in Paper II. Whereas educational

background represents formal theoretical knowledge, industry experience

informs us of how the individual’s knowledge has been practically applied.

Occupation title considers both the applied knowledge through the work tasks

and the individual’s qualification for those tasks (educational background).

These three formal measures describe parts of the individual’s total knowledge

that is codified and actively acquired. They accurately represent knowledge

that the individuals have obtained, but they do not represent all the knowledge

that the individuals possess. The formal measures of knowledge tell us little

about the social and tacit components of knowledge (Grabher 1993; Polanyi

1962). In Paper III aspects of the social character of knowledge are touched

upon through the integration of individuals’ past experiences. These past

experiences have influenced and shaped employee knowledge beyond the

information in the strictly formal records. By tracing employees’ former co-

workers an image of the context in which they have been active emerges,

which allows for an analysis of the relative similarity of previous work

experiences in the current labour force within a plant. It is suggested here that

this is a way to approach the tacit dimension of knowledge and ability in a

quantitative manner. Past work experience informs us about how formal

knowledge has been applied. It also offers insights into the knowledge

networks of individuals. The networks indicate what parts of the individual’s

own formal knowledge may have been accentuated and developed further as

a result of these interactions.

In contrast to Becker’s (1964) human capital theory, in which he claims that

individuals weigh the costs and benefits of investments in higher education,

the Swedish context offers few costs associated with higher education because

education is free. The possibility for student loans and social and economic

benefits for students are the same for the entire Swedish population. The

general educational level in the population is high, and the individual reward

(income) of additional years of schooling is below average in the European

Union (Björklund et al. 2000; Eliasson 2006). The European Commission

found the average return on each year of schooling to be a 6.5 percent increase

in personal income in the European Union (De la Fuente and Ciccone 2002).

3.2 Estimating proximity in the labour force In this section the characteristics of the labour force and the various forms

of proximity that are estimated in the papers are presented and discussed.

Throughout the empirical papers (Papers I-III) various forms of proximity

dimensions are referred to and their effects on interactive learning are

discussed. A short distance in any of the dimensions implies similarity

between agents. Boschma (2005) argues that similarity reduces uncertainty

between actors. Actors can be firms, organizations and governmental

institutions, but also individuals. Whereas Boschma (2005) mainly discusses

27

the role of the various forms of proximity on the inter-firm level, the present

thesis examines the role of proximity in interactive learning processes on the

inter-individual level within plants. The plant is an administrative unit of

inclusion in the sense that it generates cohesion between its members

(employees). This is achieved through a common understanding of firm

routines and the technological trajectory of the specific industry in a given

region. From this rationale it follows that individuals who enter the plant, as

newly employed labour, may diverge from the existing staff. Apart from

differing in formal knowledge, new employees may also have different

experiences of firm routines and institutional settings that originate from a

different firm in a different region. The efficiency of absorption of individual

knowledge into plant knowledge is affected by the degree of similarity between

the inflowing labour and the existing in-house labour. In the remainder of this

section, the operationalization of geographical, cognitive and organizational

proximity is elaborated upon and discussed.

Geographical proximity is estimated between inflowing labour and the

existing in-house labour in Paper I and Paper II. The inflowing labour is

categorized as being either intra-regional or inter-regional with regard to local

labour markets. Thus, the inflows are considered proximate (similar) or

distant (different or unrelated) in a dichotomous manner. The rationale for

this division is that the geographical proximity dimension is closely related to,

not to say intertwined with, the institutional proximity dimension (Boschma

2005; Farole et al. 2010; Rallet and Torre 1999). In the regional innovation

systems literature, the varieties in regional (or local) institutional settings are

emphasized in the competitiveness of firms and industries (Asheim and

Gertler 2005). This aspect of institutional proximity may be of little

importance to interpersonal learning processes, but belonging to a community

and enjoying insider advantages within one’s own labour market are aspects

of the geographical proximity dimension that are considered in the thesis.

Inter-regional labour mobility represents a scenario in which the individual

enters into a new community from a context that the new co-workers know

nothing or little about. Using the same logic, it is argued that intra-regional

labour flows preserve the insider advantages and remain within the same

institutional environment or community. The familiarity associated with

intra-regional labour flows presumably reduces uncertainty and generates

trust between the in-house labour and the inflowing labour.

The non-physical proximity dimension of cognitive proximity is estimated

in various ways in the papers. The reason for this is to show that different types

of knowledge can generate cognitive relatedness between individuals, through

which they can learn from each other. In Paper I the in-house knowledge

composition is calculated as an entropy measure of variety (as suggested by

Frenken et al. 2007). The estimations of relatedness in the in-house labour

force are based on educational background, whereas the relationship between

28

the inflowing labour and the existing in-house labour is based on industry

codes (as done by Boschma et al. 2009). It is assumed that a knowledge inflow

from related industries is efficiently absorbed while simultaneously

contributing new pieces of knowledge that can be combined with the existing

knowledge in new ways. In Paper II, entropy measures based on occupation

are used to determine the degree of relatedness in the in-house labour force.

The same entropy calculations as in Paper I are used to determine the degree

of similarity, relatedness and unrelatedness in the in-house composition of

knowledge. Occupation is a slightly more refined measure of knowledge than

educational background, as it reflects both work tasks and individual

qualifications. The relatedness between the inflowing knowledge and the

existing knowledge in Paper II is estimated using the same relational entropy

measures that Boschma and Iammarino (2009) used when determining the

relative similarity of trade linkages and the regional industrial setup. The

inflowing labour and its occupation titles are compared to the occupational

distribution within the plant. By so doing, the levels of similarity, relatedness

and unrelatedness are determined between the inflowing labour and the

existing labour.

In Paper III, in-house knowledge variety is estimated in three different

ways. This is done to demonstrate that cognitive relatedness can be achieved

in more than one way. Individuals’ experiences shape their cognitive abilities

(March 2010), and it is argued in Paper III that, in an evolutionary framework,

the knowledge of the labour force should be regarded as evolving over time

and, thus, as being influenced by past events. Just as firms, industries and

regions are affected by their past history, so are the micro-agents, the

individuals, which constitute the labour force. The first measure of cognitive

relatedness is an entropy measure of formal education variety, exactly like the

in-house estimations in Paper I. The degree of similarity, relatedness and

unrelatedness in formal education in the labour force are followed by an

experience measure of variety. The similarity in industry experience is

estimated in the in-house labour force. This type of knowledge represents

applied knowledge. For each employee a fifteen-year backlog is retrieved from

the database that allows us to see what industries the individual has worked

in, during the past fifteen years. The industry experiences of all employees at

the plant are combined in an “experience pool”. From this pool of industry

experiences, the degree of similarity, relatedness and unrelatedness in labour

force industry experience are determined. The third measure of cognitive

proximity is a measure of similarity in knowledge networks in the in-house

labour force. This type of knowledge represents variations in experiences in

practices and epistemic communities within the current place of work. Similar

to the industry experience measure, the knowledge network is calculated from

a pool of knowledge exposures that each employee has been subject to during

the past fifteen years. The educational background of every person who has

29

worked with the individual under study is recorded and added to the pool of

knowledge exposures at the plant. The same entropy measures, as for the

other two types of cognitive proximity, are used to estimate the degree of

similarity, relatedness and unrelatedness in knowledge exposure. The

rationale for estimating cognitive proximity in three separate ways is to

demonstrate that knowledge relatedness can be achieved in multiple ways.

Thus, a labour force that is characterized by high levels of unrelatedness in

one of these measures may be related through another type of cognitive

relatedness.

Organizational proximity is estimated in Paper III in a similar fashion to

geographical proximity, as explained above. Individuals are regarded as being

either organizationally proximate or organizationally distant in a

dichotomous manner. Intra-organizational labour inflows are individuals

changing plants within the same firm. These flows represent proximity in the

organizational dimension. The inter-organizational flows are individuals

originating from another plant and firm. These flows are organizationally

distant. Organizational proximity is considered to generate a mutual

understanding of “how things are done” and a coherent knowledge view and

awareness of firm norms and values. These aspects create combined

institutional and social cohesion within the micro-context of the firm. On the

interpersonal level, this is presumed to reduce uncertainty and contribute to

efficient communication. Inter-organizational labour flows originate from

other micro-institutional environments that may differ significantly from the

new context. Therefore, these flows need to familiarize themselves with the

new structures and internal networks and their initial interactions with the

existing in-house labour will entail more uncertainty.

3.3 Measurement of innovativeness Before proceeding to the measurement of successful learning processes at

the plant, the relationship between knowledge and innovation is considered.

According to Pavitt (2005), innovation processes contain three overlapping

sub-processes: production of knowledge, conversion of knowledge into

physical and non-physical artefacts and the matching of these artefacts to

market needs and demands. Thus, all innovation processes depend on the

acquisition, creation and application of knowledge. This knowledge is

disseminated on the market and conversed within the firm or plant (Nonaka

1994). It is important to keep in mind that knowledge creation and acquisition

do not occur at one specific point in time during the innovation process, but

rather throughout the entire process. This has been both empirically

demonstrated (see Butzin and Widmaier [2012] on innovation biographies)

and theoretically argued (e.g., Kline and Rosenberg [1986] critique the linear

model of innovation [Bush 1945]).

30

On a related note, the reverse product cycle (RPC) introduced by Barras

(1986; 1990) in the service industry also enters into traditional

manufacturing. RPC constitutes three successive phases: improve efficiency,

improve quality and, lastly, new services. The order of these stages is opposite

to that of the product cycle model for manufacturing introduced by Abernathy

and Utterback (1978). But approximately 35 years after the introduction of the

product cycle model, the reach of capitalism and the increasing pace of

Schumpeterian creative destruction have impacted the product cycle in

manufacturing. The cost associated with improving the quality of existing

products and the efficiency of production processes are substantially lower

than they are in relation to developing new products. Therefore, improving

quality, improving efficiency, finding new markets, creating services around

existing products and making incremental changes to existing products are

viable options that generate noticeable results within a foreseeable future.

This development, one of increased global competition, indicates that

collective learning processes, which generate incremental innovations within

plants, not only can but do occur in all sectors. But to pinpoint exactly how

and when essential bits of knowledge are exchanged or recombined is

impossible using quantitative methodology, and also beyond the scope of the

present thesis. Instead, the thesis rests on the assumption that learning occurs

in all sectors, and that it can be detected in the performance of plants due to

the predominance of innovations that increase efficiency.

To estimate and evaluate the potential increase in efficiency due to

collective learning processes, plant performance is specified as labour

productivity at the level of the plant in Paper I, II and III. The database does

not contain information on productivity, so performance is instead measured

as value added per employee. But the database only contains information

about value added for firms, not their respective plants. For firms with only

one plant this is unproblematic, but for multiplant firms the database ascribes

the value added of the entire firm to all of its plants. This is dealt with in Paper

I, Paper II and Paper III by distributing value added to plants in accordance

with their wage quota. By ascribing value added to plants based on their wage

quota, both the education and experiences of the employees are considered

(Wictorin 2007). As discussed in a previous section, individuals signal their

knowledge and ability (Spence 1973) and the employer is ensured of their

ability through the filtering function of higher education (Arrow 1973). High-

ability individuals are efficient (Becker 1962), which is recognized by the

employer, who awards them with higher salaries. High efficiency is a mixture

of innate ability, specific knowledge related to educational degree and generic

knowledge in communicating, learning and synthesizing knowledge inputs.

Following this logic, it is reasonable to allocate value added in proportion to

wage quota. In Papers II and III, value added was also proportionally

allocated to plants with regard to the share of firm employment (as done by

31

Martin et al. 2011). This approach is perceived as being more egalitarian,

because every employee is valued as being equally important to the firm. This

was done as a robustness check, and the allocated value added remained the

same in Paper III. The reason for this may be found in the sample of firms in

Paper III, which only consists of KIBS. Unlike a manufacturing firm with

multiple plants with different functions, firms in KIBS will most likely have

similar functions and human capital levels at all their plants. In Paper II the

models were estimated with the alternative productivity measure, and the

trends in the models did not change. The procedure for allocating value added

is therefore deemed fit for the purpose of the papers.

4. Paper summaries 4.1. Paper I: Labour mobility and plant performance: On

the (dis)similarity between labour- and capital-intensive sectors for knowledge diffusion and productivity

The first paper, Paper I, analyses the differences between capital- and labour-intensive sectors with regard to the impact of workforce knowledge composition and labour mobility on plant performance. As discussed in the theoretical background, knowledge relatedness between individuals is assumed to be beneficial to knowledge absorption and learning. But the conditions under which new knowledge can be applied and generate novelty presumably vary between different industries and firms. In the literature on proximity and relatedness, especially papers focusing on individuals and labour flows, knowledge transfer and learning are often treated as if they occur in isolation, detached from any surrounding context. This disregard of the time-place fusion in which every action is carried leaves the analyses incomplete and at risk of over- or underestimating the effects of knowledge and relatedness on plant performance. The aim of the paper is to bridge a perceived gap between micro- and macro-level theories of knowledge generation and innovation in the space economy and thereby to nuance the debate on non-physical proximity and the labour force as a knowledge resource. Whereas there is one strand of literature advocating the benefits of complementarities and relatedness for plant performance (Boschma et al. 2009), there is another strand discussing the different prerequisites for innovation in different technological regimes (Castellacci 2008; Pavitt 1984). These strands are neither contradictory nor opposing, but they rather exist in parallel and should be combined if one wishes to model the context in which individuals interact and learn to promote innovation and productivity. Through the integration of these theories in conjunction with theories of the effects of labour mobility, the paper shows that learning and innovation do not occur in isolation. Everything takes place in a context determined partly by the individuals involved, but also by the main technology used by the firm.

For the empirical part of the paper, geo-referenced longitudinal employer-employee micro-data are used. All firms in Sweden are categorized as being either capital-intensive or labour-intensive. The effects of labour flows

32

between plants, within and between labour markets, are then analysed for capital-intensive and labour-intensive sectors, respectively. The sample contains all plants in Sweden with 10 employees or more that have had a labour inflow in 2003 and value-added for both 2003 and 2005. The analysis is carried out using weighted least square (WLS) regression analysis. An additional variance analysis (ANOVA) is performed to determine the relative effect of each of the co-explaining factors on plant performance.

The aim of the paper is achieved by testing the effects of in-house constellation of knowledge on productivity, of type of knowledge inflow on productivity and of geographic origin of the inflow on productivity. By so doing, the effects of cognitive and geographical proximity in the labour force on plant performance are tested in the labour-intensive sectors and the capital-intensive sectors, respectively. The assumption is that the labour-intensive sectors will show overall higher effects of labour inflow on plant performance than will the capital-intensive sectors, which rely heavily on the efficiency of their physical capital (machinery) and supplier ingenuity for their productivity.

The results presented in Paper I show that knowledge type (similar, related or unrelated) and knowledge origin (local or non-local) have different effects on plant performance in the two sectors. A high degree of related knowledge in the in-house workforce positively impacts plant performance in both sectors, but the relative effect (when compared to other explanatory factors) is larger in the labour-intensive sectors. The analysis of labour inflow indicates that knowledge in the capital-intensive sectors is localized – only intra-regional labour flows give rise to increased plant productivity. In the labour-intensive sectors, the geographic and cognitive dimensions complement one another; similar knowledge needs to be non-local in order to be beneficial to plant performance, and unrelated knowledge mainly contributes to plant productivity growth when it originates from inside the region. The findings presented in the paper indicate that there are substantial differences between the capital-intensive sectors and the labour-intensive sectors. With regard to the analyses of in-house skill portfolio, related variety has considerably more influence on firm performance within the labour-intensive sectors.

The overall weak effects of knowledge on performance in the capital-intensive sectors may be attributed to high levels of cumulativeness of knowledge stored in the physical capital.

4.2. Paper II: Labor mobility and organizational proximity: Routines as supporting mechanisms for variety, skill integration and productivity

In Paper II the context-dependent nature of knowledge is further scrutinized through examination of the role of firm routines (proxied as intra-organizational proximity) in learning and innovation. The aim of the paper is to contextualize labour force knowledge by focusing on the socializing and structuring nature of firm routines, the goal being to offer new insights into theories of knowledge as a resource, in general, and theories of knowledge

33

transfer via labour mobility, in particular. The impact of knowledge inflow, which is assumed to have an innovative effect, is measured as changes in per capita productivity at the plant. By operationalizing familiarity with firm routines as intra-firm labour flows, and contrasting them to unfamiliarity with firm routines, operationalized as extra-firm labour flows, knowledge transfer is studied in light of not only cognitive and geographic proximity, but also organizational proximity.

Paper I showed that there are differences in the effects of related knowledge inflow in capital- and labour-intensive sectors, respectively. This supports the hypothesis that learning and innovation depend on more than just the relatedness of formal knowledge between the partaking individuals. The probability of successful application of new knowledge is also contingent on the firm’s routines, which steer the knowledge view and shape patterns and processes of communication within the firm. Whereas studies in economic geography have shown the importance of having a related knowledge base within a plant, as well as related knowledge entering the plant (Boschma et al. 2009; Eriksson 2011; Timmermans and Boschma 2013), the literature in organization studies indicates that firm routines constrain and/or enable learning through the rules, social and physical structures and cognitive scripts and abilities of organizational members (Dosi et al. 2008; Miner et al. 2008; Nelson and Winter 1982; Pentland and Reuter 1994). The firm can be perceived as a social community specialized in the generation and transfer of knowledge. In the present paper, familiarity with firm routines is perceived to generate intra-organizational proximity between organizational members. This gives internal labour flows the advantage of a shared context with the in-house staff. Internal labour flows thus have the benefit of a shared social and cognitive map and knowledge of firm goals with the in-house staff. This advantage presumably offers a shared basis of understanding and eases communication, which alleviates the need for related formal knowledge in order for learning to occur.

The analysis employs longitudinal micro-data from the ASTRID database at the Department of Geography and Economic History at Umeå University. For the empirical estimations, the Arellano-Bond dynamic panel generalized method of moments (GMM) is applied. The analysis is based on 8,633 internal job moves and 9,418 external job moves to 33 plants and four firms, offering a total of 99 observations, in Sweden between the years 2003 and 2006. Together with a group of co-explaining factors, the effects of cognitively differentiated intra- and inter-firm labour flows on plant performance are determined. The effects of the inflow variables are estimated at three points in time: the same year as the inflow occurs, the year after and two years after; this is done because internal flows are believed to be integrated more rapidly and produce, more or less, instant productivity gains.

The results indicate that the theoretically positive relationship between cognitively related labour flows and plant performance is not unaffected by the inclusion of an organizational proximity dimension in the analysis. Cognitively related knowledge inflow positively impacts productivity when it

34

originates in another firm and thus is organizationally unrelated. The positive effect on productivity is delayed, theoretically due to the longer time needed for absorption and conversion when the cognitive dimension is the only shared dimension. Cognitively related intra-firm labour flows (individuals changing plants within the same firm) are familiar with firm routines and thus organizationally similar (a short distance in the organizational dimension). This combination of similarity in the organizational dimension and relatedness in the cognitive dimension generates too much similarity for any innovative learning to occur. The findings suggest that cognitively related labour flows are primarily economically beneficial between firms. In such cases, cognitive relatedness is enriched by sets of firm-specific norms and routines associated with the previous employer that allow for rich combinations of knowledge at the new plant.

When a third proximity dimension is added to the models, namely geographical proximity, the effects of cognitively related inflows are positively reinforced in the extra-regional model. This indicates that relatedness in the cognitive dimension is mainly necessary when there are no other socializing proximities present, i.e., organizational or geographical proximity. In the absence of organizational and geographical proximity, cognitive relatedness is necessary and sufficient to generate effective learning and application of knowledge. These results imply that a cognitively related labour inflow originating from afar (firm and place) contains a greater variation, associated with the past place and firm, than does intra-firm or intra-regional inflows of cognitively related knowledge. Throughout most of the analysis there is a positive relationship between cognitively unrelated flows and plant performance in the intra-firm models. These flows are organizationally proximate and absorbed instantly and efficiently, generating productivity growth for the plant over consecutive time periods. The proximity in the organizational dimension allows for unrelated knowledge to be combined and applied in a familiar setting. When the inflow is still organizationally proximate but geographically distant, the positive effects of cognitive unrelatedness disappear. This indicates that distance (unrelatedness) in two dimensions, of which one is the cognitive dimension, cannot be overcome by proximity (similarity) in the organizational dimension.

Another outcome of Paper II is the framing of the intrinsic nature of the proximity dimensions with regard to learning and knowledge transfer in the labour force. Without any hierarchical ranking of their relative importance, it has been shown that multiple proximities exist that can promote learning and affect plant performance. The effect of the cognitive dimension can be reduced or amplified by the distance in other dimensions. In the paper, a combination of organizational and geographic proximity was shown to generate a state in which cognitively similar and related knowledge inflows either were detrimental or insignificant for productivity. It is concluded that cognitive relatedness is not a necessary condition for learning to occur in the labour force.

35

4.3. Paper III: Relatedness through experience: On the importance of collected worker experiences for plant performance

In Paper III the role of labour force experience in learning and innovation in firms is examined. In contrast to the two previous papers, Paper III does not explicitly study labour mobility, instead in-house competences are observed over time. The formal cognitive abilities of employees, as expressed by their education, are supplemented with cognitive abilities in the form of employees’ previous work experiences. These experiences are divided into industry experience and past knowledge exposure associated with previous workplaces. Theory clearly states that experience is an elemental part of individuals’ interpretative abilities (Polanyi 1962; March 2010) and should thus be considered when analysing the in-house innovative potential. Past experiences are essential in forming individuals’ abilities to seize ideas, communicate them and transform them into productive practice. Therefore, in the present paper, the cognitive proximity dimension is divided into three forms of knowledge, each with its own variety and distance. The experience variables represent socialized cognitive abilities that have been used in processes and practices at plants. Furthermore, interaction terms are constructed and their impact estimated in order to determine the potential overlap effects of different forms of cognitive proximity. By analysing the interaction effects, Paper III not only shows that the combined effects of different forms of proximity differ from the independent effects, but it also shows that the impact of one form of knowledge on plant performance is conditional on the level of another knowledge variable.

The reason for Paper III is the vast interest in knowledge as a valuable resource for firm performance. This interest has gradually shifted focus from the accumulation of human capital to the inflow and accumulation of related and complementary knowledge, as the most important determinants of knowledge transfer, learning and innovation (Boschma 2005; Harrison et al. 2001; Nooteboom et al. 2007; Tanriverdi and Venkatraman 2005). Paper III builds on the theory that a shared practice can be achieved through related cognitive abilities between individuals (Boschma 2005), represented as past work experiences. Shared practice is manifested in an understanding of how things are done, familiarity with firm goals, coherence in knowledge view and a good ability to communicate. It is further suggested, in the paper, that relatedness in formal knowledge between employees facilitates communication, whereas relatedness in industry experience and knowledge exposure offers a plethora of possible applications of formal knowledge and understandings of multiple practices. Therefore it is assumed that cognitive relatedness within the firm can be achieved in different ways by highlighting different aspects of employees’ total cognitive abilities.

The empirical analysis, a weighted pooled OLS regression with year-, industry- and region-fixed effects, is based on geo-referenced longitudinal matched employer-employee data. In the study, the abilities and experiences of employees in knowledge-intensive business services (KIBS) are analysed

36

for firms in Sweden between 2000 and 2010. Firms in KIBS depend on their employees’ ability to learn and apply knowledge for their performance, as they mainly engage in consultancy activities and research in various knowledge fields. Thus, the micro-diversity within plants is more thoroughly scrutinized in the present study than in the previous ones because the employees are, more or less, the only resource for many of these firms. In addition to the examination of staffs’ formal abilities, a 15-year backlog of industry experience and past knowledge exposure for every current employee is created and analysed to determine the degree of relatedness in industry experience and knowledge exposure.

The results show that high human capital ratios positively impact plant

performance. Considering that these firms mainly produce and sell

knowledge, the results are not surprising. The resource of knowledge

presumably promotes not only technological development, but also economic

growth in both firms and regions in all industries. Thus, high human capital

ratios represent the knowledge that has accumulated in a place over time. In

this instance, the place is a firm, or rather a plant. But when the level of human

capital is put in the same model as the various forms of knowledge variety, a

substantially more nuanced picture of the workings of knowledge on plant

performance emerges. It is shown that the effects of the different forms of

knowledge composition variables on plant performance are conditional on the

level of human capital at the plant. That is to say that the effect of, e.g., high

levels of similarity in formal knowledge on plant performance varies as a

function of level of human capital. This is something that previous theories of

proximity dynamics have put forward, but been unable to empirically test. The

results further indicate that the effects of the negative relationship associated

with high levels of similar or unrelated formal knowledge on learning are

abated as the level of human capital increases at the plant. High levels of

similar formal knowledge and human capital even generate productive

learning processes in plants in KIBS. In previous studies on the impact of

labour force knowledge on plant performance, the different forms of

knowledge have been tested separately. By combining them and creating

interaction terms, Paper III is able to more accurately represent the actual

conditions for learning and innovation at a plant. Not only are the different

forms of knowledge combined and tested simultaneously, but the multiple

ways of accomplishing and measuring cognitive relatedness are also

recognized. In most existing empirical work, high levels of similar knowledge

have negatively impacted plant performance. But by combining multiple

forms of knowledge and estimating their relationship with performance

simultaneously, this study has been able to show that the impact of similar

knowledge is conditional on the human capital ratio at the plant. The

conditional relationship between human capital levels and the cognitive

relatedness of the labour force have not been empirically tested previously.

37

Instead most studies have chosen to include either human capital or cognitive

relatedness in the labour force when estimating the effects of interactive

learning on plant performance.

5. Concluding discussion: Relatedness put in place The present thesis set out to examine the relationship between different

types of proximity in the labour force and the competitiveness of firms. The rationale for the micro-level focus is the assumption that interactive learning processes within plants occur under different conditions than inter-firm learning processes. Therefore, the relationship between interactive learning and the proximity dimensions is assumed to differ from that for inter-firm learning. Moreover, it is suggested that the functioning of the proximity dimensions varies between technological and organizational contexts. The workings of knowledge dynamics at the micro-level of the plant have been scrutinized in three empirical papers. The empirical findings indicate that the proximity dimensions are interconnected and that cognitive distance in the labour force is not the only factor that determines the efficiency of interactive learning processes. The geographical and organizational distances between inflowing labour and in-house labour also influence the potential for interactive learning. A third factor that impacts the effect of learning on performance is the degree of dependence on physical capital associated with sector belonging. It is demonstrated that capital-intensive sectors benefit substantially more from having a local labour inflow than a cognitively related inflow. Furthermore, it is established that relatedness is not necessary for interactive learning to occur, neither in the cognitive dimension nor in any of the other dimensions examined. Instead, it is the combined distances of multiple proximity dimensions that determine the degree of interpersonal learning.

The next section is structured around a discussion of the key findings in the papers and their implications for interactive learning in firms. The subsequent section discusses the findings in relation to the analytical framework of the proximity dimensions and evolutionary economic geography. It focuses on how we can understand firm competitiveness through analysis of the labour force. Moreover, the relationship between the findings and the spatial development of economic activities is outlined. The last section offers suggestions for future research.

5.1 Findings When the three papers are considered together, five main findings emerge.

First, as theoretically suggested by Boschma (2005) and Torre and Gilly

(2000), the proximity dimensions are interrelated and can influence one

another. It is evident that the geographical dimension influences the impact

38

of the cognitive dimension on plant performance. The observed relationship

is captured in Paper I and Paper II and is supported by previous empirical

findings by Boschma et al. (2009) and Eriksson (2011). It is also shown that

the organizational dimension influences the cognitive dimension. The

interdependence is manifested through the variable impact on plant

performance that a given distance in one dimension has, depending on what

other type of proximity is accounted for in the model. Put differently, a short

distance in one dimension may require a long distance in another dimension

in order for interactive learning to take place.

In Paper I and Paper II it is demonstrated that inflows of unrelated formal

knowledge positively impact plant performance if they are also geographically

proximate. This can be explained by the interrelationship of cumulative and

combinatorial knowledge dynamics (Strambach 2013; Strambach and

Klement 2012), which allows for heterogeneous knowledge to be combined in

a context where there is accumulation (similarity) in another dimension.

Proximity in the geographical dimension indicates a mutual understanding of

norms and values associated with a community (Farole et al. 2010) and also a

shared appreciation of, and involvement in, the localized capabilities (Maskell

and Malmberg 1999a). This finding demonstrates nicely how local labour

market flows reproduce and develop the localized capabilities by allowing for

combinations of unrelated pieces of knowledge through the accumulation of

place-specific values and norms

Proximity in the organizational dimension, which here is estimated as

intra-organizational proximity, is assumed to generate a coherent knowledge

view in the labour force. Moreover, intra-organizational proximity ensures a

shared understanding of the structures and procedures within the firm that

reduces uncertainty about the role of others (Weick and Roberts 1993). The

interdependence of the different proximity dimensions is found in all three

papers1 (Paper I, II and III). These findings suggest a twofold relational and

social aspect of knowledge interactions. Interactive learning occurs between

individuals who possess knowledge and work in specific socio-technical

contexts. Their knowledge can be quite similar, but it can also be completely

unrelated. The degree or relatedness determines how well knowledge is

understood and appreciated in knowledge interactions. This relational

distance determines the efficiency of communication and learning in groups,

and varies between any two individuals. This is the first relational aspect of

knowledge interactions. Concurrently, employees at a plant are also connected

through dimensions other than the cognitive2 dimension. Depending on the

1 Paper III examines different types of cognitive proximity, rather than different proximity dimensions.

But the same pattern is detected, where the influence of one knowledge type on plant performance depends

on what other type of knowledge is accounted for in the model. 2 E.g., the organizational, the geographical/institutional or the social dimension. The social dimension is

not examined in the thesis.

39

distance in these other dimensions, learning can be either facilitated or

complicated. This is the other aspect of the relational character of the

proximity dimensions.

The second finding demonstrates that the proximity dimensions have

conditional effects on learning and innovation in firms. The empirical findings

of the interdependence of the proximity dimensions in Paper I and Paper II

are further established and elaborated upon in Paper III, where it is

established that the impact of a given distance in one dimension varies with

the distance in another dimension. Whereas such relations have been

theoretically discussed in the literature (e.g., Boschma 2005), they have not

been demonstrated empirically before. The conditional effects are called

overlap effects. These overlap effects can either enhance the initial impact of

either variable – a positive overlap effect – or they can mitigate the initial

impact – a negative overlap effect. The empirical findings show that the

impact of high levels of in-house similarity in formal knowledge is conditional

on the level of in-house similarity in past knowledge exposures. High levels of

similarity in knowledge exposure substantially reduce the negative impact of

similarity in formal knowledge to the extent that high levels in both variables

positively impact plant performance. These findings could help explain the

slightly diverging results of previous studies on the effects of knowledge

variety on plant and firm performance.

Third, it is not only the proximity dimensions that are interrelated, but the

impact of the cognitive dimension on interactive learning is contingent on the

human capital ratio at the plant. The findings in Paper III demonstrate that it

is the combination of knowledge variety and human capital ratio at the plant

that determines the competitiveness of the plant. Previous studies have

argued that it is not human capital per se that influences the performance of

plants, but rather the degree of relatedness in the in-house labour force

(Boschma et al. 2009). However, the present findings suggest that the impact

of knowledge variety on plant performance is conditional on the level of

human capital at the plant. Other empirical studies have refrained from

estimating the impact of knowledge variety and human capital ratio in the

same model. But in Paper III the parallel impacts of knowledge variety and

human capital ratio are estimated together with their conditional impacts on

plant performance. It is argued that human capital represents ability, whereas

knowledge variety symbolizes specialized expert knowledge.

It is theoretically suggested that high levels of human capital represent high

generic abilities to communicate, learn and synthesize. It also signals high

individual abilities (Spence 1973) that are guaranteed through the successful

completion of higher education resulting in a diploma (Arrow 1973).

According to Becker (1964), the high abilities of highly educated individuals

render them more efficient in their execution of different tasks compared to

those without higher education. The degree of knowledge variety within a

40

plant, on the other hand, informs us of the potential for combining different

pieces of specialized knowledge into new knowledge. Paper III shows that the

impact of similar, related and unrelated formal knowledge (educational

background of the employees) is conditional on the level of human capital. The

negative impact on plant performance associated with high levels of similar or

unrelated knowledge is mitigated by high levels of human capital. This

suggests that interactive learning processes cannot be properly understood

and estimated if attention is not paid to both the generic abilities of the labour

force and the combined variety of their specialized knowledge. Furthermore,

as discussed above human capital represents an ability, presumably an innate

ability, whereas formal knowledge expresses the individual’s area of expertise.

Productive interactive learning processes within plants are influenced by both

the abilities of the labour force and the knowledge variety present in the plant.

The fourth finding shows that there are differences between sectors that

generate variable outcomes of learning processes. This implies that the

workings of the proximity dynamics also vary between sectors, not only

between different levels of analysis. In Paper I the role of knowledge variety

in interactive learning in capital-intensive sectors and labour-intensive

sectors, is examined. In line with previous empirical studies (Haltiwanger et

al. 1999; Leiponen and Drejer 2007; Rigby and Essletzbichler 1997), the

results illustrate a strong dependence on localized capabilities in the capital-

intensive sectors, where only local labour inflows positively influence plant

performance. The regional variations in production technology reinforce and

develop the localized capabilities, by allowing local inflows of unrelated

knowledge to be absorbed into the plant. Furthermore, it is established that

the overall impact of both in-house knowledge variety and inflowing

knowledge variety is smaller in the capital-intensive sectors than in the

labour-intensive sectors. The implications are that there are structural

differences between capital-intensive and labour-intensive sectors. This may

be explained, at least partially, by the dependence on physical capital

(machinery), which is produced externally, in the capital-intensive sectors.

This dependence reduces the number of innovative opportunities, because

process innovations are mainly the result of supplier innovations (Lagerholm

2007). Moreover, the physical capital used in the production of goods may

restrict the potential to apply new knowledge, generated from interactive

learning processes. The possibility to apply knowledge is presumed to

completely depend on firm routines and the technologies used within the firm.

This is a step that is often disregarded in economic geography, where

interactive learning processes, in and between firms, are treated as generating

equally applicable knowledge regardless of the sectoral and organizational

context. Both the technology used in an industry and the firm routines within

a specific firm can be viewed as structures that either hinder or promote

41

effective absorption of inflowing knowledge and the possibility to apply new

knowledge.

Lastly, the findings suggest that cognitive relatedness is not necessary for

interactive learning to occur on the micro-level of the plant. Paper I

demonstrated that the combination of either geographical proximity and

cognitive distance or geographical distance and cognitive proximity generated

significantly stronger positive impacts on plant performance than cognitive

relatedness did. In Paper II organizational and geographical proximity

allowed for unrelated knowledge to be efficiently absorbed and applied in the

plant. Within plants, interactions take place between individuals on a daily

basis. Under such circumstances, it is more important to share a short

distance in one dimension, than it is to be cognitively related. This implies that

trust and social structure are more important to learning than is sharing

formal pieces of knowledge. This is in contrast to the agglomerations literature

and the RIS literature. With regard to externalities and knowledge spill overs

between co-located firms, Neffke et al. (2012) demonstrated that a related

industry setup in the region substantially increases the survival rate of plants.

Similarly, Boschma and Iammarino (2009) show that technological

relatedness in the regional industry base positively impacts export rates. The

presumed reasons for the positive effects of a related local industry setup or

RIS are explained by supporting structures and knowledge accessibility

(Trippl 2011).

In cases of mergers and acquisitions or temporary projects between firms,

relatedness is a necessary condition for the joint ventures to be initiated and

for innovation to occur. The temporal character of such case and the potential

lack of any other proximity necessitate cognitive relatedness (Boschma 2005).

This is in contrast to the inter-individual level and labour mobility. In such

scenarios, one or a few individuals enter a context in which they are to stay for

an indefinite period of time. They need to be “accepted” by the receiving group

– the in-house labour force – and any type of proximity relation will speed up,

and ease, the integration of the inflow. Once accepted, the individual will be

able to communicate more distant pieces of knowledge, with regard to the

existing in-house composition. Thus, on the interpersonal level all proximity

dimensions seem to have mainly a socializing function, including the cognitive

dimension, whereas for cooperation between firms, relatedness in the

cognitive dimension is probably the reason why the joint project came to be in

the first place, and the other proximity dimensions can hinder or promote the

learning and innovation processes. The present results indicate that the

workings and effects of proximity dynamics are different at different levels of

analysis.

42

5.2. Discussion The above findings show that in order to explain the uneven spatial

development of economic activities it is crucial to understand how knowledge is produced and reproduced in firms. Since learning and innovation occur in plants and firms rather than regions we need to understand the interplay between internal processes and the external environment. In the present thesis it has been demonstrated that sector belonging, firm routines, and the knowledge resources in the firm interact to generate unique conditions for interactive learning and innovation. Firms and regions reproduce increasingly specialized knowledge in order to sustain a competitive advantage (Howells 2012), while utilizing the accumulated knowledge resources in the firm and region. A micro-perspective on knowledge offers significant empirical evidence that distributed knowledge production is an essential element in the learning economy. This means that knowledge is becoming increasingly spatially fragmented, due to the local or regional reproduction of knowledge. One major reason for this is the labour force and its relative fixity in space that makes it a mainly local resource. Labour force knowledge is fostered in a socio-technical context in which formal knowledge and skills are integrated with social and institutional structures that affect the view on knowledge and knowledge application (e.g., Maskell and Malmberg 1999a; Morgan 2004; Storper and Venables 2004; Strambach and Klement 2012).

The findings further suggest that for learning to occur on the interpersonal

level, it is more important to share a bond than to be cognitively related. The

interrelationship between proximity and distance that has been suggested

here is in contrast to much of the existing empirical literature on proximity

dynamics (e.g., Boschma et al. 2009). It is found that relatedness in the

cognitive dimension is not unambiguously positive for interactive learning

and innovativeness. Having similarity in one dimension and unrelatedness in

the cognitive dimension has a significantly stronger impact on interactive

learning than simply having relatedness in the cognitive dimension. It

therefore seems as if the combined distance of several proximity dimensions

should be taken into account when estimating the innovative power of a firm

or industry. The empirically demonstrated interdependence of different types

of proximities may help to explain the somewhat diverging empirical results

of the relationship between firm performance and knowledge variety in the

labour force. The findings also offer support for the notion of a relational

economy (Bathelt and Glückler 2003; Torre and Gilly 2000) in which

economic actors are unequally able to combine their knowledge and skills with

the knowledge and skills of other actors, depending on their level of

association.

The micro-foundation of innovation and economic performance is knowledge transferred between individuals. The findings indicate that innovation emerges as a result of cumulative and combinatorial knowledge dynamics. Whereas Crevoisier and Jeannerat (2009) argue that the cumulative aspect of knowledge in firms and regions is losing importance, it

43

is argued here, and in line with Strambach (2013), that the cumulative and combinatorial knowledge dynamics interact in the generation of novelty. It is the cumulative aspect of knowledge that has driven the theoretically motivated need for cognitive relatedness in all empirical work on learning so far. There have been two parallel logics to explain the need for cognitive relatedness in inter-firm and interpersonal learning. First, on the inter-firm level, firms are unlikely to diversify into activities that are completely unrelated to the existing competences (Harrison et al. 2001). The efficiency of such mergers, acquisitions, alliances and projects is too low to be economically justified (Rumelt 1974). Second, knowledge in a firm has accumulated over time and any additional contribution to the existing stock needs to be cognitively related in order to be recognized as relevant (Cohen and Levinthal 1990; Nooteboom 2009). But, instead of just focussing on cognitive relatedness as a bridge between accumulated knowledge and new knowledge, the present thesis has demonstrated that similarity in the organizational or the geographical/institutional dimension allows unrelated knowledge to be integrated into production processes. It has been shown, for instance, that organizationally proximate individuals successfully generate novelty by combining their unrelated formal knowledge (Paper II). Moreover, unrelated formal knowledge originating from the same institutional context (i.e., being intra-regional) contributed to the upgrading of skills in the capital-intensive sectors in Paper I. Thus, it is argued that the accumulation of social capital and firm-specific knowledge occurs on the level of the individual. This accumulation represents a certain degree of familiarity with firm routines and norms and values in a region that are shared with other individuals in the same context. This similarity represents a short relational distance between individuals that generates cohesion and reduces uncertainty on the interpersonal level. If there is similarity in any one of the dimensions, unrelated knowledge can be integrated into firm practices.

Combining unrelated pieces of knowledge is only possible between individuals who share some kind of common ground that enables trust and reduces uncertainty. In line with recent contributions in economic geography (Boschma 2005; Strambach and Klement 2012) and the reasoning underlying the ZPD (Vygotsky 1962), it is argued here that proximity and distance represent accumulation and combination and that interactive learning in the labour force is facilitated by close proximity in any one dimension. Furthermore, the findings in this thesis support the view that micro-agents (i.e., individuals in the present thesis) are embedded in (Granovetter 1985) and inseparable from the structures that surround them. Physical and cognitive structures influence and are influenced by the actions of micro-agents. Whereas structures (institutions, technology and firm routines) offer possibilities and impose restraints on agents, they are also interpreted every time an agent attempts to act accordingly. As time passes, the structures are incrementally changed and adapted by the individuals that live and work under them. The local environment generates relational proximity between agents through formal and informal networks (Storper 1997). This proximity

44

reproduces and rejuvenates localized capabilities by enabling the combination of heterogeneous pieces of knowledge in firms through local unrelated labour inflow. Thus, time and place can be perceived as the paramount dimensions that shape the micro-dynamics of knowledge generation and innovation.

The notion presented by Feldman and Kogler (2010), that places are not

equal and are defined by evolutionary processes, is empirically corroborated

in the present thesis. An important element in the evolution of economic

activities is the accumulated knowledge embodied in the labour force in a

region (Maskell and Malmberg 1999a). The impact that labour force

knowledge has on plant performance is substantial, multifaceted,

differentiated, and demonstrated through multiple relational measures of

knowledge and ability throughout the empirical papers. The findings in the

thesis support the claim by Storper and Scott (2009) that the productivity of

resources and skills depend on selective geographical matching processes. It

has been shown that the proximity dimensions are interdependent and it has

also been demonstrated that physical and cognitive structures are

interconnected with the proximity dimensions. These interdependencies lead

to path-dependent trajectories for firms and regions. These trajectories will

offer different absorptive capacities and innovative behaviours throughout the

economic landscape (Storper and Scott 2009).

The contextual and relational aspects of learning has been demonstrated in

the papers and emphasised throughout the thesis. The findings of the

relational character of knowledge offer significant support for the notion

presented by Bathelt and Glückler (2003) that economic actors are situated in

contexts of social and institutional relations. According to the knowledge and

competence-based theory of the firm (Kogut and Zander 1992; Teece et al.

1997), knowledge accumulation in firms is mainly driven by firm routines and

organizational practices. These are based on localized learning processes that

have emerged over time in specific institutional and technological contexts, of

which the labour force is a vital part. The specificity of these localized learning

processes helps explain why the behaviour and knowledge of successful firms

cannot easily be replicated or transferred to other contexts (Szulanski 1996;

Teece 2010). The existence of localized capabilities is also demonstrated

through the variable impact of knowledge inflow depending on the

geographical (institutional) origin of the labour inflow. It would seem that

geography is not dead (Cairncross 1997) and space is not neutral with regard

to the development and competitiveness of economic activities.

45

5.3. Directions for future research The present findings suggest that the conditions for learning and

knowledge application vary throughout the economic landscape. Therefore,

interactive learning processes and innovation should be examined within the

context of the firm and its technological trajectory and institutional

environment. Future studies on interactive learning and knowledge dynamics

should therefore consider the structures that surround the labour force and

the knowledge variety that the labour force constitutes. A further division of

sectors based on the structures that their technological trajectories dictate

could advance our understanding of how knowledge generation is impeded

and promoted in different contexts. The reciprocal relationship between

structures and agents suggests that the role of knowledge variety in firm

competitiveness should be regarded as one part of a greater system. Thus,

future research aimed at understanding local conditions for competitiveness

and growth could benefit from the integration of meso- and micro-level

analyses.

Another interesting direction for future research is to further detangle the

interdependencies and conditional influences of the proximity dimensions. In

Paper III it was shown that different types of cognitive proximity have

conditional impacts on plant performance. These findings indicate that

conditional effects may also be found between other types of proximity. This

would require the development of continuously integrated frameworks where

multiple types of proximity are analysed simultaneously. Furthermore, the

trend away from appreciating “just” high levels of human capital may be

premature due to the conditional effects of human capital and knowledge

variety. Thus, future studies on the influence of labour force knowledge on

competitiveness should integrate ability (human capital) and knowledge into

the analysis. This effect is also likely to vary between different sectors.

Moreover, the present findings indicate that the role of different proximity

dimensions may vary at different levels of analysis. For example, cognitive

relatedness is not found to be necessary for knowledge generation on the

interpersonal level of intra-firm or intra-plant learning. On the other hand, in

discussions of inter-firm relations, Boschma (2005) and others argue that

cognitive relatedness is a necessary but not sufficient requirement. Further

analytical clarity about the workings of the proximity dynamics on different

levels of analysis is therefore needed. In relation to that issue, there is a need

to continue to develop empirical tools for estimating proximity in the different

dimensions on the different analytical levels.

Our understanding of the external mechanisms that impede and/or facilitate learning (i.e., the firm and its routines, the sector and the region) should be further looked into, preferably in an increasingly integrated framework that considers multiple factors that affect learning and the innovative capabilities of firms or regions. The evolutionary framework offers

46

great opportunities for integrating various levels of study as well as different mechanisms that influence the potential for learning and innovation. Regardless of level of analysis, the passage of time leaves the observed entity or place with events and experiences that impact the ability to adapt to new circumstances in the future. Therefore, differences in the economic performance of firms, sectors and regions should be perceived as varieties of accumulated paths on several interconnected levels. Through more extensive research on each of the above-mentioned mechanisms thought to obstruct and/or promote knowledge transfer, we will eventually come closer to a complete synthesis of how and why differences in innovative capabilities exist.

6. Sammanfattning (Summary in Swedish)

I den här avhandlingen undersöks förutsättningar för lärande och

tillämpning av kunskap (innovation) i företag. I ekonomisk geografi och

besläktade ämnen har kunskap kommit att betraktas som den viktigaste

faktorn för företags och regioners konkurrenskraft. Det beror på att kunskap

är lokalt förankrad till skillnad från materiella produktionsfaktorer som kan

tillgängliggöras genom transporter över hela världen. Genom

socialiseringsprocesser kopplade till såväl platsers bransch- och

företagsstruktur som deras historia tolkas, används och reproduceras

kunskap på olika sätt på olika platser. Det handlar inte huvudsakligen om att

varor och tjänster produceras på olika platser utan snarare om att likartade

produkter produceras på olika sätt på olika platser. Den lokala kunskapen styr

hur användningen av andra resurser (produktionsfaktorer) utformas och det,

i sin tur, påverkar framtida förutsättningar för förändring och nytänkande i

såväl enskilda företag som i regioner.

Innovationskraften - förmågan att förnya produkter, processer, styrning

och marknadsföring m.m. - anses avgörande för bibehållen eller utökad

produktivitet för företag och regioner. Således blir den tillgängliga kunskapen

och förmågan att förnya den av yttersta vikt. Individer representerar olika

kunskap, vars kvalitet och kvantitet bedöms och avgörs av sammanhanget de

befinner sig i, dvs. av de andra deltagande individerna. Att skaffa sig en

konkurrensfördel genom arbetskraften handlar därför inte om att ackumulera

expertkunskap inom ett område utan snarare om att sammanföra individer

med varierad kunskap som kan kommunicera med, och lära av, varandra. På

så vis skapas en dynamisk miljö med förutsättningar för fortsatt lärande och

förnyelse som i förlängningen kan ge stärkt konkurrenskraft. Individers och

gruppers lärande är förvisso en nödvändighet, men är i sig självt otillräckligt

för att initiera en innovationskedja. För att kunskapen ska kunna omsättas i

praktik krävs en närvaro av fysiska och kognitiva strukturer som tillåter detta.

Dessa strukturer varierar mellan såväl branscher som företag och påverkar i

47

hög grad möjligheten för enskilda individer att tillämpa ny kunskap i

produktionen. Det är således inte enbart kunskapssammansättningen som är

avgörande för företagens produktivitet utan även regional bransch- och

företagsspecifika strukturer som påverkar möjligheten att omsätta

kunskapen.

Utgångspunkten i den här avhandlingen är arbetskraftens kunskap och den

variation i kunskap som finns på arbetsplatser och i deras närmiljö.

Arbetskraftens individuella och samlade förmågor utgör grunden för lärande.

Således är även de potentiella tillämpningarna (omsättningen av teori i

ekonomiskt betydelsefull praktik) av ny kunskap i företag ytterst beroende av

arbetskraftens kunskap och förmågor. Men för att implementera kunskap och

ge den ett ekonomiskt värde räcker det inte med att kunskapen existerar, utan

det måste även i företaget och branschen finnas inkluderande och flexibla

strukturer som tillåter enskildas eller gruppers teorier att bli praxis. Mycket

av den befintliga forskningen fokuserar antingen på kunskapens betydelse för

företagens produktivitet eller på de övergripande strukturernas hindrande

alternativt befrämjande egenskaper för innovationskraften. Få studier

integrerar de två perspektiven, vilka kan sägas vara olika analysnivåer av

samma fenomen, innovationskraft och konkurrenskraft. Att försöka fånga och

kartlägga interaktionen, mellan kunskapssammansättning och strukturer är

denna avhandlings huvudsakliga syfte.

Syftet uppnås genom tre artiklar som använder sig av kvantitativa

analysmetoder för att belysa olika aspekter av arbetskraftens kunskap i

relation till företagens produktivitet och konkurrenskraft. Samtliga studier

utgår från formella mått på arbetskraftens kunskap (t.ex.

utbildningsinriktning eller yrkestitel) och den variation av kunskap som finns

på arbetsplatsen. Dessa mått relateras sedan till andra strukturer som varierar

mellan arbetsplats, så som fysiska strukturer kopplade till sektoriella

skillnader, kognitiva strukturer i form av företags skilda rutiner och

professionella nätverks effekt på arbetskraftens faktiska förmågor.

Analyserna bygger på detaljerad, longitudinell socioekonomisk data ifrån

statistikdatabasen ASTRID som finns på institutionen för geografi och

ekonomisk historia vid Umeå universitet. ASTRID innehåller data från

Statistiska Centralbyrån och information finns på årsbasis om samtliga

företag, arbetsplatser och individer i Sverige.

Syftet med den första artikeln, Paper I, är att testa huruvida det finns

systematiska skillnader i möjligheten att omsätta kunskap till praktik mellan

två stora sektorer i ekonomin, nämligen den kapitalintensiva- och den

arbetskraftsintensiva sektorn. Genom att analysera effekten av såväl

sammansättningen av kunskap på arbetsplatsen som inflödet av ny

arbetskraft till arbetsstället, studeras både kunskapssammansättningens

betydelse för arbetsplatsens produktivitet och de skilda möjligheterna att

integrera inflödande kunskap med befintlig kunskap inom sektorerna.

48

Artikeln bygger dels på litteratur som diskuterar hur produktionsmetoder och

produktionsfaktorer genererar strukturella skillnader i

innovationsmöjligheter mellan branscher, och dels på litteratur om

arbetskraftsrörlighet och värdet av relaterad kunskap för företags

produktivitet. Resultaten i den första artikeln visar att det är gynnsamt för alla

arbetsplatsarbetsplatser, oavsett sektor, att ha anställda vars formella

kunskaper är besläktade (related) men inte identiska (similar). Den positiva

effekten är dock större i arbetskraftsintensiva sektorer än i kapitalintensiva

sektorerna. Den svaga positiva effekten av inflödande besläktad kunskap i de

kapitalintensiva sektorerna antas bero på att kunskap i stor utsträckning

ackumuleras i det fysiska kapitalet, vilket är kostsamt att förändra och

utveckla. I de arbetskraftsintensiva sektorerna lagras majoriteten av

kunskapen i arbetskraften och utbyte och utveckling av idéer och kunskap kan

ske genom den dagliga interaktionen.

Kunskapsinflödenas effekt på arbetsplatsens produktivitet skiljer sig åt

mellan sektorerna vad gäller såväl kunskapens sammansättning (identisk,

besläktad eller orelaterad) som geografiska ursprung (lokalt eller icke-lokalt).

Effekten av arbetskraftsinflöden, tillika kunskapsinflöden, är genomgående

större för den arbetskraftsintensiva sektorn än för den kapitalintensiva. Det

förklaras med ett beroende av fysiskt kapital (maskiner och utrustning) i de

kapitalintensiva sektorerna. Det innebär att potentiella effektiviseringar och

processinnovationer delvis kommer till stånd hos underleverantörer och

tillverkare. Således är antalet innovationsmöjligheter färre i de

kapitalintensiva sektorerna än i de arbetskraftsintensiva sektorerna där såväl

produkt- som processinnovationer sker inom företaget. Inom de

kapitalintensiva sektorerna är det endast de lokala inflödena som har en

positiv effekt på arbetsplatsens produktivitet. Det är i enlighet med tidigare

forskning som funnit att produktionstekniken inom en och samma bransch i

tillverkningsindustrin varierar kraftigt mellan olika platser. Det visar på en

socialiseringsprocess av kunskapstillämpning som innebär att i de

kapitalintensiva sektorerna är kunskap om hur saker görs och kommuniceras

på en plats av större betydelse än den formella kunskapen en individ besitter

och kan bidra med i produktionen. I de arbetskraftsintensiva sektorerna finns

det kombinerade effekter av inflödets kunskapssammansättning och

geografiska ursprung som gör att såväl lokal orelaterad kunskap som icke-

lokal identisk kunskap har en positiv effekt på produktiviteten. Resultaten för

de arbetskraftsintensiva branscherna visar att det finns andra aspekter av

arbetskraftens kunskap, än den formellt kognitiva förmågan, som påverkar

förmågan att kommunicera och implementera kunskap på arbetsplatsen. I

branscher där människor, tillika sociala kunskapsbehållare, utgör den

huvudsakliga produktionsfaktorn förekommer en större känslighet för

kunskapens geografiska ursprung. Det lokala- respektive icke-lokala inflödet

av arbetskraft representerar institutionellt lika och olika inflöden. När hänsyn

49

tas till såväl formell kunskap som det geografiska (institutionella) ursprunget

framträder en bild där kombinationen av likhet och olikhet i dessa

dimensioner genererar ett sammantaget intellektuellt avstånd som gynnar

lärande och innovationsförmågan på arbetsplatsen. De samlade resultaten i

den första artikeln indikerar att förutsättningarna för absorption och

tillämpning av kunskap skiljer sig åt mellan de två stora sektorsgrupperna.

Den andra artikeln, Paper II, handlar om hur företag socialiserar och

strukturerar kunskap genom sina rutiner och på det viset skapar en praxis som

är unik för företaget. Denna praxis skapar en organisatorisk närhet och

samhörighet mellan de anställda som tillåter effektivare kommunikation

mellan individer inom företaget än mellan individer på olika företag.

Rutinerna skapar således olika förutsättningar för de anställda att tillvarata

intern och extern kunskap. Syftet med artikeln är att jämföra effekten på

produktiviteten av interna kunskapsflöden (individer som byter arbetsplats

inom ett och samma företag) med externa kunskapsflöden (individer som

byter arbetsplats och företag). Interna och externa flöden ställs mot varandra,

samtidigt som hänsyn tas till arbetskraftens formella

kunskapssammansättning och dess geografiska ursprung. Liksom företagen

strukturerar och socialiserar kunskap gör även platser (regioner) det genom

den ackumulation av tekniskt kunnande och sociala nätverk som byggts upp i

regionen över tid. Denna samsyn kring teknisk utveckling och kunskap som

finns i regioner, sägs utgöra en del av regionens institutionella grund. Denna

formar såväl branschstrukturen i regionen som tillämpningen av kunskap

inom en specifik bransch och representerar i denna avhandling normer och

värderingar. Dessa är huvudsakligen grundade i regionens historiska

utveckling av branscher och företag. Bilden av vad som är relevant kunskap

förstärks alternativt försvagas av samarbeten mellan branscher eller mellan

företag som historiskt sett varit starka och framgångsrika. Genom att

inkludera den geografiska dimensionen i analysen framträder en bild av hur

den formella kunskapssammansättningen samspelar med såväl rumsligt som

organisatoriskt ursprung för att skapa förutsättningar för lärande och

innovationskraft.

Resultaten i denna artikel visar att inflöden av kunskap som liknar den som

finns på arbetsplatsen endast ökar produktiviteten om den samtidigt

härstammar från att annat företag med annan praxis. Den formella

kunskapen har då tillämpats inom en annan fysisk och social struktur och

eventuellt även på ett annat sätt. Den är således differentierad från den

befintliga kunskapen på den nya arbetsplatsen i ett organisatoriskt

hänseende. Detta ger den likartade kunskapen ett ytterligare djup och bredd,

vilket innebär potentiellt fler nya kombinationer och tillämpningar av

kunskap. Den positiva effekten på produktiviteten är som snabbast och mest

omfattande när det likartade kunskapsflödet härstammar från en annan

region. Det tolkas som att ett större avstånd i såväl den geografiska- som den

50

organisatoriska dimensionen skapar en stor bredd av möjliga nya

tillämpningar av den relaterade formella kunskapen. Resultaten visar också

att bekantskap med praxis (intra-organisatoriska flöden) skapar en

sammanhållen kunskapssyn. Även om individens formella kunskap är

likartad snarare än identisk med den befintliga kunskapen på den nya

arbetsplatsen så sker inget nyskapande. Det krävs att den inflödande formella

kunskapen är olik den befintliga samt att den är lokal (intraregional), och

således har samma socio-tekniska ursprung, för att ny kunskap och nya

tillämpningar ska genereras. Detta visar på en starkt konvergerade effekt av

platsspecifika värderingar på kunskap. Det indikerar även att intra-

organisatorisk närhet skapar förtroende mellan de deltagande individerna på

ett sådant sätt att mer avlägsna idéer kan mötas och diskuteras och potentiellt

implementeras. I Paper II framgår det att avstånd kan mätas på fler sätt än

som fysiskt (geografiskt) avstånd och att dessa icke-rumsliga relationella

avstånd påverkar lärande inom företag. Genom närhet i någon eller fler av

dessa dimensioner kan förtroende skapas ur vilket ny kunskap och nya

tillämpningar kan växa fram. Likartad eller identisk formell kunskap är inte

alltid nödvändig för att lärande och innovationskraft ska uppstå. Resultaten

visar att lärande och innovationsprocesser är sociala processer och som

sådana grundas de på tillit och trovärdighet, något som kan uppnås genom

närhet i den organisatoriska och den geografiska dimensionen. Huruvida det

finns någon hierarkisk struktur mellan olika former av närhet, för lärande, går

inte att säga utifrån resultaten i denna artikel. Men, det har framkommit att

korta avstånd i en socialiserande och strukturerande dimension kan

kompensera för stora avstånd i den formella kunskapsdimensionen.

I den tredje artikeln, Paper III, testas effekten av arbetskraftens kunskap i

företag inom kunskapsintensiv företagsservice (KIBS). Medan de två första

artiklarna fokuserar på effekten av inflöden av arbetskraft på arbetsplatserna

så fokuserar denna artikel på den befintliga arbetskraften på en arbetsplats.

Till kunskapsintensiv företagsservice räknas bl.a. juristfirmor, arkitektkontor,

olika IT-tjänster och forskning m.m. Dessa branscher genomsyras av lärande

och föränderlighet då de huvudsakligen säljer expertkunskap och arbetar på

uppdragsbasis mot kunder vars behov varierar. Kunskapens och lärandets roll

förväntas således vara betydande för företagens konkurrenskraft. Dessa

branscher lämpar sig därför extra bra för studier av kunskapsstrukturer,

lärande och innovationskraft. Utöver att undersöka vilken effekt den formella

kunskapen och dess sammansättning har på innovationskraften i dessa

företag så presenteras och testas en teori om att kunskap hos individer kan

mätas på fler sätt än som formell utbildning. Genom sina

arbetslivserfarenheter kommer människor i kontakt med såväl olika

branscher som arbetsplatser vars medarbetare har olika

utbildningsbakgrund. Dessa möten och erfarenheter formar individens

kunskap och förståelse av hur kunskap kan användas. Både branscherfarenhet

51

och tidigare kunskapsexponering betraktas som egna kunskapsvariabler.

Effekten på lärande och innovationskraft på arbetsplatsen beräknas för de

olika typerna av kunskap dels var för sig, dels i olika kombinationer. Detta för

att testa huruvida förekomsten av en viss typ av kunskap påverkar effekten

som en annan kunskap har på lärandet. Grundantagandet i artikeln är att

kunskap kan ta sig olika uttryck och således även bör mätas på olika sätt. Det

anses även rimligt att olika kunskapsstrukturer påverkar varandra. Detta kan

ske på ett sådant sätt att den initiala effekten av en form av kunskap på

företagets produktivitet förändras i kombination med en annan kunskapstyp.

Tidigare branscherfarenhet representerar tillämpad kunskap. Därför är det

rimligt att liknande (relaterade) branscherfarenheter hos medarbetarna på en

arbetsplats skapar förtrogenhetförtrogenhet med multipla tillämpningar av

en viss typ av formell kunskap. Likaså vittnar tidigare kunskapsexponering

om förmågan att absorbera och kommunicera kunskap. Att medarbetarna har

liknandeliknande erfarenheter av kunskapsexponering skapar förtroende för

kompetensen genom igenkänning.

Avsikten med artikeln är dels att nyansera måttet på kunskap, och dels att

testa hur kunskapssammansättningen på en arbetsplats påverkar dess

produktivitet. I teoretiska diskussioner lyfts kunskap fram som relationell och

socialt formad. Detta till trots mäts den oftast som utbildningsbakgrund i

studier av kunskapens effekter på företagens produktivitet. Artikeln

undersöker hypotesen att individer kan skaffa sig likartad kunskap genom

arbetslivserfarenheter likväl som genom utbildningsbakgrund. Resultaten

visar att företags produktivitet påverkas positivt av att ha en hög andel

anställda med högre utbildning. Med tanke på att företag inom

kunskapsintensiv företagsservice huvudsakligen säljer och producerar

kunskap är detta inte särskilt förvånande. Kunskap anses driva såväl teknisk

utveckling som ekonomiskt tillväxt i företag och regioner i alla branscher.

Andelen högutbildade representerar det humankapital som ackumulerats på

en plats, i detta fall en arbetsplats. Men när humankapitalnivån på

arbetsplatsen ställs mot olika sammansättningar av kunskap träder en mer

nyanserad bild fram. Det framgår att effekten av humankapital och de olika

formerna av kunskapssammansättning påverkar varandravarandra, dvs. de är

villkorade av varandra. Detta är något som tidigare forskning fört fram som

troligt men det har hittills inte testats. Resultaten visar att den negativa effekt

som höga nivåer av identisk och olik formell kunskap (avseende

utbildningsinriktning bland de anställda) har på lärande avtar när

humankapitalnivån ökar. Höga nivåer av såväl identisk formell kunskap som

humankapitalhumankapital genererar, till och med, produktivitetsfrämjande

lärprocesser inom KIBS. I tidigare studier av arbetskraftens kunskap och dess

effekt på företags innovationskraft har kunskapstyperna testats var för sig.

Identisk kunskap har då alltid varit skadlig för innovationskraften. Genom att

i denna artikel testa fler typer av kunskap samtidigt framgår det att effekten

52

av identisk kunskap är villkorad av utbildningsnivån på arbetsplatsen. Detta

är något som tidigare forskning inte lyckats visa.

När resultaten av de tre artiklarna beaktas i sin helhet, växer det fram en

bild av att lärande i företag är socialt snarare än kognitivt betingat. Lärande

drivs och styrs av såväl fysiska och kognitiva strukturer kopplade till specifika

branscher, som av förmågan hos de individer som deltar i aktiviteten. I denna

avhandling fokuseras på individernas förutsättningar till kollektivt lärande

och innovationskraft på arbetsplatsen genom analys av rumsliga och icke-

rumsliga avstånd mellan individerna. Avhandlingen bidrar på tre sätt till att

utveckla befintliga teorier om kunskap som resurs och effekter av rumslig och

icke-rumslig närhet på lärande. Det första bidraget är att avhandlingens

samtliga studier visar att de olika formerna av närhet samspelar för att skapa

förutsättningar för lärande och innovationskraft. Inget universellt fungerande

optimalt avstånd har framträtt i någon dimension utan studierna har snarare

visat att lärande på arbetsplatser är starkt kontextberoende. Ett kort avstånd

(närhet) i en dimension kan kompensera för ett långt avstånd i en annan

dimension. Effekten av avstånden skiljer sig åt mellan sektorsgrupper, vilket

påvisats i Paper I, liksom mellan inflöden av arbetskraft med olika kännedom

om, och förtrolighet med, företagets rutiner (Paper II). En fördjupning och ett

ytterligare bidrag, nås i Paper III där villkorade effekter (interaktionseffekter)

påvisas mellan olika former av kunskap. Dessa effekter benämns

överlappande lärandeeffekter (overlap effects). Detta betyder att den

påverkan som ett visst avstånd i en variabel har på lärande på arbetsplatsen

förändras med olika värden i en annan kunskapsvariabel. I artikeln används

fyra olika mått på kunskap varav tre är relationella avståndsmått, dvs.

varianter av kognitivt avstånd/närhet. De kunskapsavstånd som är

opåverkade av förekomsten av andra kunskaper och avstånd benämns

oberoende lärandeeffekter. Dessa resultat, de villkorade effekterna, är

sannolikt applicerbara även på övriga närhetsdimensioner (social,

organisatorisk, geografisk och institutionell närhet). Att korta avstånd i en

dimension inte bara kompenserar för långa avstånd i en annan dimension

utan också att effekten av en kunskapsform på lärande varierar med nivån på

andra kunskapsvariabler har såvitt jag vet aldrig testats och påvisats tidigare.

Dock har sannolikheten att sådana samband finns påtalats, liksom behovet av

att empiriskt testa interaktionen mellan de olika dimensionerna.

Det tredje bidraget, som har en mer implicit karaktär bygger på slutsatsen

att likhet i formell kunskap inte är nödvändig för att lärande ska äga rum

mellan individer. Det tycks viktigare, för att lärande ska uppnås, att det finns

närhet mellan individerna i någon dimension. Denna närhet skapar

förtroende mellan individerna, vilket skapar förutsättningar för att koppla

samman olika kunskap. Således är det sammantagna avståndet i flera

dimensioner av större betydelse för lärande på en arbetsplats, än vad det är

att enbart det kognitiva (kunskaps) avståndet är det rätta. Resultaten i

53

avhandlingen visar att innovationskraft kan uppstå på arbetsplatser med höga

nivåer av olik (unrelated) kunskap, dvs. stora avstånd i den kognitiva

dimensionen. Det kan vid en första anblick tyckas kontraintuitivt och

framförallt som att det står i kontrast till den befintliga litteraturen om icke-

rumsliga avstånd (t.ex. Boschma 2005). Det framhålls i litteraturen att det

teoretiska ramverket med rumsliga och icke-rumsliga avståndsdimensioner

kan användas för analyser på flera olika nivåer. Detta är än så länge främst ett

teoretiskt och generellt antagande. Men med allt fler empiriska studier av

kunskap, lärande och icke-rumsligt avstånd utkristalliseras skillnader i

effekter av närhet mellan individer och mellan företag. På mikronivån, den

nivå som studerats i denna avhandling, bidrar korta avstånd i samtliga

närhetsdimensioner med att skapa förtroende och förtrogenhet mellan

individer. En arbetsplats utgör en specifik kontext som medarbetarna är väl

bekanta med. Ibland kommer det nya medarbetare, vilkas kunskap bör

tillvaratas, införlivas i rådande rutiner och omsättas praktiskt. För att den nya

kunskapen ska kunna absorberas och tillämpas krävs att de befintliga

medarbetarna kan bedöma trovärdigheten hos den nya medarbetaren. Den

eller de nya medarbetarna kommer till en etablerad arbetsplats med

existerande rutiner och kunskapsstrukturer. Absorptionen och assimileringen

av den nya kunskapen underlättas av att nya individerna delar erfarenheter i

någon närhetsdimension med de befintliga medarbetarna. Då inflöden av

arbetskraft till en arbetsplats inte är tillfälliga samarbeten i vare sig tid eller

rum utan ofta permanenta lösningar, så är villkoren som omger lärandet i

dessa situationer annorlunda från dem som gäller vid samarbeten mellan

företag kring en produkt eller en process. De skiljer sig även från de processer

som genererar lärande i form av agglomerationsfördelar. Den kognitiva

dimensionen utgör i dessa situationer en dimension inom vilken ett kort

avstånd kan bidra till att skapa en gemensam nämnare med andra

medarbetare. Finner man en gemensam nämnare i någon dimension ökar

förtroendet för individen och dennes kunskap kan troligtvis absorberas även

om den är olik den befintliga kunskapen på arbetsplatsen.

54

7. References Abernathy, W., and Utterback, J. (1978) Patterns of innovation in

technology. Technology Review 80:40-47.

Acha, V., and Brusoni, S. (2003) Complexity and industry evolution: New

insights from an old industry, in Best Papers Proceedings of 2002. European

Association for Evolutionary Political Economy Conference, 2003.

Agrawal, A.; Cockburn, I.; and Mchale, J. (2006) Gone but not forgotten:

knowledge flows, labor mobility, and enduring social relationships. Journal

of Economic Geography 6:571–591.

Almeida, P., and Kogut, B. (1999) Localization of knowledge and the

mobility of engineers in regional networks. Management Science 45:905–917.

Almeida, P.; Song, J.; and Grant, R. M. (2002) Are firms superior to

alliances and markets? An empirical test of cross-border knowledge building.

Organization Science 13:147-161.

Amin, A., and Cohendet, P. (2005) Geographies of knowledge formations

in firms. Industry and Innovation 12:465-486.

Anderson, L., and Tegsjö, B. (2006) En rörlig arbetsmarknad – dynamiken

bland job, individer och företag [Labour market mobility – dynamics among

jobs, individuals and business]. In Statistics Sweden: Fokus på närlingsliv och

arbetsmarknad våren 2006 [Focus on economy and labor market, spring

2006]. Statistics Sweden (SCB), Stockholm.

Antonelli, C. (2000) Collective knowledge communication and innovation:

the evidence of technological districts. Regional Studies 34:535-547.

Arrow, K. (1973) Higher education as a filter. Journal of public economics

2:193-216.

Asheim, B., and Coenen, L. (2005) Knowledge bases and regional

innovation systems: Comparing Nordic clusters. Research Policy 34:1173-

1190.

Asheim, B., and Gertler, M. (2005) The geography of innovation: Regional

Innovation Systems, in: J. Fagerberg, D.C. Mowery, and R.R. Nelson (eds):

The Oxford handbook of innovation. Oxford: Oxford University Press, pp.

291-317.

Asheim, B.; Smith, H.; and Oughton, C. (2011) Regional Innovation

Systems: Theory, empirics and policy. Regional Studies 45:875-891.

Barras, R. (1986) Towards a theory of innovation in services. Research

Policy 15:161-173.

Barras, R. (1990) Interactive innovation in financial and business services:

The vanguard of the service revolution. Research Policy 19:215-237.

Bathelt, H., and Glückler, J. (2003) Toward a relational economic

geography. Journal of Economic Geography 3:117-144.

55

Bathelt, H., and Glückler, J. (2005) Resources in economic geography:

from substantive concepts towards a relational perspective. Environment and

Planning A 37:1545-1563.

Bathelt, H., and Glückler, J. (2013) Institutional change in economic

geography. Progress in Human Geography Online First Version of Record -

Oct 18, 2013:1-24

Bathelt, H.; Malmberg, A.; and Maskell, P. (2004) Clusters and knowledge:

local buzz, global pipelines and the process of knowledge creation. Progress

in Human Geography 28:31-56.

Bathelt, H., and Taylor, M. (2002) Clusters, power and place: inequality

and local growth in time-space. Geografiska Annaler B 84:93-109.

Becker, G. (1962) Investment in Human Capital: A Theoretical Analysis.

Journal of Political Economy 70:9-49.

Becker, G. (1964) Human capital. New York: Columbia University Press.

Becker, M. C. (2004) Organizational routines: A review of the literature.

Induatrial and Corporate Change 13:643-677.

Becker, M. C., and Zirpoli, F. (2008) Applying organizational routines in

analysing the behaviour of organizations. Journal of Economic Behavior and

Organization 66:128-48.

Bell, D. (1973) The coming of post-industrial society. New York: Basic

Books.

Bhaskar, R. (1978) A realist theory of science. Brighton: Harvester Press.

Bienkowska, D. (2007) Arbetskraftens rörlighet och klusterdynamik: En

studie av IT- och telekomklustren i Kista och Mjärdevi [Labour mobility and

cluster dynamics. A study of ICT clusters in Kista and Mjärdevi, Sweden].

PhD-thesis, Geografiska regionstudier 75, Depertment of Social and

Economic Geography, Uppsala University, Uppsala.

Björklund, A.; Edin, P-A.; Holmlund, B.; and Wadensjö, E. (2000)

Arbetsmarknaden, (The labor market), Second edition. Kristianstad:

Kristianstads boktryckeri.

Blanc, H., and Sierra, C. (1999) The internationalization of R&D by

multinationals: a trade-off between external and internal proximity.

Cambridge Journal of Economics 23:187-206.

Boschma, R. (1999) Learning and regional development. GeoJournal

49:339-343.

Boschma, R. A. (2004) Competitiveness of regions from an evolutionary

perspective. Regional Studies 38:1001-1014.

Boschma, R. A. (2005) Proximity and Innovation: A Critical Assessment.

Regional Studies 39:61-74.

Boschma, R.; Eriksson, R.; and Lindgren, U. (2009) How does labour

mobility affect the performance of plants? The importance of relatedness and

geographical proximity. Journal of Economic Geography 9:169-90.

56

Boschma, R.; Eriksson, R.; and Lindgren, U. (2014) Labour market

externalities and regional growth in Sweden: The importance of labour

mobility between skill-related industries. Regional Studies DOI:

10.1080/00343404.2013.867429.

Boschma, R., and Frenken, K. (2006) Why is economic geography not an

evolutionary science? Towards an evolutionary economic geography. Journal

of Economic Geography 6:273-302.

Boschma, R., and Labooy, J.G. (1999) Evolutionary economics and

economic geography. Journal of Evolutionary Economics 9:411-429.

Boschma, R., and Martin, R. (2007) Editorial: Constructing an evolutionary

economic geography. Journal of Economic Geography 7:537-548.

Boschma, R., and Martin, R. (2010) (eds) The Handbook of Evolutionary

Economic Geography. Cheltenham: Edward Elgar.

Boschma, R., and Iammarino, S. (2009) Related variety, trade linkages and

regional growth in Italy. Economic geography 85:289-311. Brenner, T., and Broekel, T. (2011) Methodological Issues in Measuring

Innovation Performance of Spatial Units. Industry and Innovation 18:7-37.

Breschi, S., and Lissoni, F. (2009) Mobility of skilled workers and co-

invention networks: an anatomy of localized knowledge flows. Journal of

Economic Geography 9:439-68.

Bush, V. (1945) Science: The endless frontier. Washington: United States

Government Printing Office.

Butzin, A., and Widmaier, B. (2012) The study of time-space dynamics of

knowledge with innovation biographies. Working Papers on Innovation and

Space 07.12. Philipps University, Marburg.

Cairncross, F. (1997) The death of distance. How the communication

revolution will change our lives. Boston: Harvard Business School Press. Castellacci, F. (2006) A critical realist enterpretation of evolutionary

growth theorising. Cambridge Journal of Economics 30:861-880. Castellacci, F. (2008) Technological paradigms, regimes and trajectories:

manufacturing and service industries in a new taxonomy of sectoral patterns of innovation. Research Policy 37: 978–994.

Cohen, M. D.; Burkhart, R.; Dosi, G.; Egidi, M.; Marengo, L.; Warglien, M.; and Winter, S. (1996) Routines and other recurring action patterns of organizations: Contemporary research issues. Industrial and Corporate Change 5:653-98.

Cohen, M. D., and Levinthal, D. A. (1990) Absorptive capacity: A new

perspective on learning and innovation. Administrative Science Quarterly

35:128-52.

Collier, A. (1994) Critical Realism – An Introduction to Roy Bhaskar’s

Philosophy. Oxfordshire: Verso.

Cooke, P. (2001) Regional innovation systems, clusters and the knowledge

economy. Industrial and Corporate Change 10:945-974.

57

Cooke, P. (2004) Regional knowledge capabilities, embeddedness of firms

and industry organization: Bioscience megacentres and economic geography.

European Planning Studies 12:625-641.

Crevoisier, O. (1999) Two ways to look at learning regions in the context of

globalization: The homogenizing and particularizing approaches. GeoJournal

49:353-361.

Crevoisier, O., and Jeannerat, H. (2009) Territorial Knowledge Dynamics:

From the Proximity Paradigm to Multi-location Milieus. European Planning

Studies 17:1223-1241.

David, P.A., and Foray, D. (1995) Accessing and expanding the science and

technology knowledge base. STI review 16:13-68.

De la Fuente, A., and Ciccone, A. (2002) Human Capital in a Global and

Knowledge-based Economy. Final Report for the DG Employment and Social

Affairs, European commission.

Dicken, P. (2007) Global shift, mapping the changing contours of the

world economy, 5th edition. London: Sage publications.

Dosi, G. (1982) Technological paradigms and technological trajectories: a

suggested interpretation of the determinants and directions of technical

change. Research Policy 11:147–162.

Dosi, G.; Faillo, M.; and Marengo, L. (2008) Problem solving and

governance in the capability-based view of the firm: the roles and theoretical

representations of organizational routines. In Handbook of Organizational

Routines, ed. M. C. Becker, 107-21. Cheltenham, U.K.: Edward Elgar

Publishing Limited.

Duranton, G., and Puga, D. (2004) Micro-foundations of urban

agglomeration economies, in Henderson J.V. and Thisse J.V. (Eds) Handbook

of Regional and Urban Economics, Vol. 4, pp. 2063–2117. North-Holland,

Amsterdam.

Eliasson, K. (2006) College choices and earnings among university

graduates in Sweden. Nationalekonomiska institutionen (Department of

Economics), Umeå University, Umeå.

Eriksson, R. (2011) Localized spill overs and knowledge flows: How does

proximity influence the performance of plants? Economic Geography 87:127-

52.

Eriksson, R.; and Lindgren, U. (2009) Localized mobility clusters: Impacts

of labour market externalities on firm performance. Journal of Economic

Geography 9:33-53.

EUROFOUND (2006) Mobility in Europe. Analysis of the 2005

Eurobarometer survey on geographical and labour market mobility. Office

for Official Publications of the European Communities, Luxembourg.

Farole, T.; Rodriguez-Pose, A.; and Storper, M. (2010) Human geography

and the institutions that underlie economic growth. Progress in Human

Geography 35:58-80.

58

Feldman, M. P., and Kogler, D. (2010) Stylized facts in the geography of

innovation, in: B. H. Hall and N. Rosenberg (Eds) Economics of Innovation.

Amsterdam: Elsevier, pp. 411–427.

Feldman, M. S. (2000) Organizational routines as a source of continuous

change. Organization Science 11:611-629.

Feldman, M. S. (2003) A performative perspective on stability and change

in organizational routines. Industrial and Corporate Change 12:727-52

Feldman, M. S., and Rafaeli, A. (2002) Organizational routines as sources

of connections and understandings. Journal of Management Studies 39:309-

31.

Florida, R. (2002) The rise of the creative class. New York: Basic Books.

Frenkel, A. (2001) Why high-technology firms choose to locate in or near

metropolitan areas. Urban Studies 38:1083-1101.

Gertler, M. (2001) Best practice? Geography, learning and the institutional

limits to strong convergence. Journal of Economic Geography 1:5-25.

Gertler, M. S. (2003) Tacit knowledge and the economic geography of

context, or The undefinable tacitness of being (there). Journal of Economic

Geography 3:75–99.

Giddens, A. (1990) The consequences of modernity. Stanford: Stanford

University Press.

Grabher, G. (2002a) Cool projects, boring institutions: Temporary

collaboration in social context. Regional Studies 36:205-214.

Grabher, G. (2002b) The project ecology of advertising: Tasks, talents and

teams. Regional Studies 36:245-262.

Grabher, G. (2004) Temporary architectures of learning: Knowledge

governance in project ecologies. Organization Studies 25:1491-1514.

Granovetter, M. (1985) Economic action and social structure: the problem

of embeddedness. American Journal of Sociology 91:481–510.

Grant, R. M. (1996a) Toward a knowledge-based theory of the firm.

Strategic Management Journal 17:109-122.

Grant, R. M. (1996b) Prospering in dynamically-competitive environments:

Organizational capability as knowledge integration. Organization Science

7:375-387.

Hall, P., and Soskice, D. (2001) Varieties of Capitalism: The Institutional

Foundations of Comparative Advantage. Oxford University Press.

Haltiwanger, J. C.; Lane, J. I.; and Spletzer, J. R. (1999) Productivity

differences across employers: the roles of employer size, age, and human

capital. American Economic Review 89:94–98.

Hamilton, B. H.; Nickerson, J. A.; and Owan, H. (2003) Team incentives

and worker heterogeneity: An empirical analysis of the impact of teams on

productivity and participation. Journal of Political Economy 111:465-497.

59

Harrison, J.; Hitt, M.; Hoskisson, R.; and Ireland, R. (2001) Resource

complementarity in business combinations: Extending the logic to

organizational alliances. Journal of Management 27:679-690.

Heimeriks, G., and Boschma, R. (2013) The path- and place-dependent

nature of scientific knowledge production in biotech 1986-2008. Journal of

Economic Geography, Advance Access; doi:10.1093/jeg/lbs052, 1-26.

Hine, D.; Parker, R.; Pregelj, L.; and Verreynne, M-L. (2013)

Deconstructing and reconstructing the capability hierarchy. Industrial and

Corporate Change (Advance Access doi:10.1093/icc/dtt046).

Hipp, C.; Tether, B.; and Miles, I. (2000) The incidence and effects of

innovation in services: Evidence from Germany. International Journal of

Innovation Management 4:417-454.

Howells, J. (2002) Tacit knowledge, innovation and economic geography.

Urban Studies 39:871-884.

Howells, J. (2012) The geography of knowledge: never so close but never so

far apart. Journal of Economic Geography 12:1003-1020.

Huckman, R. S., and Pisano, G. P. (2003) The effects of organizational

context on individual performance. NBER Working Paper No. 10027.

Hudson, R. (2009) From knowledge-based economy to … knowledge-based

economy? Reflections on change in the economy and development policies in

the north east of England. Regional Studies 45:997-1012.

Jaffe, A. B.; Trajtenberg, M.; and Henderson, R. (1993) Geographic

localization of knowledge spill overs as evidenced by patent citations.

Quarterly Journal of Economics 108:577–598.

Jansson, F. (1997) Åldersstrukturens betydelse för arbetskraftens

rörlighet. Arbetsmarknad och Arbetsliv 3: 295-306.

Kenney, M., and von Burg, U. (1999) Technology, entrepreneurship and

path dependence: Industrial clustering in Silicon Valley and Route 128.

Industrial and Corporate Change 8:67-103.

Kirat, T., and Lung, Y. (1999) Innovation and proximity: Territories as loci

of collective learning processes. European Urban and Regional Studies 6:27-

38.

Klepper, S. (2010) The origin and growth of industry clusters: The making

of Silicon Valley and Detroit. Journal of Urban Economics 67:15-32.

Kline, S.J., and Rosenberg, N. (1986) An overview of innovation, in R.

Landau, and N. Rosenberg, (eds) The positive sum strategy: Harnessing

technology for economic growth. Washington DC: National Academy Press,

pp. 275-304.

Knoben, J. (2009) Localized inter-organizational linkages, agglomeration

effects, and the innovative performance of firms. Annals of Regional Science

43:757-779.

60

Kogut, B., and Zander, U. (1992) Knowledge of the Firm, Combinative

Capabilities, and the Replications of Technology. Organization Science 3:383-

397.

Lacetera, N; Cockburn, I. M.; and Henderson R. (2004) Do firms change

capabilities by hiring new people? A study of the adaption of science-based

drug discovery. Advances in Strategic Management 21:133-159.

Lagerholm, M. (2007) Kunskap och innovation i ett moget kluster. En

ekonomisk-geografisk studie av aluminiumindustrin i Småland-Blekinge.

Geografiska regionstudier 74, Uppsala Universitet, Uppsala.

Lam, A. (1997) Embedded firms, embedded knowledge: Problems of

collaboration and knowledge transfer in global cooperative ventures.

Organization Studies 18:973-996.

Lam, A. (2000) Tacit knowledge, organizational learning and societal

institutions: An integrated framework. Organization Studies 21:487-513.

Lawson, T. (1997) Economics and reality. London: Routledge.

Lawson, C. and Lorenz, E. (1999) Collective learning, tacit knowledge and

regional innovative capacity. Regional Studies 33:305-317.

Leiponen, A., and Drejer, I. (2007) What exactly are technological regimes?

Intra-industry heterogeneity in the organization of innovation activities.

Research Policy 36:1221-1238.

Lucas, R. (1988) On the mechanics of economic development. Journal of

Monetary Economics 22:3-42.

Lundvall, B-Å. (1992) (ed) National systems of innovation. London: Pinter.

Lundvall, B-Å., and Johnson, B. (1994) The learning economy. Journal of

Industry Studies 1:23-42.

MacKinnon, D.; Cumbers, A.; and Chapman, K. (2002) Learning,

innovation and regional development: a critical appraisal of recent debates.

Progress in Human Geography 26:293-311.

Maillat, D., and Lecoq, B. (1992) New technologies and transformation of

regional structures in Europe: The role of the milieu. Entrepreneurship and

Regional Development: An International Journal 4:1-20.

Malerba, F. (2006) Sectoral systems: How and why innovation differs

across sectors, in J. Fagerberg, D. Mowery and R. Nelson, (eds): The Oxford

Handbook of Innovation. New York: Oxford University Press, pp 380-406.

Maliranta, M.; Mohnen, P.; and Rouvinen, P. (2009) Is inter-firm labor

mobility a channel of knowledge spillovers? Evidence from a linked

employer–employee panel. Industrial and Corporate Change 18:1161–1191.

Malmberg, A., and Maskell, P. (2002) The elusive concept of localization

economies: towards a knowledge-based theory of spatial clustering.

Environment and Planning A 34:429-449.

March, J. (2010) The ambiguities of experience. Ithaca: Cornell University

Press.

61

Markusen A. (1996) Sticky places in slippery space: A typology of industrial

districts. Economic geography 72:293-313.

Marshall, A. (1890) Principles of Economics: An Introductory Volume.

London: MacMillan.

Martin, R. (2000) Institutional approaches in economic geography, in E.

Sheppard, and T.J. Barnes, (eds): A companion to Economic Geography.

Malden: Blackwell Publishing, pp 77-94.

Martin, P.; Mayer, T.; and Mayneris, F. (2011) Spatial concentration and

plant-level productivity in France. Journal of Urban Economics 69:182-195.

Maskell, P. (2001) The firm in economic geography. Economic Geography

77:329-344.

Maskell, P.; Bathelt, H.; and Malmberg, A. (2006) Building global

knowledge pipelines: The role of temporary clusters. European Planning

Studies 14:997-1013.

Maskell, P., and Malmberg, A. (1999a) Localised learning and industrial

competitiveness. Cambridge Journal of Economics 23:167-185.

Maskell, P., and Malmberg, A. (1999b) The competitiveness of firms and

regions – ‘Ubiquitification’ and the importance of localized learning.

European Urban and Regional Studies 6:9-25.

Massey, D. (1984) The spatial division of labour. Social structures and the

geography of production. Basingstoke: Macmillan.

Miles, I. (2006) Innovation in services, in J. Fagerberg, D. Mowery and R.

Nelson, (eds): The Oxford Handbook of Innovation. New York: Oxford

University Press, pp 433-458.

Miner, A. S.; Ciuchta, M. P.; and Gong, Y. (2008) Organizational routines

and organizational learning, in: M. C. Becker (ed) Handbook of

organizational routines. Cheltenham, U.K.: Edward Elgar Publishing

Limited, pp. 152-186.

Morgan, K. (1997) The learning region: Institutions, innovation and

regional renewal. Regional Studies 31:491-503.

Morgan, K. (2004) The exaggerated death of geography: learning,

proximity and territorial innovation systems. Journal of Economic

Geography 4:3-21.

Neffke, F., and Henning, M. (2013) Skill relatedness and firm

diversification. Strategic Management Journal 34:297-316.

Neffke, F.; Henning, M.; and Boschma, R. (2011) How Do Regions Diversify

over Time? Industry Relatedness and the Development of New Growth Paths

in Regions. Economic Geography 87:237-265.

Neffke, F.; Henning, M.; and Boschma, R. (2012) The impact of aging and

technological relatedness on agglomeration externalities: a survival analysis.

Journal of Economic Geography 12:485-517.

Neffke, F.; Svensson Henning, M.; and Boschma, R. (2008) Surviving in

agglomerations: plant evolution and the changing benefits of the local

62

environment. Papers in Evolutionary Economic Geography 08.20, Utrecht

University, Utrecht.

Nelson, R. (1995) Recent evolutionary theorizing about economic change.

Journal of Economic Literature XXXIII:48-90.

Nelson, R., and Winter, S. (1982) An evolutionary theory of economic

change. Cambridge, MA: Harvard university Press.

Nonaka, I. (1994) A dynamic theory of organizational knowledge creation.

Organization Science 5:14-37.

Nonaka, I.; Ryoko, T.; and Akiya, N. (2000) A firm as a knowledge-creating

entity: A new perspective on the theory of the firm. Industrial and Corporate

Change 9:1-20.

Nooteboom, B. (2000) Learning and innovation in organizations and

economies. Oxford: Oxford University Press.

Nooteboom, B. (2009) A cognitive theory of the firm, learning,

governance and dynamic capabilities. Cheltenham: Edward Elgar.

Nooteboom, B.; Van Haverbeke, W.; Duysters, G.; Gilsing, D.; and van den

Oord, A. (2007) Optimal cognitive distance and absorptive capacity. Research

Policy 36:1016–1034.

Ottaviano, G. I. P. (2011) ‘New’ new economic geography: firm

heterogeneity and agglomeration economies. Journal of Economic

Geography 11:231-240.

Paci, R., and Usai, S. (1999) Externalities, knowledge spill overs and the

spatial distribution of innovation. GeoJournal 49:381-390.

Patel, P., and Pavitt, K. (1997) The technological competencies of the

world’s largest firms: complex and path-dependent, but not much variety.

Research Policy 26:141–156.

Pavitt, K. (1984) Sectoral patterns of technical change: towards a taxonomy

and a theory. Research Policy 13:343–373.

Pavitt, K. (1998) Technologies, products and organization in the innovating

firm: what Adam Smith tells us and Joseph Schumpeter doesn’t. Industrial

and Corporate Change 7:433–452.

Pavitt, K. (2005) Innovation Processes, in J. Fagerberg, D.C. Mowery, and

R.R. Nelson, (eds) The Oxford handbook of innovation. Oxford: Oxford

University Press, pp. 86-114.

Penrose, E. (1959) The theory of the growth of the firm. Fourth edition.

Introduction by C. N. Pitelis (Reprint 2013) Oxford: Oxford University Press.

Pentland, B. T. (1995) Grammatical models of organizational processes.

Organization Science 6:541-56.

Pentland, B. T., and Reuter, H. H. (1994) Organizational routines as

grammars of action. Administrative Sciences Quarterly 39:484-510.

Pinch, S., and Henry, N. (1999) Paul Krugman’s geographical economics,

industrial clustering and the British motor sport industry. Regional Studies

33:815–827.

63

Piore, M. J., and Sabel, C. F. (1984) The second industrial divide:

Possibilities for prosperity. New York: Basic Books.

Polanyi, M. (1962) Tacit knowing: Its bearing on some problems of

philosophy. Reviews of Modern Physics 34:601-616.

Polanyi, M. (1966) The logic of tacit inference. Philosophy 41:1-18.

Porter, M. (1990) The competitive advantage of nations. New York: The

Free Press.

Puga, D. (2010) The magnitude and causes of agglomeration economies.

Journal of Regional Science 50:203-219.

Rallet, A., and Torret, A. (1999) Is geographical proximity necessary in the

innovation networks in the era of global economy?. GeoJournal 49:373-380.

Rigby, D., and Essletzbichler, J. (1997) Evolution, process variety and

regional trajectories of technological change in U.S. manufacturing. Economic

Geography 73:269-284.

Romer, P. M. (1990) Endogenous technological change. The Journal of

Political Economy 98:S71-S102.

Rumelt, R. (1974) Strategy, structure and economic performance. Boston:

Harvard Business School Press.

Saxenian, A. (1994) Regional advantage: Culture and competition in

Silicon Valley and Route 128. Cambridge, MA: Harvard University Press.

Sayer, A. (1981) Abstraction: a realist interpretation. Radical Philosophy

28:6-15.

Schumpeter, J. (1942) Capitalism, socialism and democracy. New York:

Harper & Row.

Scott, A. (2004) A perspective of economic geography. Journal of Economic

Geography 4:479-499.

Solow, R.M. (1956) A contribution to the theory of economic growth.

Quarterly Journal of Economics 70:27-50.

Song, J.; Almeida, P.; and Wu, G. (2003) Learning-by-Hiring: When is

mobility more likely to facilitate interfirm knowledge transfer? Management

Science 49:351-65.

Spence, M. (1973) Job market signaling. The Quarterly Journal of

Economics 87:355-374.

Statistics Sweden (2014) Fokus på närlingsliv och arbetsmarknad hösten

2014 [Focus on business and labor market]. Information om utbildning och

arbetsmarknad 2014:3. Örebro: SCB-Tryck.

Sternberg, R. (1996) Reasons for the genesis of high-tech regions –

Theoretical explanations and empirical evidence. Geoforum 27:205-223.

Storper, M. (1997) The regional world: Territorial development in a global

economy. New York, NY: The Guilford Press.

Storper, M., and Scott, A.J. (2009) Rethinking human capital, creativity

and urban growth. Journal of Economic Geography 9:147-167.

64

Storper, M., and Venables, A.J. (2004) Buzz: face-to-face contact and the

urban economy. Journal of Economic Geography 4:351-370.

Strambach, S. (2013) Micro-dynamics of knowledge: actors, processes and

territorial organization. Working Papers on Innovation and Space 01.13.

Philipps University, Marburg.

Strambach, S., and Klement, B. (2012) Cumulative and combinatorial

microdynamics of knowledge: The role of space and place in knowledge

integration. European Planning Studies 20:1843-1866.

Szulanski, G. (1996) Exploring internal stickiness: Impediments to the

transfer of best practice within the firm. Strategic Management Journal

17:27-43.

Tanriverdi, H., and Venkatraman, N. (2005) Knowledge relatedness and

the performance of multibusiness firms. Strategic Management Journal

26:97-119.

Timmermans, B., and Boschma, R. (2013) The effect of intra- and inter-

regional labour mobility on plant performance in Denmark: the significance

of related labour inflows. Journal of Economic Geography (Advance Access

doi:10.1093/jeg/lbs059).

Teece, D. J.; PISANO, G. P.; and SHUEN, A. (1997) Dynamic capabilities

and strategic management. Strategic Management Journal, 18:509-533.

Teece, D. J. (2010) Technological Innovation and the Theory of the Firm:

The Role of Enterprise-Level Knowledge, Complementarities, and (Dynamic)

Capabilities, in: B. H. Hall, and N. ROSENBERG, (Eds) Economics of

Innovation. Amsterdam: Elsevier, pp. 679-730.

Tether, B.; Hipp, C: and Hull, F. (2001) Standardization and

particularization in service: Evidence from Germany. Research Policy 30:115-

138.

Torre, A., and Gilly, J.P. (2000) On the analytical dimension of proximity

dynamics. Regional Studies 34:169-180.

Torre, A., and Rallet, A. (2005) Proximity and localization. Regional

Studies 39:47-59.

Trippl, M. (2011) Regional innovation systems and knowledge-sourcing

activities in traditional industries – evidence from the Vienna food sector.

Environment and Planning A 43:1599-1616.

Von Thünen, J. H. (1966) Von Thünen’s The Isolated State: An English

Edition of Der Isolierte Staat. Oxford: Pergamon.

Von Tunzelmann, N., and Acha, V. (2006) Innovation in “low-tech”

industries, in J. Fagerberg, D. Mowery and R. Nelson, (eds): The Oxford

Handbook of Innovation. New York: Oxford University Press, pp 407-432.

Vygotsky, L. (1962) Thought and language. Cambridge, MA: The MIT

Press.

Weber, A. (1929) Theory of the location of industries. Chicago: University

of Chicago Press.

65

Weick, K. E., and Roberts, K. H. (1993) Collective mind in organizations:

Heedful interrelating on flight decks. Administrative Sciences Quarterly

38:357-81.

Wernerfelt, B. (1984) A resource-based view of the firm. Strategic

Management Journal 5:171-180.

Wictorin, B. (2007) Är kluster lönsamma? EN undersökning om platses

betydelse för företags produktivitet [Are clusters profitable? The role of place

for firm productivity]. Doctoral Thesis No. 124, Department of Business

Studies, Uppsala University, Sweden.

Widerstedt, B (1998) Job Mobility, Wage Growth and Match Quality in

Sweden. Doctoral thesis: Moving or Staying? Job Mobility as a Sorting

Process. Umeå Economic Studies No. 464.

Yeung, H. W-C. (1998) The socio-spatial constitution of business

organizations: a geographical perspective. Organizations 5:101-128.

Zeller, C. (2004) North Atlantic innovative relations of Swiss

pharmaceuticals and the proximities with regional biotech arenas. Economic

Geography 80:83-111.

GERUM kulturgeografi (1997-20xx)

(ISSN 1402-5205)

1997:1 Holm, E. (ed.): Modelling Space and Networks. Progress in Theoretical and

Quantitative Geography. 1997:2 Lindgren, U.: Local Impacts of Large Investments. (Akad. avh.) 1998:1 Stjernström, O.: Flytta nära, långt bort. De sociala nätverkens betydelse för val

av bostadsort. (Akad. avh.) 1998:2 Andersson, E.; L.-E. Borgegård och S. Hjort: Boendesegregation i de nordiska

huvudstadsregionerna. 1999:1 Stjernström, O. & I. Svensson: Kommunerna och avfallet. Planering och hante-

ring av farligt avfall, exemplen Umeå och Gotland. 1999:2 Müller, D.: German Second Home Owners in the Swedish Countryside. (Akad.

avh.) 2000:1 Håkansson, J.: Changing Population Distribution in Sweden – Long Term Trends

and Contemporary Tendencies. (Akad. avh.) 2000:2 Helgesson, L.: Högutbildad, men diskvalificerad. Några invandrares röster om

den svenska arbetsmarknaden och vägen dit. 2000:3 Tollefsen Altamirano, A.: Seasons of Migrations to the North. A Study of

Biographies and Narrative Identities in US-Mexican and Swedish-Chilean Return Movements. (Akad. avh.)

2001:1 Alatalo, M.: Sportfisketurism i Västerbottens läns inlands- och fjällområde. Om

naturresursanvändning i förändring. (Lic. avh.) 2001:2 Nilsson, K.: Migration among young men and women in Sweden. (Akad. avh.) 2001:3 Appelblad, H.: The Spawning Salmon as a Resource by Recreational Use. The case

of the wild Baltic salmon and conditions for angling in north Swedish rivers. (Akad. avh.)

2001:4 Pettersson, R.: Sami Tourism - Supply and Demand. Two Essays on Indigenous

Peoples and Tourism in Sweden. (Lic. avh.) 2002:1 Alfredsson, E.: Green Consumption Energy Use and Carbon Dioxide Emissions.

(Akad. avh.) 2002:2 Pettersson, Ö.: Socio-Economic Dynamics in Sparse Regional Structures. (Akad.

avh.)

2002:3 Malmberg, G. (red.): Befolkningen spelar roll.

2002:4 Holm, E.; Holme, K.; Mäkilä, K.; Mattsson-Kauppi, M. & G. Mörtvik: The SVERIGE spatial microsimulation model. Content validation, and example applications.

2003:1 Abramsson, M.: Housing Careers in a Changing Welfare State. (Akad. avh.)

2003:2 Strömgren, M.: Spatial Diffusion of Telemedicine in Sweden. (Akad. avh.)

2003:3 Jonsson, G.: Rotad, rotlös, rastlös – Ung mobilitet i tid och rum. (Akad. avh.)

2004:1 Pettersson, R.: Sami Tourism in Northern Sweden – Supply, Demand and

Interaction. (Akad. avh.)

2005:1 Hjort, S.: Rural migration in Sweden: a new green wave or a blue ripple? (Lic. avh.)

2005:2 Malmberg, G.: Sandberg, L. & Westin, K.: Den goda platsen. Platsanknytning och

flyttningsbeslut bland unga vuxna i Sverige.

2005:3 Lundgren, A.: Microsimulation and tourism forecasts. (Lic. avh.)

2005:4 Brandt, B. F.: Botniabanan – Förväntningar i tid och rum på regional utveckling och resande. (Akad. avh.)

2006:1 Helgesson, L.: Getting Ready for Life – Life Strategies of Town Youth in Mozam-

bique and Tanzania. (Akad. avh.) 2006:2 Lundmark, L.: Restructuring and Employment Change in Sparsely Populated

Areas. Examples from Northern Sweden and Finland. (Akad. avh.)

2007:1 Li, W.: Firms and People in Place. Driving Forces for Regional Growth. (Akad. avh.)

2007:2 Lundholm, E.: New Motives for Migration? On Interregional Mobility in the

Nordic Context. (Akad. avh.) 2007:3 Zillinger, M.: Guided Tourism - the Role of Guidebooks in German Tourist

Behaviour in Sweden. (Akad. avh.) 2008:1 Marjavaara, R.: Second Home Tourism. The Root to Displacement in Sweden?

(Akad. avh.) 2008:2 Sörensson, E.: Making a Living in the World of Tourism. Livelihoods in

Backpacker Tourism in Urban Indonesia. (Akad. avh.) 2009:1 Hjort, S.: Socio-economic differentiation and selective migration in rural and

urban Sweden. (Akad. avh.) 2009:2 Eriksson, R.: Labour Mobility and Plant Performance. The influence of proximity,

relatedness and agglomeration. (Akad. avh.)

2010:1 Khouangvichit, D.: Socio-Economic Transformation and Gender Relations in Lao PDR. (Akad. avh.)

2010:2 Eriksson, M.: (Re)producing a periphery - popular representations of the Swedish

North. (Akad. avh.)

2010:3 Phouxay, K.: Patterns of Migration and Socio-Economic Change in Lao PDR.

(Akad. avh.) 2011:1 Hjälm, A.: A family landscape. On the geographical distances between elderly

parents and adult children in Sweden. (Akad. avh.) 2011:2 Sandow, E.: On the road. Social aspects of commuting long distances to work.

(Akad. avh.) 2011:3 Sandberg, L.: Fear of violence and gendered power relations. Responses to threat

in public space in Sweden. (Akad. avh.) 2011:4 Phommavong, S.: International Tourism Development and Poverty Reduction in

Lao PDR. (Akad. avh.) 2012:1 Haugen, K.: The accessibility paradox – everyday geographies of proximity,

distance and mobility. (Akad. avh.) 2012:2 Olofsson, J.: Go West – East European Migrants in Sweden. (Akad. avh.) 2013:1 Ouma, A.: From Rural Gift to Urban Commodity – Traditional Medicinal

Knowledge and Socio-spatial Transformation in the Eastern Lake Victoria Region. (Akad. avh.)

2013:2 Åkerlund, U.: The Best of Both Worlds – Aspirations, Drivers and Practices of

Swedish Lifestyle Movers in Malta. (Akad. avh.) 2014:1 Olsson, O.: Out of the Wild: Studies on the forest as a recreational resource for

urban residents. (Akad. avh.) 2014:2 Gustafsson, C.: ”For a better life…” A study on migration and health in Nicaragua.

(Akad. avh.) 2015:1 Östbring, L.: Relatedness put in place. On the effects of proximity on plant

performance. (Akad. avh.)