Knowledge flows across space and firms DS

Knowledge flows across space and firms DS

Knowledge flows across space and firms

LINA BJERKE

This dissertation consists of four separate papers and an introductory chapter.

The four papers can be read independently of each other but are held together by concepts around embodied knowledge: knowledge embodied in products and embodied knowledge flows. Thus the papers mainly contribute to the empirical literature on firm and regional knowledge. The rapid growth of knowledge-based industries is one of the prominent features of post-industrialism and economic growth in the industrialised part of the world.

The first paper investigates the residential choice of Swedish university graduates after graduation. It also analyses what factors make them move away from their graduation region. In addition to individual characters such as age and gender, there are also regional characteristics that can either retain graduates or make them choose another residence region. The results of this paper show that large and growing regions are good at keeping their graduates but are also good at attracting graduates from other regions.

The second paper examines what regional characteristics are preferable attributes in order to renew regional exports in the manufacturing sector with export products from other regions. The results indicate that to do so, regions need a specialised export support system and a large amount of sector-related knowledge.

The third paper deals with the issue of how industries and regions absorb new knowledge. Focusing on the role of regional high-quality import flows, the results of this paper show that imports play an important role in regional high-quality export renewal.

The fourth paper investigates how creative labour inflow affect the productivity in firms in knowledge-intensive business services (KIBS). Labour inflow bring new knowledge and increase firm productivity but only if the incoming knowledge is firm-related, which means that the firm can absorb this new knowledge and incorporate and add it into the existing knowledge stock.

ISSN 1403-0470 ISBN 978-91-86345-30-3

JIBS

Knowledge flows across space and firms

LINA BJERKE

Jönköping International Business School

Jönköping University

JIBS Dissertation Series No. 078 • 2012

Knowledge flows across space and firms

LINA BJERKE

This dissertation consists of four separate papers and an introductory chapter.

The four papers can be read independently of each other but are held together by concepts around embodied knowledge: knowledge embodied in products and embodied knowledge flows. Thus the papers mainly contribute to the empirical literature on firm and regional knowledge. The rapid growth of knowledge-based industries is one of the prominent features of post-industrialism and economic growth in the industrialised part of the world.

The first paper investigates the residential choice of Swedish university graduates after graduation. It also analyses what factors make them move away from their graduation region. In addition to individual characters such as age and gender, there are also regional characteristics that can either retain graduates or make them choose another residence region. The results of this paper show that large and growing regions are good at keeping their graduates but are also good at attracting graduates from other regions.

The second paper examines what regional characteristics are preferable attributes in order to renew regional exports in the manufacturing sector with export products from other regions. The results indicate that to do so, regions need a specialised export support system and a large amount of sector-related knowledge.

The third paper deals with the issue of how industries and regions absorb new knowledge. Focusing on the role of regional high-quality import flows, the results of this paper show that imports play an important role in regional high-quality export renewal.

The fourth paper investigates how creative labour inflow affect the productivity in firms in knowledge-intensive business services (KIBS). Labour inflow bring new knowledge and increase firm productivity but only if the incoming knowledge is firm-related, which means that the firm can absorb this new knowledge and incorporate and add it into the existing knowledge stock.

ISSN 1403-0470 ISBN 978-91-86345-30-3

JIBS

Knowledge flows across space and firms

LINA BJERKE

Jönköping International Business School

Jönköping University

JIBS Dissertation Series No. 078 • 2012

Knowledge flows across space and firms

LINA BJERKE

Jönköping International Business School

P.O. Box 1026

SE-551 11 Jönköping

Tel.: +46 36 10 10 00

E-mail: [email protected] www.jibs.se

Knowledge flows across space and firms

JIBS Dissertation Series No. 078

© 2012 Lina Bjerke and Jönköping International Business School

ISSN 1403-0470

ISBN 978-91-86345-30-3

Printed by ARK Tryckaren AB, 2012

2

Acknowledgement

I would like to express my gratitude to a number of people. First and foremost

I wish to thank my supervisor Professor Charlie Karlsson, who had the patience to persistently give feedback on how to structure my work and in what direction I should go when I wavered. He introduced me to research and I am very happy that I accepted his offer and returned home from Spain to begin working at the Department of Economics at Jönköping International Business

School. Many thanks also to my second supervisor Professor Börje Johansson, who has been a major source of inspiration. He has also created the very unique environment at the Economics Department, which I deeply appreciate. I am so glad to be a part of that. All doors are open, no questions are ignored, and all strive together in the same direction. I am forever grateful to Professor Åke E.

Andersson for leading me into the world of knowledge. He also chaired our

Friday seminars and enriched the discussions with new insights and comments.

A special thanks to all who participated in these seminars. Further, I would also like to thank my discussant at the final seminar, Professor Peter Egger, who really dug into my work and gave me the most valuable comments and encouraged me to work hard with my dissertation.

Working together with Professor Martin Andersson has always meant interesting discussions for which I would like to thank him. My thanks also go to Professor Ghazi Shukur and Dr. Kristofer Månsson, who provided me with valuable advice on statistics. I would also like to thank Associate Professors

Charlotta Mellander, Johan Klaesson and Dr. Lars Pettersson for their kindness and constant support. Thanks also to Kerstin Ferroukhi and Katarina Blåman for being at the heart of things at the Economics Department in Jönköping, solving all practical problems. I am also thankful for support from the Centre for Excellence for Science and Innovation Studies (CESIS).

The stimulating research environment in Jönköping is very much created by all my fellow colleagues, to whom I am truly indebted and thankful. You are all fantastic! Some of you have worked closely with me. Mikaela Backman, who I also have had the joy to write together with, is a friend I have needed in happiness and sorrow. Andreas Högberg gave his strongest support and made my time as a PhD candidate even more fun and productive. Pia Nilsson has been my companion during late nights and long summer days at the office.

3

During my time as a PhD candidate I have also had the privilege to travel and attend conferences that have been a source of inspiration. Once more, I thank the professors at Jönköping International Business School for giving me these opportunities.

Lastly, and most importantly, I wish to thank my family. The one who I know would have been the proudest person today is probably sitting on a cloud watching my hard work and stubbornness. However, there are three very special persons who in their everyday lives have had to put up with my hard work and stubbornness: Johan, Joar and Liv. They love me, support me and teach me what is important in life. To them I dedicate this thesis.

Jönköping, April 2, 2012

Lina Bjerke

4

Abstract

This dissertation consists of four separate papers and an introductory chapter.

The four papers can be read independently of each other but are held together by concepts around embodied knowledge: knowledge embodied in products and embodied knowledge flows. Thus the papers mainly contribute to the empirical literature on firm and regional knowledge. The rapid growth of knowledge-based industries is one of the prominent features of postindustrialism and economic growth in the industrialised part of the world.

The first paper investigates the residential choice of Swedish university graduates after graduation. It also analyses what factors make them move away from their graduation region. In addition to individual characters such as age and gender, there are also regional characteristics that can either retain graduates or make them choose another residence region. The results of this paper show that large and growing regions are good at keeping their graduates but are also good at attracting graduates from other regions.

The second paper examines what regional characteristics are preferable attributes in order to renew regional exports in the manufacturing sector with export products from other regions. The results indicate that to do so, regions need a specialised export support system and a large amount of sector-related knowledge.

The third paper deals with the issue of how industries and regions absorb new knowledge. Focusing on the role of regional high-quality import flows, the results of this paper show that imports play an important role in regional highquality export renewal.

The fourth paper investigates how creative labour inflow affect the productivity in firms in knowledge-intensive business services (KIBS). Labour inflow bring new knowledge and increase firm productivity but only if the incoming knowledge is firm-related, which means that the firm can absorb this new knowledge and incorporate and add it into the existing knowledge stock.

5

Table of Contents

Introduction and summary of the thesis ................................................. 11

1. Introduction and overview ................................................................. 11

2. Basic concepts in the economics of knowledge .............................. 12

3. Knowledge production ........................................................................ 16

3.1 Regional knowledge assets and regional agglomeration ............ 17

3.2 Regional agglomeration and knowledge diffusion ...................... 18

4. Microeconomic perspectives on knowledge production and diffusion

4.1 Individuals ......................................................................................... 21

4.2 Labour migration ............................................................................. 22

4.3 Firms .................................................................................................. 24

4.4 Knowledge absorption and learning from trade ......................... 25

5. Knowledge and economic growth ..................................................... 26

6. Empirical concerns and methodology .............................................. 28

6.1 Knowledge flows ............................................................................. 30

6.2 Firm performance ............................................................................ 31

6.3 Data .................................................................................................... 32

7. Outline, summary results and contributions of the thesis ............. 33

References .......................................................................................................... 37

Collection of Papers ................................................................................ 47

Paper 1 Where do university graduates go? .......................................... 49

1. Introduction .......................................................................................... 51

2. Concepts and theories ......................................................................... 53

3. Data ........................................................................................................ 57

3.1 Exposition of data ........................................................................... 61

3.2 Method and analysis ........................................................................ 65

4. Regression results ................................................................................. 67

4.1 Individual characteristics................................................................. 70

7

4.2 Characteristics of graduation region ..............................................72

4.3 Where do university graduates go and what are the push

...............................................................................................74

5. Conclusions ............................................................................................76

References ...........................................................................................................78

Appendix .............................................................................................................81

Paper 2 Export dynamics and product relocation ................................. 83

1. Introduction ...........................................................................................85

2. Regional export base renewal ..............................................................86

2.1 Location, relocation and entry barriers of trade ..........................87

2.2 Multi-product firms and export base renewal ..............................88

3. Hypotheses and data .............................................................................90

3.1 Methodology .....................................................................................91

3.2 Descriptives .......................................................................................95

4. Results .................................................................................................. 101

4.1 Regional absorptive capacity ........................................................ 103

4.2 Regional export support system .................................................. 105

5. Conclusions and further research .................................................... 107

References ........................................................................................................ 110

Appendix A1 ................................................................................................... 113

Appendix A2 ................................................................................................... 114

Appendix A3 ................................................................................................... 115

Paper 3 Imports, knowledge flows and renewal of regional exports .... 117

1. Introduction ........................................................................................ 119

2. Imports, knowledge flows and export renewal-conceptual

framework

3. Data and construction of variables .................................................. 124

3.1 Measurement of high-quality imports and export renewal ..... 125

3.2 High quality imports and export renewal across Swedish

regions

8

4. Empirical strategy, model and results.............................................. 129

4.1 Results .............................................................................................. 133

5. Conclusions and future research ...................................................... 138

References ........................................................................................................ 140

Appendix .......................................................................................................... 143

Paper 4 Knowledge-intensive business services, creative labour inflow and firm productivity ................................................................... 147

1. Introduction ........................................................................................ 149

2. Creative labour inflow and KIBS productivity .............................. 151

2.1 Creative labour and knowledge diffusion .................................. 152

2.2 Regional economic milieu ............................................................. 153

3. Method, data and model formulation ............................................. 154

4. Empirical results ................................................................................. 160

4.1 Labour inflow ................................................................................. 162

4.2 Variables controlling for firm characteristics and regional milieu

5. Conclusions and suggestions for future research .......................... 167

References ........................................................................................................ 169

Appendix .......................................................................................................... 173

JIBS Dissertation Series ............................................................................................ 177

9

Introduction and summary of the thesis

1. Introduction and overview

“The phenomenon of human knowledge is no doubt the greatest miracle in our universe”

(Popper, 1972 p.VII)

Knowledge is the joint feature of the four papers in this dissertation. The rapid expansion of education and the fast growth of knowledge-based industries are two of the most striking changes associated with post-industrialism and the role of knowledge in economic growth have largely been accentuated. In economic analysis, knowledge can be defined as sets of related ideas and associated facts, which are brought into the world through material objects and embodiment in persons (Andersson & Beckmann, 2009). This dissertation deals knowledge in regions, knowledge in firms, knowledge embodied in individuals and embodied knowledge flows. The introductory chapter provides the reader with the background theory of the four from each other independent papers. They can be read and understood independently but are held together by concepts around embodied knowledge and highlight four research questions:

1. How does embodied knowledge locate in space? This question is dealt with in the first paper entitled Where do university graduates go? It explores the residence location of Swedish graduates and what regional factors

2. possibly push them away from their graduation region. The analysis is based on data on all graduates in Sweden, and a multinomial logit model on location is applied to estimate their residence decisions.

What role do regions play in knowledge absorption? This paper is named

Export dynamics and product relocation and examines what regional characteristics are preferable attributes to renew regional exports with export products from other regions. This is analysed by a zero-inflated

Poisson model where the number of relocated export products in

11

Jönköping International Business School

3. sectors and regions is tested, controlling for relevant regional characteristics.

How do industries and regions absorb new knowledge? This question is dealt with in the third paper titled Imports, knowledge flows and the renewal of

regional exports focusing on the role of regional import flows for highquality export renewal. The paper makes use of data on imports and exports across Swedish regions and uses a fractional logit model to analyse this research question.

How is embodied knowledge transmitted between individuals and firms? The final 4. paper is named Knowledge intensive business services, creative labour inflow and

firm productivity and analyses creative labour inflow into firms in knowledge-intensive business services (KIBS) and its effect on firm productivity. This is analysed by a pooled OLS from 2001 to 2008, while controlling for firm- and regional characteristics.

The following sections provide the theoretical framework for all four papers in the dissertation. The ambition is not to provide a complete overview of the economics of knowledge but to highlight the essential contributions for this thesis.

2. Basic concepts in the economics of knowledge

Knowledge is a wide and extensively used concept in a wide range of disciplines. In philosophy, knowledge has for long been discussed but in its modern form, aspects of knowledge were first introduced by Adam Smith

(1776/1904), Karl Marx (1904) and Joseph Schumpeter (1939). However, in a broader context, the importance of knowledge was first recognized by Hayek

(1937), Arrow (1962a, 1962b) and Machlup (1962). Economic literature on knowledge has a number of concepts that need to be disentangled; one can start with human capital.

Human capital is an intangible form of capital and a generic term for skills, knowledge, capacities, education, training, learning-by-doing and schooling.

Everything that improves earnings, health, ethics and good habits can be considered as investments in human capital. It is convenient to use the concept human capital since people cannot be separated from their knowledge, skills,

12

Introduction and summary of the thesis

health and values in the same way as they can be separated from their financial and physical assets (Becker, 1964).

For Hayek and Machlup, knowledge and information were similar concepts but research soon evolved towards a more precise definition (Andersson, 1985;

Andersson & Beckmann, 2009). Information is data, facts and figures and is passive as long as it is not interpreted. Contrary to this, knowledge is cognitive skills, which are important for the capacity of people to interpret passive information and accumulate more knowledge. One central difference between knowledge and information is the marginal costs of copying. For information, this cost is close to zero, especially in modern times of globalisation and modern technology when we can store and share data with modern computer software and telecommunication. Copying of knowledge can, on the other hand be very expensive and is a complex interactive process between a source and an absorber. Another diverging feature between information and knowledge is the way they are transmitted. The marginal cost of transporting information has low (if any) distance sensitivity, while the marginal cost of diffusing knowledge is an increasing function of distance (Audretsch & Feldman, 1996).

On this point, knowledge is immensely different. Knowledge transmission requires social interaction between a sender and a recipient, and knowledge needs to be decodified before it can be absorbed. Information in documents such as notes, books, working papers, journal articles and computer files are sources of knowledge but also the output of knowledge. Such material knowledge is very much like consumption goods or factors of production and can be counted, measured and compared.

Absorption and transfer of material knowledge can be classified into different types with respect to the level of formalisation. On the one end, it can be general and absorbed without any specific knowledge or tools. On the other end, it is fully formalised with the implication that it has only a limited possibility to be absorbed, i.e. tacit knowledge, which was popularised by Nelson and Winter (1982). They presented an evolutionary way of combining information and the outcomes of learning but it is derived from early works by

Polányi (1958; 1962, 1966a; 1966b; 1969), who gave tacit knowledge a central role in the learning economy and innovation. Tacit knowledge is frequently used as a general term describing all uncodified knowledge but Polányi formalises the difference between tacit knowledge and codified knowledge.

Codified knowledge is closer to information in the sense that it can to a larger extent be transmitted and duplicated if the recipient has awareness of the specific topic. Tacit knowledge is intangible in the sense that it involves skills such

13

Jönköping International Business School

as knowing how to ride a bicycle or how to acquire scientific intuition. The more implicit it is, the more difficult it is to absorb (Cohen & Levinthal, 1990).

To be more precise, “tacitness” arises when knowledge is not yet explicitly stated while codified knowledge can be shared without human support (Foray

& Lundvall, 1996; Polanyi, 1966b). In traditional definitions tacit knowledge requires some form of face-to-face interaction whereas codified knowledge can be distributed freely through channels such as emails and letters. This assumption may be too strict if we introduce concepts related to knowledge diffusion and knowledge flows.

Knowledge flows is a generic term describing inflows and outflows of knowledge from a firm, region or nation. It can either be deliberate market activities between suppliers and receivers of knowledge or it can be pure spillovers (Johansson, 2005). The latter is related to the fact that knowledge can, to some extent be associated with non-excludability. This means that it is difficult to keep as a private asset and spills over to other parts of the economy.

Knowledge also has an element of non-rivalry which means that it can be consumed and accumulated irrespective of population size.

These public-good characteristics of knowledge lead to the argument that knowledge transmission has a non-zero marginal cost but the knowledge good itself still remains free (Stiglitz, 1999). Already Nelson (1959) and Arrow

(1962b) discuss this in terms of knowledge spillovers and its public character and to what extent it is a non-rival input in production. The cumulative result of all knowledge flows (inflows and outflows) can be defined as the knowledge stock.

Knowledge flows can therefore be adjusted instantaneously whereas knowledge stocks cannot (Dierickx & Cool, 1989).

The idea that knowledge is a pure public good is not prevalent in modern economic literature but knowledge spillovers can instead be presented as: “…a prototypical externality, by which one or few agents investing in research or technology development will end up facilitating other agents’ innovation efforts...” (Breschi & Lissoni, 2001a, p. 975). Research in knowledge spillovers has now grown in the direction of spatial dependencies and the belief that knowledge spillovers are actually knowledge externalities bounded in space

(Breschi & Lissoni, 2001b).

Knowledge externalities are not new in literature and has traditionally been grouped into two categories of externalities, pecuniary externalities and technological externalities (Scitovsky, 1954). Pecuniary externalities are byproducts of deliberate market interactions. Technological externalities materialise through non-market interactions and are considered as pure spillovers. Since

14

Introduction and summary of the thesis

they are by-products of non-market interactions, technological externalities are sometimes related to pure market failures such as congestion.

When knowledge can be passed on to a new recipient, geography becomes an important factor and one can discuss differences between global knowledge and local-specific knowledge. One might believe that globalisation reduces distances and spurs knowledge sharing. Contradictory to this, there are still knowledge relationships that need very close distances (Porter, 1998). As an aid to exploring knowledge relationships and extent of knowledge networks we can introduce the term sticky information. This concept can be defined as the

“…incremental expenditure required to transfer that unit of information to a specified locus in a form usable by a given information seeker” (von Hippel,

1994). When this diffusion cost is low, stickiness is low and when the costs are high, stickiness is also high.

Some argue that tacit and codified knowledge are the disembodied knowledge that can be placed in a context of localised learning. Disembodied knowledge is the know-how that is the result of positive externalities of the innovation process. This is generally based on individuals and their skills and experiences, a collective technical milieu and a specialised institutional framework which all are highly immobile (Asheim, 1999; Castro & Jensen-Butler, 1993). In contrast to this, embodied knowledge is the way knowledge is brought into the real world; first in persons and then in material objects such as documents, machinery and products. The discussion of knowledge concepts and knowledge characteristics can now neatly be placed in a spatial setting of knowledge accumulation. New technology and social media enable the world to be highly interconnected but this does not mean that we can discard the importance of the spatial dimension and knowledge diffusion.

Recent interest and focus have also been directed towards other ways to think about and reflect upon knowledge. Already Andersson (1985) emphasised that knowledge workers with specific competences are separated from information and data. Knowledge workers have the ability to combine their knowledge and competences and create something new, an ability called

creativity. This is a dynamic concept and Sternberg and Lubart (1999) explain creativity as the ability to produce work that is novel and adaptable and by that important in solving complex problems.

15

Jönköping International Business School

3. Knowledge production

A formal way to treat knowledge production is the constant elasticity of substitution (CES) function, which is a general production function where one can substitute inputs across different types of production (Arrow et al., 1961).

1

In microeconomic theory of production, this is a convenient and flexible choice of modelling. Equation 1 presents a way to describe knowledge production in a micro-economic setting and allows us to vary combinations of inputs

, into knowledge production, such as skilled labour, university research or knowledge acquired from R&D investments (Andersson & Beckmann, 2009).

, = +

, > 0 ≤ 1

(1)

The production function formulated in Equation 1 is concave when

0 < ≤ 1

, which implies that if

> 0

with only one input factor,

> 1

or

> 1

, there will anyway be a positive output. When

< 0

, then both input factors are necessary to yield a positive output. It is easy to understand that

> 0

is the normal case in knowledge economics since we assume that new knowledge can be created through embodied knowledge (labour) from a given stock of knowledge without any outher input factors and in an extreme case we can note:

= −∞ , = ,

(2)

With Leontief fixed coefficients,

= 0

, there is a production function such that, and the CES function is linear and homogenous in all cases there are decreasing, constant or increasing returns to scale.

Just as it is tricky to measure inputs in production functions of normal goods, it is also difficult to measure inputs in knowledge production. The nonrival element of knowledge implies a non-convex production function and one cannot say that actors are price-takers at a competitive market (Arrow, 1962b;

Dasgupta & Stiglitz, 1988; Romer, 1990; Schumpeter, 1943). Here, absorptive capacity is a central concept in understanding knowledge accumulation and knowledge production.

The capability to adopt and assimilate knowledge from external sources can be defined as absorptive capacities that are intangible results of knowledge production. It can either be a by-product of own R&D investments (Mowery,

1

This model was preceded by two competing alternatives: the Walras-Leontief-Harrod-Domar assumption of constant input coefficients and the Cobb-Douglas production function with a unitary elasticity of substitution between labour and capital. The latter was first presented in the late 19 th

century by Knut Wicksell (1898).

16

Introduction and summary of the thesis

1983), a by-product of production experiences (Abernathy, 1978; Rosenberg,

1982), or a result of agglomeration of educated individuals. It can also be a combination of all these.

3.1 Regional knowledge assets and regional agglomeration

The accumulated output of knowledge production are sometimes referred to as

knowledge assets. It can be physical capital such as industrial structure, property rights, network of firms and networks of suppliers, customers and competitors.

Knowledge assets are difficult to copy and there is increasing recognition that the competitive advantage of firms is their ability to create, transfer, utilise and protect these assets (Teece, 2000). Regional characteristics can act as knowledge assets playing the role of inputs in knowledge production; much research shows that the regional economic milieu can foster innovative activity (Brennenraedts et al., 2006; Henderson et al., 1998; Johansson & Andersson, 1998; Zucker et al., 1998b). One such regional characteristic can be the mere presence of universities and other educational institutions. These attract various forms of human capital and the concentration of knowledge can contribute to regional innovation performance (Acs et al., 1992; Anselin et al., 1997; Bradley &

Taylor, 1996; Faggian & McCann, 2006b).

Network creations around innovation activities are sometimes referred to as

innovation systems, facilitating generation, transmission and assimilation of knowledge (Fischer & Fröhlich, 2001).

2

Linkages between regional clusters and partners elsewhere are facilitated trough formal and informal innovation networks or embodied through exports and imports. Silicon Valley in

California, the electronic cluster in central parts of Japan, the financial districts in London and New York and the leather fashion cluster in Italy are all examples of concentration of innovations.

Johansson and Andersson (1998) summarise regional knowledge assets into five intensity attributes and emphasise the role of these intensities in the production of knowledge and knowledge outputs. The first is import intensity, which enables a constant supply of new information and knowledge. The second is knowledge intensity, which is required to absorb incoming knowledge and transmit it into production. The third is R&D intensity and the fourth is

2

The Dixit-Stiglitz model of monopolistic competition clarifies how agglomeration can be explained by such inter-regional mobility of goods and production factors (Dixit & Stiglitz, 1977). By combining monopolistic competition with traditional models of transportation costs, models of new economic geography emphasise the importance of pecuniary externalities and strong home-market effects. Locations with sufficiently large demand attract a relatively large share of production.

17

Jönköping International Business School

customer intensity. The final attribute is supplier intensity describing the value of complementary firms with a good supply of input goods and business services.

In fact, knowledge- and R&D activities tend to agglomerate in space for a number of reasons, of which knowledge spillovers are one. The identification of knowledge spillovers is central in the economics of agglomeration and suggests that knowledge investments can spill over locally but also to other geographical locations. Much of the idea is that knowledge activities perform better with concentration and the benefits are the externalities generated from firms and individuals.

3.2 Regional agglomeration and knowledge diffusion

The relationship between knowledge assets and knowledge diffusion was largely omitted in the literature until the second wave of endogenous growth theories

(Aghion & Howitt, 1992). Knowledge flows can take various forms but have principally been discussed from either a firm perspective or from an individual perspective. A flow of knowledge can be a transaction similar to other market transactions with a receiver of knowledge who pays the supplier of knowledge.

This can for instance be consultants hired by a firm to solve a complex problem. It can also be a joint action between firms in a region, for instance collaboration in product development, internal networks or cooperation with a specific input supplier. Pure knowledge spillovers are the knowledge that firms acquire through non-deliberate actions such as product and process imitations.

They can still occur through links and cooperation, but a large part of spillovers require face-to-face contacts and geographic proximity (Andersson &

Mantsinen, 1980; Ciccone & Hall, 1996; Coe & Helpman, 1995; Jacobs, 1969;

Jaffe et al., 1993; Karlsson & Manduchi, 2001; Keller, 2004).

On the supply side of regional agglomeration, spillovers are by-products of highly specialised environments, i.e. localisation economies or, Marshall-Arrow-

Romer externalities (MAR) (Marshall, 1890/1920).

The idea is the idea that magnitude and effectiveness of face-to-face interactions decreases with geographical distances. Information, ideas and knowledge are weightless but the main belief is anyhow that knowledge externalities are more easily transmitted over smaller geographical distances (Andersson & Beckmann, 2009; Krugman,

1991a).

3

It can also be university spillovers and knowledge production, which sometimes are called local academic spillovers (Jaffe, 1989; Varga, 2000).

3 ICT (Information and Communication Technology) is naturally mentioned here. Communication costs have decreased and a demand and supply schedule implies that demand for face-to-face interaction should fall. On

18

Introduction and summary of the thesis

Another feature is that labour demand and labour supply are larger in agglomeration economies. Labour market pooling generates a more efficient matching process between employers and employees. This reduces frictional unemployment as well as reduces the risk of being employed in a sector which does not perfectly match the education, skills and experiences (Krugman,

1991a, 1991b). A final advantage of agglomeration is input sharing and sharing of business services and infrastructure (Duranton & Puga, 2004; Fujita &

Thisse, 2002; Johansson & Quigley, 2003).

Porter (1990) argues that knowledge spillovers in specialised clusters stimulate economic growth. Porter further notes that local competition, as opposed to local monopolies, spurs the pace and extent of knowledge absorption and innovations. Thus, these externalities are maximised in cities with competitive industries in geographic clusters. Another type of spillovers is the one illustrated by Jacobs (1961, 1969). She argues that geographical concentration of different industries instead creates a heterogeneous economic milieu where diversity of knowledge and ideas foster new ideas, innovations and economic growth.

A part of economic literature discusses urban agglomeration and emphasises the role of cities and metropolitan regions for knowledge production and knowledge diffusion. The congestion costs of being located in a highly diversified regions are outweighed by the reduced search costs for the ideal production inputs. These regions can act as nursery cities for new products and innovations.

4

They are characterised by external economies, a diversified industrial environment and a large share of skilled labour, which makes them favourable locations for innovation and product development (Andersson &

Johansson, 1984; Duranton & Puga, 2001; Krugman, 1979).

Johansson and Karlsson (2003) discuss the diffusion from metropolitan regions from the aspect that they are particularly good breeding places for innovations. The diffusion process can take various forms: (i) Firms in the metropolitan region decentralise activities, either to lower production costs or because they are growing, (ii) they can change their internal division of labour when they have production units in several different regions, (iii) they can outsource part or all of their production to independent firms (suppliers) in other regions, (iv) they can make it possible for firms in other regions to use the contrary, it could be that the quantity of information flows increases, leaving quality at the same level.

Then, ICT functions is a complement rather than substitute in sharing information (Nonaka, 1994).

4

Export relocation is another type of regional knowledge diffusion and access to human capital is a critical factor for the capacity of smaller regions to adopt new export products from larger regions (Andersson et al.,

2007; Gråsjö, 2006).

19

Jönköping International Business School

their business concept via licensing, franchising, etc., (v) other regions can imitate their products, and finally (vi) they can developed new products based on new knowledge and locate the production of these products into another region.

The diffusion process described above is a framework which is rather spatially specific and it follows assumptions of the product cycle theory as presented by Vernon (1966), Norton & Rees (1979), and Andersson and

Johansson (1984). Larger regions can improve their relative situation by increasing the rate of new product development, while smaller regions can improve their situation by speeding up imitation of new products. The product life cycle is a dynamic chain of events where the division of labour and knowledge determines the competitive advantages of locations. The dependence on specific regional attributes changes over time when products go from newly invented to obsolete on the market.

5

The early phase is that of creativity, innovation and conceptualisation, typical of a process with high knowledge intensity. Location is concentrated to only a few advanced regions.

The following is the growth stage, a phase with diminishing knowledge intensity, and firms seek lower production costs and standardised production processes. Domestic and foreign demand increases and the product technology gradually becomes available to the rest of the economy.

4. Microeconomic perspectives on knowledge production and knowledge diffusion

A general belief is that knowledge is expensive to produce, possible to duplicate and depreciates over time, and that only a fraction of all knowledge flows are accumulated into the knowledge stock. Why are there still incentives to invest in knowledge? The answer is that knowledge assets create opportunities for further development and growth on the micro, meso, and macro levels of the economy. Thus, individuals benefit from higher incomes, firms generate innovations and better firm performance, regions increase their productivity

5 This dynamic relationship between product cycles and technology transfer determines patterns of trade and these modelling assumptions differ significantly from conventional trade theory as described by Ricardo or

Heckscher (1931). The theory of product cycles is thoroughly scrutinised and discussed by Andersson and

Johansson (1984), Johansson and Andersson (1998) and Johansson and Karlsson (1987).

20

Introduction and summary of the thesis

levels and competitive positions and nations experience a higher economic growth.

4.1 Individuals

Schooling, training and labour migration are different types of human capital investments and earnings from these can be monetary (e.g. wages) and nonmonetary (self-employment or better residence options) (Becker, 1964;

Gottlieb, 1995; Mincer, 1958; Roback, 1982;1988). While studying, earnings are lower which means that future earnings need to be sufficiently high to cover costs associated with higher education. The marginal returns on higher education is a complex issue and one also need to contemplate the reduced risk of unemployment (Mincer, 1991). Another type of risk is the one of being employed in the “wrong” sector where they cannot fully use their knowledge and skills which means that if individuals can freely choose location they seek to maximise expected income, given the regional share of unemployment and by that, incorporating the employment possibilities in a region (Backman &

Bjerke, 2009).

While education is a common way to measure human capital investments, there are alternative measures.

6

Similar to the gains from education there are also monetary gains to reap from having an occupation with work tasks associated with creativity and cognitive abilities (Stolarick et al., 2010). The interest for other human capital measures has increased since production in the industrialised part of the world gradually has changed from manufacturing to the service sector. Labour demand has shifted towards a new type of skilled labour with occupations associated with complex problem solutions and cognitive knowledge input (Florida, 2002c; Glaeser & Saiz, 2004; Saxenian,

1994). A large number of creative individuals are employed in knowledgeintensive business services (KIBS) and Figure 1 illustrates the rapid percentage growth of number of work places and number of employees in this sector in

Sweden. The KIBS number of employees (workplaces) has increased by 119

(103) per cent compared to the 17 (21) per cent growth in the rest of the

Swedish economy.

6 Education and training are the most significant parts of human capital investment, and early studies show that such investments increase labour productivity and labour income (Becker, 1962; Schultz, 1961).

An alternative to theories on productivity-adding human capital there are theories on education as a screening or

filtering device by which education sorts out individuals with different abilities and functions as information for those that demand labour. Education can also be productivity-adding on a private level but the social productivity of education is more problematic (Arrow, 1973).

21

Jönköping International Business School

250

225

200

175

150

125

100

KIBS Employees Sweden Employees

KIBS Workplaces Sweden Workplaces

Figure 1 Growth of number of employees and number of firms in the KIBS sector and in Sweden between 1997 and 2010. Indexed base year =1997

4.2 Labour migration

The mere fact that knowledge is embodied in people implies that migration can matter for regional performance (Faggian & McCann, 2006a; Gertler, 2003;

Hudson, 2005; Lawson, 1999; Rodriguez-Pose & Vilalta-Bufi, 2005). Yet the key mechanisms and causality behind embodied knowledge mobility and performance of regions are still not fully explored. Literature on causes and effects of labour mobility has shown that national migrates are far from being a homogeneous group of people and one way to examine the relation between knowledge diffusion and labour migration is to study the migration of graduate students. While some students choose to stay in their exam region after graduation, others seek a job in another region. Where do they choose to reside and work after graduation? Are there regions that are particularly good at attracting graduates? Previous literature actually show that individuals with higher education tend to be more mobile and are more willing to move longer distances than non-educated individuals (Becker, 1964; Johansson & Quigley,

2003; McCann, 2001; Schwartz, 1976; Sjaastad, 1962). In relation to this, the migration of university graduates can be placed in a context of equilibrium issues around what drives labour migration.

If the regional labour market demand were perfectly synchronised with the type of students graduating locally there would be a zero migration flow and

22

Introduction and summary of the thesis

the system of regional labour markets would be in equilibrium.

7

All graduates would be absorbed by the regional labour market. If individuals are satisfied with their location, a Pareto equilibrium prevails and no one would have the intention to migrate. The implication of this is that inter-regional migration and equilibrium is a non-feasible combination.

8

If there is zero regional net migration in equilibrium, one also has to assume that the ratio of out-migration probabilities and population shares are mutually related in the long run

(Harrigan & McGregor, 1993). Then two equally sized regions with the same number of possible migrants would have a gross migration flow that selfcancels in equilibrium. This means that a nation (economic system) with a fixed number of regions can be in equilibrium at the aggregate level while there is inter-regional migration within the nation. So, if population shares are distributed across age of individuals and across regions at the year (t+1). Then, the following must hold:

+ 1 = Α + Μ

, where equilibrium occurs if

= Α + Μ

(3)

The first variable matrix, А holds information on birth rates and death rates distributed across regions at the year (t+1). The matrix М contains information on migration probabilities between regions distributed across age of individuals.

If the sum of the eigenvalues of regional population shares are larger than 1 i.e.,

=

> 1

, the equilibrium is unstable. This implies that there is population growth at the same rate in all regions and that

+ 1 =

sum up to 100 (in percentages). Thus, migration would be modified such as there is some regional population growth and the main interest is to find the answer to what the process looks like in order to reach a constant population share in all regions.

So, in a system where population shares are not uniformly distributed across regions, one expects positive and negative figures of regional net-migrations.

Individual equilibrium can be thought upon as a constant probability to move but disequilibrium arises when whichever of the migration decision factors change. One type of decision factors is the one associated with the regional labour market. The consequence of the equilibrium discussion on migration is that, in order to generate a long-run equilibrium we must relax the assumption that all graduates have identical characteristics and that all regions are of equal

7 Some of the first ones to explicitly state the forces on the demand side were Haig (1926) and Lampard

(1955). Migration theory and spatial concentration of individuals were formally introduced by Muth (1971), who stated that migration and employment growth mutually affect each other

8

Markov processes are repeatedly used in migration studies using wage differentials as equilibrium variables.

These studies tend to underestimate the necessary population changes in order to reach stochastic equilibrium

(Kelley & Weiss, 1969).

23

Jönköping International Business School

size (Evans, 1990; Harrigan & McGregor, 1993). These asymmetries and their impact on migration can be divided into individual factors, for instance family relations, age, gender, ethnicity (Faggian et al., 2006, 2007) but also on regional factors such as different labour market conditions.

4.3 Firms

In terms of firm innovation, there is a well-established and positive relationship between firm R&D investments and firm performance (Eklund & Wiberg,

2008; Griliches, 1979, 1984; Gråsjö, 2005; Johansson & Lööf, 2008; Klette &

Kortum, 2004; Kortum, 2008). However, the effects of firm R&D on knowledge production, innovation and firm performance is dependent on type of development.

9

An influential strand of literature differentiates between types of innovations with respect to knowledge input and innovation output

(Abernathy & Utterback, 1978; Porter, 1986).

First, incremental innovations builds on existing knowledge and resources of firms and a radical innovations are associated with completely new knowledge, and incoming knowledge and new resources are, in that case knowledgedestroying. These types of innovations can also be differentiated with respect to external effects. An incremental innovation is associated with modest technological changes, and existing market products remain competitive.

Radical innovations are associated with large technological advancements, which implies that existing products become obsolete. Following this, incumbent firms will be in a better position if the innovation is incremental, since they can use their built-up knowledge and resources. In contrast, new entrants will have an advantage if the innovation is radical because they will not need to change their knowledge assets.

The primary rationale for firm knowledge absorption and firm innovation is the possibility to gain monopoly profits. The uniqueness of a new product fades over time and it becomes obsolete when new innovations enter the market. As above discussed, product life cycle theories illustrate the fact that market-leading positions need to be constantly supported by new knowledge and new innovations. To do this, firms rely on their own R&D activities but also on external sources of knowledge input, which can be described as innovation systems (Acs et al., 1992, 1994; Feldman, 1994a, 1994b; Jaffe, 1989).

9 Consumer preferences are associated with firm product supply and firm innovation. Consumers have preferences for specific product characteristics and different combinations of such characteristics (Becker,

1965; Hicks, 1965; Lancaster, 1966). Associated with this is the literature on horizontally and vertically differentiated products (Lancaster, 1975).

24

Introduction and summary of the thesis

Regional knowledge assets are important for firm innovativeness but firm attributes are equally important in order to stay innovative and once a firm is classified as innovative, firm attributes are even more important (Johansson &

Lööf, 2006). Firm attributes associated with the ability to take in and adapt to new knowledge specify the absorptive capacity and one part of this is embodied knowledge in terms of skilled labour. External knowledge is also a function of prior knowledge which means that absorptive capacity can be highly path dependent (Allen, 1977; Almeida & Kogut, 1999; Cohen & Levinthal, 1989,

1990; Mowery, 1983). Others suggest that absorptive capacities are generated by production processes (Abernathy & Utterback, 1978; Rosenberg, 1982), or results of a high density of educated individuals (Ciccone & Hall, 1996) or a combination of individuals’ capacities (Roper & Love, 2006; Van den Bosch et al., 2003; Zucker et al., 1998a; Zucker et al., 1998b).

4.4 Knowledge absorption and learning from trade

Firms and regions only produce a small fraction of world knowledge and innovation output and can be framed into a global setting. Coe and Helpman conclude their analysis by stating that:

not only does a county’s total factor productivity depend on its own R&D capital stock,

but, as suggested by the theory, it also depends on the R&D capital stock of its trade partners”

(Coe & Helpman, 1995, p. 875)

On the export side, this is sometimes described as learning by exporting

(Almeida & Kogut, 1999; Grossman & Helpman, 1991a, 1991b; Nelson, 1993).

Exports also widen the scope and scale of demand, and greater consumer demand has a multiplier effect on output and investment, reallocating resources to firms with the highest productivity (Kaldor, 1970).

10

Yet, in contemporary literature the role of imports on regional innovation is rarely mentioned but has gained recent interest.

The role of imports goes back to historical studies by Eli F. Heckscher, who argued that imports vary significantly more than exports (Heckscher, 1957); the growth-enhancing effects of imports are highly predominant compared to exports (Falvey et al., 2004). Imports are equally valuable knowledge assets,

10 Studies show that falling trade costs create this kind of selective effect on those firms being less productive and force them to exit the market (Bernard et al., 2006; Clerides et al., 1998; Pavcnik, 2002).

25

Jönköping International Business School

affecting R&D, innovation and economic development (Coe & Helpman, 1995;

Grossman & Helpman, 1991a, 1991b; Keller, 2004). The basic idea is that import networks constitute a pertinent feature of knowledge transmission. A large and diverse import is strategically important; import linkages can be channels of new knowledge, new technology and new products.

To what extent firms and regions actually can learn from trade is very much related to regional absorptive capacity but also on durable regional characteristics. These comprise access to local and external market potentials for different types of products but also supply of durable capacities. Durable capacities are the regionally trapped resources, such as material and nonmaterial infrastructure and the regional sector composition, which at least in the short run is a trapped resource (cf. Johansson et al., 2001; Maurseth &

Verspagen, 1999).

5. Knowledge and economic growth

Regional competitiveness is a complex and cumulative outcome of factors which is explained in three major strands of literature. The first is the neoclassical theory, placing emphasis on technological progress, how it exogenously affects economic growth and that much of regional advantages originates from export specialisation (Solow, 1956,1957, Lucas, 1988, Romer,

1990). The second is endogenous growth theory which incorporates investments and knowledge assets generated within economies. It explains the existence of clusters of knowledge, technological development and innovation, i.e. local innovation systems. Thus, production of human capital, R&D investments, new technologies and knowledge spillovers is assigned a central role for economic growth (Arrow, 1962b; Harrod, 1939; Hicks, 1965; Solow,

1957). The third is the theory on increasing returns, which explains agglomeration economies but irrespective of theoretical framework, much of regional competitiveness is generated from attracting input factors into innovation processes where knowledge is one type of input (Kitson et al.,

2004).

A striking feature of knowledge and innovation is the variability in time and space; economies rather tend to experience structural changes than smooth paths of innovations. Also, knowledge investments have the intricate character of diminishing returns which means that twice as much input does not generate twice as much output. Thus, regional economic growth is increasingly driven by

26

Introduction and summary of the thesis

endogenous or decentralised factors such as industrial structure, innovation,

R&D investments, competitiveness, governmental institutions and culture

(Romer, 1994). Economic growth is associated with external returns where, from each new idea, a number of others arise, triggering innovations and driving the long-run economic growth and Griliches (1979, 1986) is often credited with the role of pathfinder on how to explain the relation between

R&D and productivity growth. Abilities to exploit regional assets determine firm competitiveness and urban competitiveness and it has been shown that denser regions actually have a higher rate of innovation (Antonelli, 1994;

Feldman & Audretsch, 1999; Glaeser, 1999). The problem arises when one would like to compare the competitiveness of firms with the competitiveness of regions.

11

Regions are neither macro oriented (national level) nor micro oriented (firm level). Nor do regions run out of business as firms can do, and region’s overall performance cannot directly be measured as productivity.

12

However, innovative activity is more likely to appear in some regions than in others and we can note that through history, regions with significant economic growth have always had a combination of successful technology advances and large investments in education and training (Becker, 1993).

Regional differences in innovation are primarily variations in knowledge assets, such as research universities, private and public R&D, infrastructural networks and characteristics of labour markets. A wide base of knowledge assets enables absorption, transmission and accumulation of embodied knowledge and new technology. At a regional level, the accumulation of knowledge assets and how it relates to economic growth is a dual problem where both parts are, to some extent policy related. The first part of the problem is to increase the regional stock of human capital, for instance by facilitating higher educations. The second part of the problem is to retain these accumulated knowledge assets in the region (Bradley & Taylor, 1996).

A sub-set of models in economic growth suggests that R&D investments are endogenously determined with a dependence on embodied knowledge

(Shell, 1966; Uzawa, 1965). New knowledge can diffuse through spillovers but requires real absorptive capacities in the labour force (Braconier, 1998; Nelson

& Rosenberg, 1993; Young, 1995). Another set of theories argues that the

11

For the historical development of regional competitiveness see e.g. Budd and Hirmis (2004) or Kitson et al.

(2004). Krugman heavily criticises Porter’s ideas on competitiveness (Dornbusch, 1994; Krugman, 1996).

12

Schumpeter (1943) and Bain (1956) early noted the role of competitiveness between industries and how there are variations in innovative capabilities. Williamsson (1985), Porter (1990; 2003) and Caves (1980) develop this further and argue that competitiveness is related to institutions and industrial organisations.

27

Jönköping International Business School

endogenous formation of knowledge by R&D is free to be used by all firms in a nation (Romer, 1990; Grossman & Helpman, 1991a).

A parallel track to the above is the Schumpeterian alternative to economic growth developed by evolutionary economists (Schumpeter, 1934). A major flaw in neoclassical economic growth models is the rather strict assumptions on equilibrium. The evolutionary branch of economics partly emerged to give an alternative explanation according to which economic change is historically determined and a continuous evolutionary process (McKelvey, 1994; Nelson &

Winter, 1974, 1982).

6. Empirical concerns and methodology

This dissertation comprises two main aspects of knowledge: knowledge of firms and industries and knowledge of individuals. A general and all-embracing way to measure knowledge and its effects has not yet been presented. However, proxies for knowledge have long been used in literature preceded by lively debates, and we need to make a distinction between knowledge inputs and knowledge outputs. With the risk of starting on the wrong side of the production function, patent counts were perhaps the first quantitative proxy for knowledge output (Griliches, 1984; Scherer, 1965; Schmookler, 1966). Also the influential papers by Lucas (1988) and Jaffe (1989) on the effects of academic research and patent activities inspired a range of literature on innovation counts, patent citations (Jaffe et al., 1993) and regional cross-citations

(Maurseth & Verspagen, 1999; Verspagen & Schoenmakers, 2000).

Yet, quantitative measures have some limitations. The primary issue is the difficulty to allocate counts in space to exactly one relevant firm or industry.

The second flaw is the wide differences in quality and economic significance across counts such as granted patents.

These problems illustrate the difficulties of many quantitative proxies on knowledge outputs and the fact that it is tremendously problematic to estimate the effects on economic growth.

13

Also, patents is a measure in the borderland between knowledge input and knowledge output and cannot be completely separated from each other

(Griliches, 1990). Instead, R&D expenditures can be applied as an alternative input measure while output can be measured as the number of innovations,

13

Data on patent renewals and patent citations are sometimes applied to solve the issue of quality heterogeneity (Albert et al., 1991; Pakes & Schankerman, 1984; Trajtenberg, 1990). However, not all innovations are patentable and patents can therefore only explain a fraction of all innovative activity.

28

Introduction and summary of the thesis

firm productivity growth, profit change, new firm settlements or stock market value (Kleinknecht & Poot, 1992).

14

Turning now to knowledge inputs, a common way to measure these is to discern the share of the labour force with at least a bachelor degree i.e. education

(Glaeser, 2005; Glaeser & Saiz, 2004). There are other ways to think upon knowledge and already Andersson (1985) emphasises that knowledge workers with specific competences are separated from information and data.

15

Knowledge workers have the ability to combine their knowledge and competences and create something new, which is called creativity.

People with creative occupations in science, engineering, arts, culture, entertainment and knowledge-based professions of management, finance, law healthcare and educations have in recent literature been defined as the creative

class (Florida, 2002a, 2002b, 2002c). By focusing on an occupational-based creative class measure, one can scrutinise what employees actually do at work and better capture the effect of the absorptive capacity. There is a debate around creativity that mostly deals with how to separate it from educationbased measures and whether creative individuals are the engines of economic growth. Hoyman and Faricy (2009) find no evidence for this. Also, Markusen

(2006) disapproves of the index structure; she argues that the creative class is highly heterogeneous and argues that occupations are largely based on educational attainment and grouping workers into a creative class is insignificant when controlling for education.

However, for data on Sweden, Mellander (2009) shows that only 25 per cent of creative occupations hold a degree with three or more years of schooling. By that she illustrates that a long education is no prerequisite for being creative but rather that the probability to have a creative occupation increases with education.

Despite the choice of measure, knowledge input and knowledge output have shown to be spatially dependent on each other. Empirical findings show that patent citations (Acs et al., 2002; Jaffe et al., 1993), R&D expenditures

(Acs, 2002), manufacturing activities (Gordon & McCann, 2000), highly educated labour (Faggian & McCann, 2006b), skilled labour (Florida, 2002c;

Glaeser & Saiz, 2004), and import activities (Johansson, 1993) are all geographically localised.

14 R&D expenditures are frequently used as a measure of input in the innovation process. This can be an imperfect measure since particularly small firms do not separate R&D expenditures from other investments.

Also, a successful innovation is not necessarily preceded by a large R&D investment and some large R&D investments never result in successful innovations.

15 Thompson and Thompson (1985) introduced the idea of occupations within industries with the argument that it would better measure skills in industries.

29

Jönköping International Business School

6.1 Knowledge flows

Spillovers can be measured as tangible resources such as patents and patent citations, treated as endogenous to R&D input (Fischer & Varga, 2003; Jaffe,

1989). One type of knowledge spillovers if the one embedded in people. A second type of spillover is knowledge embedded in high-quality products which spills over through imports and can be transmitted into the firm’s scope of products and exports

Qualitative measures of export performance such as international reputation and foreign perception are rarely used in economic literature. Instead, most analysts use quantitative measures such as export intensity, export participation, number of exporting countries, export value, export volume and export unit value. By using export unit values we account both values and volumes and reduce the risk of asymmetries across regions, industries and firms. Unit analysis is a widely used proxy for export performance but early critical studies highlight a few problems dealing with product heterogeneity (Kravis & Lipsey,

1985). One problem related to unit values is the growth of the service sector, which has increased its share of exports. The quality of products is problematic to compare in the service sector but even more so to compare them with manufacturing products. The second issue is the possible problem of using time series. To obtain accurate unit values over time there is a trade-off between frequency of shipments, many export destinations and industry variation (Knetter, 1993).

To follow up the problems of heterogeneity, we can initially assume that larger regions export more in absolute values. Regardless of measurement technique we refer to an export differentiation on the intensive margin (Besedeš &

Prusa, 2011; Helpman et al., 2008). That is, when export is proportionate to regional size, we do not give any weight to the number of export varieties.

There is no consensus around this concept in the literature on international trade and instead, emphasis is also placed on export differentiation at the

extensive margin. Models of trade and monopolistic competition commonly emphasise export differentiation in terms of both size and variety of products and stress differentiation at the extensive margin and its significance for export growth (Hummels & Klenow, 2005).

16

Measures of output of knowledge production are often accused of measuring quantity rather than quality. Models of vertical differentiation argue

16

Armington (1969) emphasise the intensive margin in terms of national differentiation. In contrast, by defining trade in models of monopolistic competition Krugman (1981) stress the importance of differentiation at the extensive margin where economies twice the size produce twice the range of products.

30

Introduction and summary of the thesis

that richer countries export products of higher quality (Flam & Helpman, 1987;

Grossman & Helpman, 1991a). Differentiation at the quality margin can be interpreted as an opportunity to charge higher prices, and large exporters can, through scale economies, supply many export varieties to many recipients, keeping their high prices without affecting demand. The present thesis deals with export heterogeneity and quality differences by using high-quality exports.

This rests on the assumption that the effects of knowledge input also accounts for quality of knowledge output. We assume that import acts as knowledge input in the production process and is a way to treat dynamics in the stock of knowledge assets. High-quality exports are the knowledge output in this production process. Indeed, we lose the renewal of non-exported products but this is a smaller issue compared to if we would also account for products with no commercial substance.

6.2 Firm performance

The problem of analysis knowledge inflows is how to measure the outcome, i.e., knowledge output. The fifth paper of this dissertation deals with how labour inflow affects the firm productivity when productivity is measured as valued added per employee. If one uses productivity to measure the increased quality per unit of knowledge input, product innovation can be one way t raise firm productivity.

Syverson (2011) argues that such innovation can be captured in standard revenue-based productivity measures. The effect of firm productivity by product innovations has been analysed in several recent studies. First,

Balasubramanian and Sivadasan (2011) relate patents to production activities and show that the number of patents granted is positively related to firm size, scope and total factor productivity (TFP). Lentz and Mortensen (2008) use firm-level data to show that greater part of productivity growth comes from reallocation of employment to innovating firms. Finally, Bartel et al. (2005) show that recent advances in technology have caused an IT-based productivity growth which is generated by an increased ability to customise products and services.

Value added is a well-established measure of productivity and this is used as the firm performance measure in Paper 5 in this dissertation. However, one needs to emphasize some drawbacks, especially when making cross-regional and cross-industrial comparisons. The first is the extent of industry concentration and variations in market power across existing producers. The second is the risk of incorporating structural differences in productivity levels

31

Jönköping International Business School

across regions and industries. Differences in competitive level, use of production factors and consumer demand can for instance affect wage levels and also productivity levels. The third possible drawback is related to the fast growth of the service sector in all developed economies. As claimed by

Griliches (1979), output and quality are more difficult to measure in firms in the service industry than in firms in the manufacturing industry. Service output cannot be quantified in volumes since many of services have become knowledge-intensive, limiting the usage of productivity growth. This means that, for instance product renewals or product innovations (quality improvements) may not raise the quantity of output (measured in physical units) per unit input but increases the product price (Syverson, 2011).

6.3 Data

This thesis consists of four empirical papers. Papers 3 and 4 are based on

Swedish trade data provided by Statistics Sweden, covering the period 1997 to

2003. The data hold extensive and detailed information on exports and imports for all traded products defined as 8-digit CN codes. For each export product one can, in addition to finding the location of the exporting firm, also extract export volume, export value and export destination. Equivalent to this, for each import product we can find value, volume and country of import origin.

Papers 2 and 5 are based on longitudinal data also provided by Statistics

Sweden. The data contain information on all employed individuals in Sweden between 1986 and 2008 with information on individual characteristics such as gender and age but also on work-place, occupation, work-place location residence location and education. The first year available for occupational data is 2001 and this is the starting year of the empirical analysis in Paper 5. Work place data and regional data can thereafter be linked to work place characteristics and regional characteristics respectively.

In an analysis of knowledge diffusion and knowledge flows geographical scale need to be considered with care. The geographical unit of analysis in all papers in this dissertation is functional regions. These are referred to as local labour market regions (LLM) equivalent to metropolitan areas in the USA.

These regions are composed of a number of municipalities (urban regions). A functional region is characterised by high intensity of intra-regional commuting flows and are delineated based on the intensity of observed commuting flows between municipalities (Nutek, 1998). The approximated travel time distance is

20 to 30 minutes within the LLMs and travel time between the two locations within a functional region rarely exceeds 50 minutes (Johansson et al., 2002,

32

Introduction and summary of the thesis

2003; Karlsson & Manduchi, 2001). This structuring of regions implies that the central municipality is characterised by not only its concentration of production and consumption activities but also the way infrastructure is construed. There is a high interaction frequency within the borders of functional regions while the interaction drops sharply when exiting the regional borders.

7. Outline, summary results and contributions of the thesis

Paper 2, Where do university graduates go? raises three questions associated with

Swedish graduates: (i) Where do university graduates prefer to reside after graduation?, (ii)

Are the residence decisions of graduate movers different across study areas? and (iii) How do

push factors, in this case the characteristics of the graduation region, affect the residence decision for graduate movers?

Knowledge is important for regional economic growth, and one type of knowledge is the knowledge embodied in university graduates. The Swedish educational system has gone through major changes during the past 30 years in the direction towards decentralisation and increased access to colleges and universities. In Government Bill 1975:9 one of the main purposes of such reform was to increase access to education and by doing so increase the social levelling.

For knowledge distribution this can have an impact on local knowledge supply but only as long as the region is an attractive residential choice also after graduation. If regions cannot produce knowledge internally, they become highly dependent on the distribution of new graduates. Previous literature on migration and location decisions predominantly focuses on characteristics of the destination regions and what regional attributes can be defined as pull factors (Florida, 2002c; Glaeser et al., 2001; Gottlieb & Joseph,

2006). Compared to this, there is little information on what makes people stay , what attributes can be defined as push factors and how the decisions differ across study areas. This is where Paper 2 positions itself in the existing literature. To answer the stated questions, Sweden is divided into three types of regions where graduates, if they do not choose to stay in their graduation region, choose to reside after graduation; (i) Large regions, comprising Stockholm,

Göteborg and Malmö; (ii) Growth regions, which are characterised by a growing population and positive economic growth and; (iii) Decline regions, which are those with a decreasing population. Using a multinomial logit model, these

33

Jönköping International Business School

residence choices are tested against a number of individual characteristics and a number of push factors describing the graduation region.

From the results in Paper 2 one can conclude that decline regions (rural areas) in Sweden have major difficulties in attracting new graduates. In general terms one can say that push factors associated with larger and denser regions, such as a large and differentiated labour market and a large service sector, keep graduates in their graduation region. The residence decisions differ across study areas and students who are likely to end up with a job in the public sector are the only graduates that are pushed out to more rural areas in Sweden. Declining regions are not very successful in attracting graduates with jobs in more competitive sectors. These people, for instance graduates in social- and natural sciences, rather stay in their graduation region or move to large or growing regions. From the point of view of regional economic policy, there is one indirect conclusion one can draw from these results: a decentralisation of higher education does not solve the issue of regional knowledge supply per se.

The purpose of Paper 3, Export dynamics and product relocation, is to an analysis export dynamics in the manufacturing industry and to examine what regional characteristics attract relocated exports coming from other regions in Sweden between 1997 and 2003. Export products enter and exit regions as a natural consequence of product maturity, profit maximising firms, and regional characteristics. Being involved in international trade forces firms to be good at export base renewal and one way to renew the regional export base is to absorb products coming from other regions. Thus, a region’s composition of export products is dynamic and over time, old export products are replaced by new export products or new export product varieties.

In this paper, an export product is defined as a combination of a firm identity code and an 8-digit product identity code and the products is completely relocated if it leaves the export base in the exit region and is new for the entry region. Hence, it is not a matter of product imitation and this is partly the research contribution of this paper. It gives new insights into the relocation of products from the perspective of sectors in the Swedish manufacturing industry. To the author’s knowledge, no one has previously studied the interregional relocation of combinations of firms and products, that is where one can exclude a pure imitation of an export product by a firm in one region from another firm in another region.

The relationship between the number of relocated export products entering a region and regional pull factors is analysed by a zero-inflated Poisson model.

The results indicate that the regional share of within-sector knowledge is a

34

Introduction and summary of the thesis

significant factor of a high absorption capacity. The result is even further strengthened since the results also show that non-sector-related knowledge can have a negative effect on absorptive capacity. Looking at the distribution of export product relocation in Sweden one can see that many regions are entry regions and exit regions while other regions fail in both categories.

From the point of view of regional economic policy, the results in Paper 3 show that entry regions are not necessarily characterised by a high regional knowledge base per se but what attracts relocated exports is an specialised export support system and access to sector specific knowledge. Much export product relocation seems to follow the product cycle theory according to which products relocate when they are in a phase of standardisation where new knowledge is less important. Thus, these products need access to specialised trade infrastructure networks and an industrial business milieu.

Paper 4, Imports, knowledge flows and the renewal of regional exports (co-authored with Martin Andersson and Charlie Karlsson), is a study of the relationship between regional import flows and the renewal of regional exports. It is related to Paper 3 in the sense that they both deal with regional export renewal and the regional economic milieu. However, while Paper 3 focuses on export renewal through inter-regional export product relocation, Paper 4 sees imports as one of the main input factors in export renewal. The main contribution of Paper 4 is the way it focuses on the role of high-quality imports in high-quality export renewal. That is the role of high-quality knowledge inflows and high-quality knowledge output, where input and output in knowledge production are measured with high-quality trade.

While the main focus is on the effect of import diversity and high-quality imports, the analysis also controls for other characteristics that may influence regional renewal, such as regional knowledge intensity, measures of agglomeration economies, distance to metropolitan regions and local presence of advanced business services. To examine potential spillover effects across sectors in a region, high-quality imports and human capital in other sectors are also included as regressors and the results show that it is a positive relation between high-quality imports and novel export products. The results also suggest spillover effects within sectors but not across sectors, and the firms involved in high-quality novel exports do not seem to be dependent on knowledge-intensive business services. Firm size is also controlled for and is negatively correlated with high-quality export renewal. One can therefore argue that small and medium-sized firms are relatively better on new exports than their larger competitors.

35

Jönköping International Business School

Hence, from the results one can conclude that the quality of a region’s imports has a significant influence on the export renewal of industries and regions. This means that conceptualisations and models of regional growth need to a larger extent merge the ideas of local and global innovation networks and their interplay. From the point of view of regional economic policy, this suggests that imports should be emphasised as playing an important role in regional innovation and high-quality exports renewal.

Paper 5, Knowledge intensive business services, creative labour inflow and firm

productivity attempts to find the relation between firm productivity and creative labour inflows in knowledge-intensive business services. The underlying hypothesis is that labour acts as a channel for diffusion of knowledge and it is an empirical study using Swedish data for the seven year period 2001-2008. The results show that labour inflow per se does not have a positive effect on firm productivity but rather that it depends on the type of labour inflow. Creative labour inflow has a positive effect on KBS- and non-KIBS firms, but the effect is significantly stronger for firms in the KIBS sector than for firms outside the

KIBS sector, controlling for firm and regional characteristics. This paper is novel in the sense that it examines the effect on firm productivity on KIBS firms and non-KIBS firms. This places it in the literature on KIBS firms and how they are characterised as intensive users of a specific type of knowledge.

They are primarily involved in tasks around complex problem solving in very close collaboration with their customers and clients. Thus the second distinct feature of Paper 5 is how it relates the characteristics of KIBS firms to the characteristics of creative labour.

36

Introduction and summary of the thesis

References

Abernathy, W. J. (1978). The Productivity Dilemma: Roadblock to Innovation in the Automobile

Industry. Baltimore, MD: Johns Hopkins University Press.

Abernathy, W. J., & Utterback, J. M. (1978). Patterns of Industrial Innovation.

Technology Review, 80(7), 40-47.

Acs, Z. (2002). Innovation and the Growth of Cities. Cheltenham: Edward Elgar.

Acs, Z., Anselin, L., & Varga, A. (2002). Patents and innovation counts as measures of regional production of new knowledge. Research Policy, 31(7), 1069-1085.

Acs, Z. J., Audretsch, D. B., & Feldman, M. P. (1992). Real Effects of Academic

Research: Comment. The American Economic Review, 82(1), 363-367.

Acs, Z. J., Audretsch, D. B., & Feldman, M. P. (1994). R & D Spillovers and Recipient

Firm Size. The Review of Economics and Statistics, 76(2), 336-340.

Aghion, P., & Howitt, P. (1992). A Model of Growth through Creative Destruction.

Econometrica, 60(2), 323-351.

Albert, M. B., Avery, D., Narin, F., & McAllister, P. (1991). Direct validation of citation counts as indicators of industrially important patents. Research Policy, 20(3), 251-259.

Allen, T. J. (1977). Managing the Flow of Technology. Cambridge, MA: MIT Press.

Almeida, P., & Kogut, B. (1999). Localization of Engineers in Regional Networks.

Management Science, 45(7), 905-917.

Andersson, M., Gråsjö, U., & Karlsson, C. (2007). Human Capital and Productivity in a

Spatial Economic System: Relating the Extent of Spatial Dependence to Localities'

Position. Annales d'Economie et de Statistique(87-88), 125-144.

Andersson, Å. E. (1985). Kreativitet- storstadens framtid: En bok om Stockholm [in Swedish].

Stockholm: Prisma.

Andersson, Å. E., & Beckmann, M. J. (2009). Economics of Knowledge: Theory, Models and

Measurements. Cheltenham, UK: Edward Elgar.

Andersson, Å. E., & Johansson, B. (1984). Knowledge Intensity and Product Cycles in

Metropolitan Regions IIASA Working Paper 84-13. Laxenburg, Austria: International

Institute for Applied Systems Analysis.

Andersson, Å. E., & Mantsinen, J. (1980). Mobility of resources, accessibility of knowledge, and economic growth. Behavioral Science, 25(5), 353-366.

Anselin, L., Varga, A., & Acz, Z. (1997). Local Geographic Spillovers between

University Research and High Technology Innovations. Journal of Urban Economics,

42(3), 422-448.

Antonelli, C. (1994). Technological districts, localized spillovers and productivity growth: The Italian evidence on technological externalities in the core regions.

International Review of Applied Economics, 8(1), 18-30.

Armington, P. S. (1969). A Theory of Demand for Products Distinguished by Place of

Production. International Monetary Fund Staff Papers, 16(1), 159-178.

Arrow, K. J. (1962a). The Economic Implications of Learning by Doing. Review of

Economic Studies, 29(3), 155-173.

Arrow, K. J. (1962b). Economic Welfare and the Allocation of Resources for

Invention. In R. Nelson (Ed.), The rate and direction of inventive activity: Economic and

social factors (pp. 609-626). Princeton, NJ: Princeton University Press.

Arrow, K. J. (1973). Higher education as a filter. Journal of Public Economics, 2(3), 193-

216.

37

Jönköping International Business School

Arrow, K. J., Chenery, H. B., Minhas, B. S., & Solow, R. M. (1961). Capital-Labor

Substitution and Economic Efficiency. Review of Economics and Statistics, 43(3), 225-

250.

Asheim, B. T. (1999). Interactive learning and localised knowledge in globalising learning economies. GeoJournal, 49(4), 345-352.

Audretsch, D. B., & Feldman, M. P. (1996). R&D Spillovers and the Geography of

Innovation and Production. American Economic Review, 86(3), 630-640.

Backman, M., & Bjerke, L. (2009). Returns to Higher Education-a regional perspective.

CESIS Electronic Working Paper Series, Paper No.171. Centre of Excellence for

Science and Innovation Studies, Royal Institute of Technology. Stockholm.

Bain, J. S. (1956). Barriers to new Competition: Their Character and Consequences in

Manufacturing Industries. Cambridge, MA: Harvard University Press.

Balasubramanian, N., & Sivadasan, J. (2011). What Happens When Firms Patent? New

Evidence from U.S. Economic Census Data. Review of Economics and Statistics, 93(1),

126-146.

Bartel, A. P., Ichniowski, C., & Shaw, K. L. (2005). How Does Information Technology Really

Affect Productivity? Plant-Level Comparisons of Product Innovation, Process Improvement and

Worker Skills. NBER Working Paper, No. 11773. Cambridge, MA: National Bureau of Economic Research

Becker, G. S. (1962). Investment in Human Capital: A Theoretical Analysis. Journal of

Political Economy, 70(5), 9-49.

Becker, G. S. (1964). Human Capital: A Theoretical and Empirical Analysis with Special

Reference to Education. Chicago, IL: Chicago University Press.

Becker, G. S. (1965). A Theory of the Allocation of Time. The Economic Journal, 75(299),

493-517.

Becker, G. S. (1993). Human Capital: A Theoretical and Empirical Analysis with Special

Reference to Education (3rd ed.). Chicago, IL: The University of Chicago Press.

Bernard, A. B., Jensen, J. B., & Schott, P. K. (2006). Trade costs, firms and productivity. Journal of Monetary Economics, 53(5), 917-937.

Besedeš, T., & Prusa, T. J. (2011). The role of extensive and intensive margins and export growth. Journal of Development Economics, 96(2), 371-379.

Braconier, H. (1998). Essays on R&D and growth. Lund Economic Studies. Department of Economics. Lund University, Lund.

Bradley, S., & Taylor, J. (1996). Human Capital Formation and Local Economic

Performance. Regional Studies, 30(1), 1-14.

Brennenraedts, R. M. F., Bekkers, R., & Verspagen, B. (2006). The different channels of

university-industry knowledge transfer: Empirical evidence from Biomedical Enginering.

Eindhoven Centre for Innovation Studies Working Paper 06.04.

Breschi, S., & Lissoni, F. (2001a). Knowledge Spillovers and Local Innovation Systems:

A Critical Survey. Industrial and Corporate Change, 10(4), 975-1005.

Breschi, S., & Lissoni, F. (2001b). Localised knowledge spillovers vs. innovative milieux: Knowledge “tacitness” reconsidered. Papers in Regional Science, 80(3), 255-

273.

Budd, L., & Hirmis, A. (2004). Conceptual Framework for Regional Competitiveness.

Regional Studies, 38(9), 1015-1028.

Castro, E. d., & Jensen-Butler, C. (1993). Flexibility, routine behaviour and the neoclassic model in the analysis of regional growth. University of Aarhus, Denmark:

Department of Political Science.

38

Introduction and summary of the thesis

Caves, R. E. (1980). Industrial Organization, Corporate Strategy and Structure. Journal of

Economic Literature, 18(1), 64-92.

Ciccone, A., & Hall, R. E. (1996). Productivity and the Density of Economic Activity.

American Economic Review, 86(1), 54-70.

Clerides, S. K., Lach, S., & Tybout, J. R. (1998). Is Learning by Exporting Important?

Micro-Dynamic Evidence from Colombia, Mexico, and Morocco. Quarterly Journal of

Economics, 113(3), 903-947.

Coe, D. T., & Helpman, E. (1995). International R&D spillovers. European Economic

Review, 39(5), 859-887.

Cohen, W. M., & Levinthal, D. A. (1989). Innovation and Learning: The Two Faces of

R & D. The Economic Journal, 99(397), 569-596.

Cohen, W. M., & Levinthal, D. A. (1990). Absorptive Capacity: A New Perspective on

Learning and Innovation. Administrative Science Quarterly, 35(1), 128-152.

Dasgupta, P., & Stiglitz, J. (1988). Learning-by-Doing, Market Structure and Industrial and Trade Policies. Oxford Economic Papers, 40(2), 246-268.

Dierickx, I., & Cool, K. (1989). Asset Stock Accumulation and Sustainability of

Competitive Advantage. Management Science, 35(12), 1504-1511.

Dixit, A. K., & Stiglitz, J. E. (1977). Monopolistic Competition and Optimum Product

Diversity. American Economic Review, 67(3), 297-308.

Dornbusch, R. (1994). The Fight over Competitiveness: A Zero-sum Debate? Foreign

Affairs, 73 (4), 186-197.

Duranton, G., & Puga, D. (2001). Nursery Cities: Urban Diversity, Process Innovation, and the Life Cycle of Products. American Economic Review, 91(5), 1454-1477.

Duranton, G., & Puga, D. (2004). Micro-foundations of Urban Agglomeration

Economies. In J. V. Henderson & J.-F. Thisse (Eds.), Handbook of Regional and Urban

Economics. Vol. 4: Cities and Geography (pp. 2063-2117). Amsterdam: Elsevier.

Eklund, E. J., & Wiberg, D. (2008). R&D and the Persistence of Profits. Journal of

Managerial Economics, 6(2), 40-53.

Evans, A. W. (1990). The assumption of equilibrium in the analysis of migration and interregional differences: a review of some recent research. Journal of Regional Science,

30(4), 515-531.

Faggian, A., & McCann, P. (2006a). Human Capital Flows and Regional Knowledge

Assets: A Simultaneous Equation Approach. Oxford Economic Papers, 58(3), 475-500.

Faggian, A., & McCann, P. (2006b). Human capital flows and regional knowledge assets: a simultaneous equation approach. Oxf. Econ. Pap., 58(3), 475-500.

Faggian, A., McCann, P., & Sheppard, S. (2006). An analysis of ethnic differences in

UK graduate migration behaviour. Annals of Regional Science, 40(2), 461-471.

Faggian, A., McCann, P., & Sheppard, S. (2007). Some evidence that women are more mobile than men: gender differences in U.K. graduate migration behavior. Journal of

Regional Science, 47(3), 517-539.

Falvey, R., Foster, N., & Greenaway, D. (2004). Imports, exports, knowledge spillovers and growth. Economics Letters, 85(2), 209-213.

Feldman, M. P. (1994a). The Geography of Innovation. Dordrecht: Kluwer Academic

Publishers.

Feldman, M. P. (1994b). Knowledge complementarity and innovation. Small Business

Economics, 6(5), 363-372.

Feldman, M. P., & Audretsch, D. B. (1999). Innovation in cities: Science-based diversity, specialization and localized competition. European Economic Review, 43(2),

409-429.

39

Jönköping International Business School

Fischer, M. M., & Fröhlich, J. E. (2001). Knowledge, Complexity and Innovation Systems.

Germany and New York, NY: Springer-Verlag.

Fischer, M. M., & Varga, A. (2003). Spatial knowledge spillovers and university research: Evidence from Austria. Annals of Regional Science, 37(2), 303-322.

Flam, H., & Helpman, E. (1987). Vertical Product Differentiation and North-South

Trade. American Economic Review, 77(5), 810-822.

Florida, R. (2002a). Bohemia and economic geography. Journal of Economic Geography.

Florida, R. (2002b). The Economic Geography of Talent. Annals of the Association of

American Geographers, 92(4), 743-755.

Florida, R. (2002c). The Rise of the Creative Class: and how it's transforming work, leisure,

community & everyday fife. New York, NY: Basic Books.

Foray, D., & Lundvall, B. Å. (1996). The Knowledge-based Economy: From the

Economics of Knowledge to the Learning Economy. In D. Foray & B. Å. Lundvall

(Eds.), Employment and Growth in the Knowledge-based Economy. Paris: OECD.

Fujita, M., & Thisse, J.-F. (2002). Economics of Agglomeration: Cities, Industrial Location, and

Regional Growth. Cambridge, UK: Cambridge University Press.

Gertler, M. S. (2003). Tacit knowledge and the economic geography of context, or the undefinable tacitnessof being (there). Journal of Economic Geography, 3(1), 75-99.

Glaeser, E. (1999). Learning in Cities. Journal of Urban Economics 46(2), 254-277.

Glaeser, E. L. (2005). Review of Richard Florida's The Rise of the Creative Class. Regional

Science and Urban Economics, 35(5), 593-596.

Glaeser, E. L., Kolko, J., & Saiz, A. (2001). Consumer city. Journal of Economic Geography,

1(1), 27-50.

Glaeser, E. L., & Saiz, A. (2004). The Rise of the Skilled City [with Comments by G.T

Burtless & W.S Strange]. Brookings-Wharton Papers on Urban Affairs, 5, 47-105.

Gordon, I. R., & McCann, P. (2000). Industrial Clusters: Complexes, Agglomeration and/or Social Networks? Urban Studies, 37(3), 513-532.

Gottlieb, P. D. (1995). Residential Amenities, Firm Location and Economic

Development. Urban Studies, 32(9), 1413-1436.

Gottlieb, P. D., & Joseph, G. (2006). College-to-work migration of technology graduates and holders of doctorates within the United States. Journal of Regional

Science, 46(4), 627-659.

Griliches, Z. (1979). Issues in Assessing the Contribution of Research and

Development to Productivity Growth. The Bell Journal of Economics, 10(1), 92-116.

Griliches, Z. (1984). R&D, Patents and Productivity. Chicago, IL: University of Chicago

Press.

Griliches, Z. (1986). Productivity, R and D, and Basic Research at the Firm Level in the

1970's. American Economic Review, 76(1), 141-154.

Griliches, Z. (1990). Patent Statistics as Economic Indicators: A Survey. Journal of

Economic Literature, 28(4), 1661-1707.

Grossman, G. M., & Helpman, E. (1991a). Innovation and Growth in the Global Economy.

Cambridge, MA: MIT Press.

Grossman, G. M., & Helpman, E. (1991b). Trade, knowledge spillovers, and growth.

European Economic Review, 35(2-3), 517-526.

Gråsjö, U. (2005). The importance of accessibility to R&D on patent production in Swedish

municipalities.

Gråsjö, U. (2006). University Educated Labour, R&D and Regional Export

Performance Spatial Spillovers of Knowledge Production- An Accessibility Approach (Vol.

No.034): JIBS Dissertation Series

40

Introduction and summary of the thesis

Haig, R. M. (1926). Toward an Understanding of the Metropolis: I. Some Speculations

Regarding the Economic Basis of Urban Concentration. Quarterly Journal of

Economics, 40(2), 179-208.

Harrigan, F. J., & McGregor, P. G. (1993). Equilibrium and Disequilibrium

Perspectives on Regional Labor Migration. Journal of Regional Science, 33(1), 49-67.

Harrod, R. F. (1939). An Essay in Dynamic Theory. Economic Journal, 49(193), 14-33.

Hayek, F. A. (1937). Economics and Knowledge. Economica, 4(13), 33-54.

Heckscher, E. F. (1931). Merkantilismen [in Swedish] Stockholm: Norstedt & Söner.

Heckscher, E. F. (1957). Svenskt arbete och liv- från medeltiden till nutiden [in Swedish].

Stockholm: Bonnier.

Helpman, E., Melitz, M., & Rubinstein, Y. (2008). Estimating Trade Flows: Trading

Partners and Trading Volumes. Quarterly Journal of Economics, 123(2), 441-487.

Henderson, R., Jaffe, A. B., & Trajtenberg, M. (1998). Universities as a Source of

Commercial Technology: A Detailed Analysis of University Patenting, 1965-1988.

Review of Economics and Statistics, 80(1), 119-127.

Hicks, J., R., (1965). Capital and growth. Oxford, UK: Clarendon.

Hoyman, M., & Faricy, C. (2009). It Takes a Village: A Test of the Creative Class, Social

Capital, and Human Capital Theories. Urban Affairs Review, 44(3), 311-333.

Hudson, R. (2005). Economic Geographies: Circuits, Flows and Spaces. London, UK: SAGE

Publications.

Hummels, D., & Klenow, P. J. (2005). The Variety and Quality of a Nation's Exports.

American Economic Review, 95(3), 704-723.

Jacobs, J. (1961). The Death and Life of Great American Cities. New York, NY: Random

House.

Jacobs, J. (1969). The Economy of Cities. New York, NY: Random House.

Jaffe, A., B., Trajtenberg, M., & Henderson, R. (1993). Geographic Localization of

Knowledge Spillovers as Evidenced by Patent Citations. Quarterly Journal of

Economics, 108(3), 577-598.

Jaffe, A. B. (1989). Real Effects of Academic Research. American Economic Review, 79(5),

957-970.

Johansson, B. (1993). Ekonomisk dynamik i Europa: nätverk för handel, kunskapsimport och

innovationer [in Swedish]. Malmö: Liber-Hermods i samarbete med Institutet för framtidsstudier.

Johansson, B. (2005). Parsing the Menagerie of Agglomeration and Network

Externalities. In C. Karlsson, B. Johansson & R. R. Stough (Eds.), Industrial Clusters

and Inter-firm Networks (pp. 107-147). Cheltenham, UK: Edward Elgar.

Johansson, B., & Andersson, Å. E. (1998). A Schloss Laxenburg Model of Product

Cycle Dynamics. In M. J. Beckmann, B. Johansson, F. Snickars & R. Thord (Eds.),

Knowledge and Networks in a Dynamic Economy: Festschrift in Honor of Ake E. Andersson

(pp. 181-219). Berlin: Springer Verlag.

Johansson, B., & Karlsson, C. (1987). Processes of Industrial Change: Scale, Location and Type of Job. In M. M. Fisher & P. Nijkamp (Eds.), Regional Labour Markets:

analytical contributions and cross-national comparisons (pp. 139-165). Amsterdam: North-

Holland.

Johansson, B., & Karlsson, C. (2003). Växande branscher-om Stockholmsregionens samspel med

övriga landet [in Swedish]. Storstadspolitik 8:2003. Stockholms läns landsting:

Regionplane- och trafikkontoret.

Johansson, B., Karlsson, C., & Stough, R. (2001). Theories of Endogneous Regional Growth.

Lessons for Regional Policies. Berlin: Springer-Verlag.

41

Jönköping International Business School

Johansson, B., Klaesson, J., & Olsson, M. (2002). Time distances and labor market integration. Papers in Regional Science, 81(3), 305-327.

Johansson, B., Klaesson, J., & Olsson, M. (2003). Commuters’ non-linear response to time distances. Journal of Geographical Systems, 5(3), 315-329.

Johansson, B., & Lööf, H. (2006). Innovation activities explained by firm attributes and location. In CESIS Electronic Working Paper Series, No.63 (Ed.). Stockholm: Royal

Institute of Technology, Centre of Excellence for Studies in Innovation and

Science.

Johansson, B., & Lööf, H. (2008). The impact offFirm’s R&D strategy on profit and productivity. In CESIS Electronic Working Paper Series, No.156 (Ed.): Royal Institute of Technology, Centre of Excellence for Studies in Innovation and Science.

Johansson, B., & Quigley, J. M. (2003). Agglomeration and networks in spatial economies. Papers in Regional Science, 83(1), 165-176.

Kaldor, N. (1970). The case for regional policies. Scottish Journal of Political Economy,

17(3), 337-348.

Karlsson, C., & Manduchi, A. (2001). Knowledge Spillovers in a Spatial Context: A

Critical Review and Assessments. In M. Fischer & J. Fröhlich (Eds.), Knowledge,

Complexity and Innovation Systems (pp. 101-123). Berlin and Heidelberg, Germany and

New York, NY: Springer.

Keller, W. (2004). International Technology Diffusion. Journal of Economic Literature,

42(3), 752-782.

Kelley, A. C., & Weiss, L. W. (1969). Markov Processes and Economic Analysis: The

Case of Migration. Econometrica, 37(2), 280-297.

Kitson, M., Martin, R., & Tyler, P. (2004). Regional Competitiveness: An Elusive yet

Key Concept? Regional Studies, 38(9), 991 - 999.

Kleinknecht, A., & Poot, T. P. (1992). Do Regions Matter for R&D? Regional Studies,

26(3), 221-232.

Klette, T. J., & Kortum, S. (2004). Innovating Firms and Aggregate Innovation Journal

of Political Economy, 112(5), 986-1018.

Knetter, M. M. (1993). International Comparisons of Pricing-to-Market Behavior.

American Economic Review, 83(3), 473-486.

Kortum, S. (2008). Exploring Innovation with Firm Level Data The Conference Board

Economics Program Working Papers No. 08-11. New York, NY: The Conference Board.

Kravis, I. B., & Lipsey, R. E. (1985). Prices and Terms of Trade for Developed-Country Exports

of Manufactured Goods. National Bureau of Economic Research. Cambridge, MA.

Krugman, P. (1991a). Geography and Trade. Cambridge, MA: MIT Press.

Krugman, P. (1991b). Increasing Returns and Economic Geography. Journal of Political

Economy, 99(3), 483-499.

Krugman, P. (1996). Making sense of the competitiveness debate. Oxford Review of

Economic Policy, 12(3), 17-25.

Krugman, P. R. (1979). Increasing returns, monopolistic competition, and international trade. Journal of International Economics, 9(4), 469-479.

Krugman, P. R. (1981). Intraindustry Specialization and the Gains from Trade. Journal of

Political Economy, 89(5), 959-973.

Lampard, E. E. (1955). The History of Cities in the Economically Advanced Areas.

Economic Development and Cultural Change, 3(2), 81-136.

Lancaster, K. (1975). Socially Optimal Product Differentiation. American Economic

Review, 65(4), 567-585.

42

Introduction and summary of the thesis

Lancaster, K. J. (1966). A New Approach to Consumer Theory. Journal of Political

Economy, 74(2), 132-157.

Lawson, C. (1999). Towards a competence theory of the region. Cambridge Journal of

Economics, 23(2), 151-166.

Lentz, R., & Mortensen, D. T. (2008). An Empirical Model of Growth Through

Product Innovation. Econometrica, 76(6), 1317-1373.

Lucas Jr., R. E. (1988). On the mechanics of economic development. Journal of Monetary

Economics, 22(1), 3-42.

Machlup, F. (1962). The Supply of Inventors and Inventions. In R. R. Nelson (Ed.), The

Rate and Direction of Inventive Activity: Economic and Social Factors (pp. 143-170).

Cambridge, MA: National Bureau of Economic Research.

Markusen, A. (2006). Urban development and the politics of a creative class: evidence from a study of artists. Environment and Planning A, 38(10), 1921-1940.

Marshall, A. (1890/1920). Principles of Economics. London, UK: Macmillan.

Marx, K. (1904). A Contribution to the Critique of Political Economy. Chicago, IL: Kerr.

Maurseth, P. B., & Verspagen, B. (1999). Europe: one or several systems of innovation?

An analysis based on patent citations. In J. Fagerberg, P. Guerrieri & B. Verspagen

(Eds.), The Economic Challenge for Europe: Adapting to Innovation-Based Growth (pp. 149-

174). Cheltenham, UK and Northampton, MA: Edward Elgar.

McCann, P. (2001). Urban and Regional Economics. Oxford, UK: Oxford University Press.

McKelvey, M. (1994). Evolutionary Innovation: Early Industrial Use of Genetic Engineering.

Ph.D. Dissertation, Linköping University, Linköping.

Mellander, C. (2009). Creative and Knowledge Industries: An Occupational

Distribution Approach. Economic Development Quarterly, 23(4), 294-305.

Mincer, J. (1958). Investment in Human Capital and Personal Income Distribution.

Journal of Political Economy, 66(4), 281-302.

Mincer, J. (1991). Education and unemployment. NBER Working Paper No. 3838.

Cambridge, MA: National Bureau of Economic Research.

Mowery, D. C. (1983). The relationship between intrafirm and contractual forms of industrial research in American manufacturing, 1900-1940. Explorations in Economic

History, 20(4), 351-374.

Muth, R. F. (1971). Migration: Chicken or Egg? Southern Economic Journal, 37(3), 295-

306.

Nelson, R. (1993). National Innovation Systems: A Comparative Analysis. Oxford, UK and

New York, NY: Oxford University Press.

Nelson, R. R. (1959). The Simple Economics of Basic Scientific Research. Journal of

Political Economy, 67(3), 297-306.

Nelson, R. R., & Rosenberg, N. (1993). Technical Innovation and Natinoal Systems. In

R. R. Nelson (Ed.), National Innovation Systems: A Comparative Analysis (pp. 3-21).

Oxford, UK and New York, NY: Oxford University Press.

Nelson, R. R., & Winter, S. G. (1974). Neoclassical vs. Evolutionary Theories of

Economic Growth: Critique and Prospectus. The Economic Journal, 84(336), 886-905.

Nelson, R. R., & Winter, S. G. (1982). An Evolutionary Theory of Economic Change

Cambridge, MA Harvard University Press.

Nonaka, I. (1994). A Dynamic Theory of Organizational Knowledge Criterion.

Organization Science, 5(1), 14-37.

Norton, R. D., & Rees, J. (1979). The product cycle and the spatial decentralization of

American manufacturing. Regional Studies, 13(2), 141-151.

43

Jönköping International Business School

Nutek. (1998). Småföretag och regioner i Sverige 1998: med ett tillväxtperspektiv för hela landet.

Report B 1998:10. [in Swedish]. Stockholm: Närings- och teknikutvecklingsverket:

Nutek.

Pakes, A., & Schankerman, M. (1984). The Rate of Obsolescence of Patents, Research

Gestation Lags, and the Pivate Rate of Return to Research Resources. In Z.

Griliches (Ed.), R&D, Patents and Productivity (pp. 73-88). Chicago, IL: University of

Chicago Press.

Pavcnik, N. (2002). Trade Liberalization, Exit, and Productivity Improvements:

Evidence from Chilean Plants. Review of Economic Studies, 69(1), 245-276.

Polanyi, M. (1958). Personal Knowledge: Towards a Post-Critical Philosophy. London, UK:

Routledge &Kegan Paul.

Polanyi, M. (1962). Tacit Knowing: Its Bearing on Some Problems of Philosophy.

Reviews of Modern Physics, 34(4), 601-616.

Polanyi, M. (1966a). The Logic of Tacit Inference. Philosophy, 41(155), 1-18.

Polanyi, M. (1966b). The Tacit Dimension. London, UK: Routledge & Kegan Paul.

Polanyi, M. (1969). Knowing and being: Essays by Michael Polanyi. London: Routledge &

Kegan Paul.

Popper, K. R. (1972). Objective Knowledge: An Evolutionary Approach (1 ed.). Oxford, UK:

Oxford University Press.

Porter, M. E. (1986). Competition in Global Industries. Boston, MA: Harvard Business

Press.

Porter, M. E. (1990). The Competitive Advantage of Nations. New York, NY: Free Press and

London, UK:Macmillan.

Porter, M. E. (1998). Clusters and the New Economics of Competition. Harvard Business

Review, November-December 1998, 77-90.

Porter, M. E. (2003). The Economic Performance of Regions. Regional Studies, 37(7),

545 - 546.

Roback, J. (1982). Wages, Rents, and the Quality of Life. Journal of Political Economy,

90(6), 1257-1278.

Roback, J. (1988). Wages, Rents, and Amenities: Differences among Workers and

Regions. Economic Inquiry, 26(1), 23-41.

Rodriguez-Pose, A., & Vilalta-Bufi, M. (2005). Education, migration, and job satisfaction: the regional returns of human capital in the EU. Journal of Economic

Geography, 5(5), 545-566.

Romer, D. M. (1990). Endogenous Technological Change. Journal of Political Economy,

98(5), 71-102.

Romer, P. M. (1994). The Origins of Endogenous Growth. Journal of Economic

Perspectives, 8(1), 3-22.

Roper, S., & Love, J. (2006). Innovation and regional absorptive capacity: the labour market dimension. Annals of Regional Science, 40(2), 437-447.

Rosenberg, N. (1982). Inside the Black Box: Technology and Economics. Cambridge, UK and

New York, NY: Cambridge University Press.

Saxenian, A. (1994). Regional Advantage: Culture and Competition in Silicon Valley and Route

128. Cambridge, MA: Harvard University Press.

Scherer, F. M. (1965). Corporate Inventive Output, Profits, and Growth. Journal of

Political Economy, 73(3), 290-297.

Schmookler, J. (1966). Invention and Economic Growth Cambridge, MA: Harvard

University Press.

44

Introduction and summary of the thesis

Schultz, T. W. (1961). Investment in Human Capital. American Economic Review, 51(1), 1-

17.

Schumpeter, J. A. (1934). The Theory of Economic Development: An Inquiry into Profits, Capital

Credit, Interest, and the Business Cycle. Cambridge, MA: Harvard University Press.

Schumpeter, J. (1939). Business Cycles: A Theoretical, Historical and Statistical Analysis of the

Capitalist Process. New York, NY: McGraw Hill.

Schumpeter, J. A. (1943). Capitalism, Socialism and Democracy. London, UK: Allen and

Unwin.

Schwartz, A. (1976). Migration, Age and Education. Journal of Political Economy, 84(4),

701-720.

Scitovsky, T. (1954). Two Concepts of External Economies. Journal of Political Economy,

62(2), 143-151.

Shell, K. (1966). Toward a Theory of Inventive Activity and Capital Accumulation.

American Economic Review, 56(1/2), 62-68.

Sjaastad, L. A. (1962). The Costs and Returns of Human Migration. Journal of Political

Economy, 70(5, Part 2), 80-93.

Smith, A. (1776/1904). An Inquiry into the Nature and Causes of the Wealth of Nations (Fifth ed.). London, UK.

Solow, R. M. (1956). A Contribution to the Theory of Economic Growth. Quarterly

Journal of Economics, 70(1), 65-94.

Solow, R. M. (1957). Technical Change and the Aggregate Production Function. Review

of Economics and Statistics, 39(3), 312-320.

Sternberg, R. J., & Lubart, T. I. (1999). The Concept of Creativity: Prospects and

Paradigms. In R. J. Sternberg (Ed.), Handbook of Creativity (pp. 3-15). Cambridge, UK and New York, NY: Cambridge University Press.

Stiglitz, J. E. (1999). Knowledge As a Public Good. In I. Kaul, I. Grunberg & M. A.

Stern (Eds.), Global Public Goods: International Cooperation In The 21st Century (pp. 308-

325). New York, NY: Oxford University Press.

Stolarick, K., King, K., & Mellander, C. (2010). Occupational and Industrial

Distribution in Denmark- A Comparison with the United States, Canada and

Sweden. In M. P. Institute (Ed.), Working Paper: Rotman School of Management,

University of Toronto.

Syverson, C. (2011). What Determines Productivity? Journal of Economic Literature, 49(2),

326-365.

Teece, D. J. (2000). Strategies for Managing Knowledge Assets: the Role of Firm

Structure and Industrial Context. Long Range Planning, 33(1), 35-54.

Thompson, W. R., & Thompson, P. R. (1985). From industries to occupations: rethinking local economic development. Economic Development Commentary, 9(3), 12-

18.

Trajtenberg, M. (1990). A penny for your quotes: patent citations and the value of innovations. RAND Journal of Economics, 21(1), 172-187.

Uzawa, H. (1965). Optimum Technical Change in An Aggregative Model of Economic

Growth. International Economic Review, 6(1), 18-31.

Van den Bosch, F. A. J., Van Wijk, R., & Volberda, H. W. (2003). Absorptive Capacity:

Antecedents, Models and Outcomes. In Report Series in Management: Erasmus University

Rotterdam.

Varga, A. (2000). Local Academic Knowledge Spillovers and the Concentration of

Economic Activity. Journal of Regional Science, 40(2), 289-309.

45

Jönköping International Business School

Vernon, R. (1966). International Investment and International Trade in the Product

Cycle. Quarterly Journal of Economics, 80(2), 190-207.

Verspagen, B., & Schoenmakers, W. (2000). The Spatial Dimension of Knowledge Spillovers in

Europe: Evidence from Firm Patenting Data. Paper prepared for the AEA Conference in

Intellectual Property Econometrics 19-20 April 2000. Alicante, Spain.

Wicksell, K. (1898). Geldzins und Güterpreise: Eine Studie über die den Tauschwert des Geldes

bestimmenden Ursachen (in German). Jena: Fischer.

Williamson, O. E. (1985). The Economic Institutions of Capitalism. New York, NY: Free

Press. von Hippel, E. (1994). "Sticky Information" and the Locus of Problem Solving:

Implications for Innovation. Management Science, 40(4), 429-439.

Young, A. (1995). Growth Without Scale Effects. NBER Working Paper No.5211. National

Bureau of Economic Research Cambridge, MA.

Zucker, L. G., Darby, M. R., & Armstrong, J. (1998a). Geographically Localized

Knowledge: Spillovers or Markets? Economic Inquiry, 36(1), 65-86.

Zucker, L. G., Darby, M. R., & Brewer, M. B. (1998b). Intellectual Human Capital and the Birth of U.S. Biotechnology Enterprises. American Economic Review, 88(1), 290-

306.

46

Collection of Papers

Paper 1

Where do university graduates go?

Lina Bjerke

Paper 2

Export dynamics and product relocation

Lina Bjerke

Paper 3

Imports, knowledge flows and renewal of regional exports

Martin Andersson, Lina Bjerke & Charlie Karlsson

Paper 4

Knowledge intensive business services, creative labour inflow and firm productivity

Lina Bjerke

47

Paper 1

Where do university graduates go?

Lina Bjerke

.

1 …… .……….

49

Where do university graduates go?

Lina Bjerke a b

ABSTRACT

This paper examines three questions: Where do university graduates reside after

graduation? Are the residence decisions of graduate movers different across study areas? How

do push factors, in this case the characteristics of the graduation region, affect the residence

decision for graduate movers? It uses data for all Swedish university graduates in

2007 and analyses how their individual characteristics and the characteristics of their graduation region influence where they reside in 2008. Demographic variables such as age, gender, creative occupation and self-employment follow prior expectations. The results show that declining regions have difficulties to attract graduates irrespective of the characteristics of the graduation region.

High employment in the graduation region makes it attractive to stay. Higher average house prices in the graduation region increases the probability to move and to reside in a growing region but not in a declining region.

Keywords: university graduate migration, regional push factors, embodied knowledge flows

JEL classification codes: R23, J01, J24, J60 a

A research grant from the Centre of Excellence for Science and Innovation Studies (CESIS) at the Royal

Institute of Technology in Stockholm is gratefully acknowledged.

b

I appreciate PhD Candidate Mikaela Backman for being the co-author on the very first version of this paper.

50

Where do university graduates go?

Lina Bjerke

1. Introduction

This paper answers three questions related to migration of university graduates:

Where do university graduates prefer to reside after graduation? Are the residence decisions of

graduate movers different across study areas? How do push factors, in this case the

characteristics of the graduation region, affect the residence decision for graduate movers? In

2007, slightly more than 60 000 individuals graduated in Sweden, and a bit more than 28 000 of these do not live in their graduation region in 2008. This paper focuses on Swedish graduates in 2007, where they reside in 2008 and what factors push them away from their graduation region. Lee outlines the migration decision in a very precise way:

“No matter how short or how long, how easy or how difficult, every act of migration involves

an origin, a destination, and an intervening set of obstacles.” (Lee, 1966, p. 49)

Sweden is a heterogeneous country in terms of population density, net migration and regional economic growth. Regions with problems related to depopulation, a decreasing share of university graduates and an increasing average age of population most often also have problems with economic decline (OECD, 2006). Regional depopulation is a priority policy problem for the Swedish government, and many rural development policies are directed towards this and towards regions with low or negative economic growth.

Knowledge is one of the predominant factors in economic growth and regional competitiveness and a critical type of knowledge is the one embodied in university graduates (Abramovsky et al., 2007; Anselin et al., 1997; Eckel &

Neary, 2010).

1

If regions cannot produce this knowledge internally, they become highly dependent on university graduates from other regions.

1 One of the earliest and most predominant contributors on the relation between knowledge and economic growth is Lucas (1988).

51

Jönköping International Business School

The Swedish educational system has gone through major reforms during the past 4 decades with a decentralisation of institutions of higher education, regulations, admissions. The number of study places in research programs were, on the request upon the Swedish government more than doubled by the year

2000. In Government Bill 1975:9, it is expressed that the main purposes of the reform are to increase access to higher education and by doing so increasing the social levelling.

The number of institutions of higher education and the number of universities graduates has, since the mid-1970s, increased substantially. The number of institutions of higher education are now located in all 23 counties in

Sweden. Such a decentralisation could be an important tool to increase regional knowledge supply but only if the regions can retain the locally produced graduates. A region needs to offer enough suitable jobs for university graduates and attractive residential opportunities. In a recent study using US data, Winters

(2011) shows that a noticeably large share of young in-migrants moving with the purpose of studying, stay after graduation and play a prominent role in the local provision of knowledge.

The existing literature offers explanations of the spatial distribution of embodied knowledge and the focus has predominantly been on characteristics of the destination region, i.e., pull factors (Florida, 2002c; Glaeser et al., 2001;

Gottlieb & Joseph, 2006). Theories on agglomeration, localised knowledge and localised knowledge externalities have served as leading explanations of graduate labour migration showing that size and density create attractive labour markets with better matching between demand and supply of graduates

(Almeida & Kogut, 1999; Duranton & Puga, 2004).

2

Today we know substantially less about what factors make people stay in their education region or what factors that push them out. To answer this we subdivide Sweden into three types of functional regions. If the university graduates do not stay in their graduation region, they move and reside in any of three types of regions; (i) Large regions comprising the three largest regions in

Sweden, Stockholm, Göteborg and Malmö. In terms of population, they are the three largest regions and contain the three largest cities, and they are major nodes for commuting, in-migration and trade; (ii) Growth regions, that are characterised by a growing population; (iii) Decline regions, which are those with a decreasing population. The majority of decline regions can also be referred to as rural areas in Sweden.

2

This is derived from early contributions by Marshall (1890/1920) and Jacobs (1961, 1969).

52

Where do university graduates go?

Studies like this has, for instance been performed for the UK (Faggian et al.,

2006, 2007a) and Australia (Corcoran et al., 2010) but, to the author’s knowledge not for Sweden.

Graduates are heterogeneous and accordingly one can classify them into study areas, distinguishing between graduates in pedagogics (teacher training), social science, natural science and medicine.

3

Graduates are movers if they live in another functional region 2008 than the functional region from where they graduated in 2007. The results show that the residence decision after graduation is affected by demographic variables such as age and gender, as well as by whether the individual gets creative work tasks or not and, if he or she is an immigrant or self-employed. There are also variables describing the graduation region, such as degree of regional specialisation, employment share and housing market, which influence the migration decision. Declining regions have the largest difficulties to absorb university graduates no matter of push factors in the graduation region.

2. Concepts and theories

The migration decision problem can be formalised in a simple way as,

⋛ + +

(1)

,where region and

is the expected utility associated with staying in the graduation

is the expected utility associated with moving to another region. Mobility costs are the monetary and social costs associated with moving.

Equation 1 also has a risk component but graduates stand out in the perspective that they have a higher employment search efficiency, which can reduces such transaction costs of migration (Schwartz, 1973, 1976; Sjaastad,

1962).

4

On Swedish data, it has been shown that 43 per cent of university graduates thought that their jobs were only in part suited to their university education, while 14 per cent thought that their jobs were not at all suited for their education (Statistics Sweden, 2009).

5

3

Natural science category comprises natural science, mathematics and computer engineering but also architecture and engineering (“civilingenjör”).

4

Individuals with a higher education tend to be efficient in searching for an employment and they often choose location before they find a partner and raise a family (if they ever will) (Pleiborn & Strömquist, 1997).

Hence, the problem of asymmetric information may be smaller for university graduates than for others.

5 An immediate consequence of this is that they are overqualified for their jobs, having lower incomes than those who have jobs within their educational fields. The report is limited to eleven areas of education. Among

53

Jönköping International Business School

Besides transaction costs there are location-specific attributes that pull individuals into regions and studies on this mainly focus on incomes, labour demand and housing markets (Herzog & Schlottmann, 1986; Tiebout, 1956;

Whisler et al., 2008). However, there are also factors that push graduates away from their graduation region. Mellander et al (2011) show, by a large-scale survey sample, that factors related to local community satisfaction are more important than economic conditions or individual demographic variables for explaining the likelihood that individuals stay in a region.

One can assume that characteristics of the graduation region are, ceteris

paribus, more familiar to the individual than the characteristics of other possible residence regions. Normally, a graduate has spent at least three years in their study region and has had time to learn whether this is an attractive future work and residence location. To some extent, the decision process can be described as a discrete choice with asymmetric information but one characteristic that is rather transparent to all individuals, is regional size.

A large region comprises a number of attributes that are attractive to people with higher education which is why graduates tend to agglomerate in large and dense regions. These regions supply better opportunities to find an employment matching their education, which lowers the search costs

(Acemoglu, 1996; Helsley & Strange, 1990). Denser regions also tend to offer higher wages, i.e., urban premium, for skilled persons, not least since human capital is more efficiently matched, which makes it possible to be more productive than in low-density regions (Ciccone & Hall, 1996; Glaeser, 1998;

Glaeser & Mare, 2001).

The more contemporary path of urban economics argues that there are high-amenity cities which are attractive to people, causing the desire to live in cities to increase for reasons beyond increasing urban wages (Florida, 2002b,

2002c; Glaeser & Gottlieb, 2006; Glaeser et al., 2001; Lloyd & Clark, 2001).

There are interesting relations between individual incomes and amenities.

Regional income variations can partly be explained by variations in the regional supply of amenities. This is sometimes referred to as amenity rationing. Higher incomes represent a compensation for disamenities, whereas individuals can accept smaller incomes if a region is characterised by a larger supply of desired amenities, ceteris paribus. Looking at the demand side of this, there is also a relation between income variations and variations in the demand of amenities. those fields with a low percentage of persons working in their educational field were the humanities, natural science, languages and political science.

54

Where do university graduates go?

If amenities are normal goods and can be categorised into normal, superior or inferior goods, then when average income rises, one would expect that the demand increases for normal and superior bundles of amenities (Graves, 1983;

Graves & Linneman, 1979; Roback, 1982;1988). If the demand for amenities is related to income, it is also reasonable to assume that university graduates have a relatively high demand for amenities and that migration of university graduates is sensitive to the supply of amenities, such as quality of life, architecture, good environment, and supply of culture and shopping possibilities (Florida, 2002a). However, some people move to rural areas as a lifestyle decision, while retaining urban employment (Ferguson et al., 2007;

Partridge et al., 2010).

6

Another perspective on migration and regional attributes is residential expenditures and total costs of living. Regional house prices and rent differentials are important determinants of migration.

7

The structure of the housing market (supply and price) can have discouraging effects on inmigration into regions with a relatively high labour demand (Gabriel et al.,

1992). For young persons in particular, the expensive housing market may cause impediments to residing in the largest regions simply due to lack of capital and insufficient rental housing. Also, for people with mortgages, asymmetric demand shocks (e.g. demand at a specific labour market) between regions may cause major difficulties moving from smaller regions with lower housing prices to more buoyant regions. The migration decision rather becomes a question of access to capital than access to jobs (McCann, 2001).

While the literature on agglomeration and urbanisation generally focuses on variables that are exogenous to individuals, there are factors that are to individuals, such as age and gender. These, together with characteristics such as marriage or co-habitation status, number of children and type of education provide understandings of the social aspects of moving. A well-established strand of literature shows that mobility generally increases with education, that men are less keen to migrate, that women tend to move longer distances and that the probability to move diminishes with age and distance between regions

(Becker, 1964; Costa & Kahn, 2000; Detang-Dessendre & Molho, 2000;

Edlund, 2005; Faggian et al., 2007b; Krieger, 2008; Mincer, 1977; Pleiborn &

Strömquist, 1997; Sjaastad, 1961, 1962). Figure 1 demonstrates the relationship

6

Both housing market and mobility is related to opportunities to commute. Eliasson et al., (2003) show that the likelihood of choosing commuting as an alternative to migrating increases significantly with the access to job opportunities.

7

Related to this is also the taxation system (Day, 1992; Gabriel et al., 1992; Niedomysl, 2004). Tax differentials have been shown to affect domestic human capital distribution and regional tax differentials and housing costs affect university graduates (Haque & Kim, 1994).

55

Jönköping International Business School

between ages and inter regional migration for all individuals in Sweden over a life-cycle and it emphasises the differences in migration patterns over the life cycle

14000

12000

10000

t

2

8000

6000

4000

t

1

2000

t

3

0

0 10 20 30 40 50 60 70 80 90 100+ t

Figure 1 Relationship between number of domestic in-migrants (national migration) and age in 2009 (data provided by Statistics Sweden)

The vertical axis shows the national migration M (t) and the horizontal axis shows the age t of individuals. The peaks at t

2

and t

3

are the median age of labour force entry and retirement, respectively. The migration rate is low in the adolescent years but increases gradually when reaching the age of leaving the family home. Peak t

1

illustrates that households find it easier to move with younger children, before school becomes compulsory (at age 7). In addition, employment-related mobility diminishes over time. Individuals tend to be more mobile earlier in their careers and stabilise over time.

8

In general, if individuals maximise their expected life-time earnings, they migrate earlier in life, when the expected remaining work time is sufficiently long to recoup the migration costs. A typical highly educated individual at time t

2

has made only minor investments in on-the-job training but large investments in formal education. At t

3

, migration is not driven by employment or income but rather by consumption preferences, reduced travel costs and positive amenities. There are social aspects of migration such as younger people seeking denser areas to find city life and elderly seeking slower pace and

8 Topel and Ward (1988) show that a typical worker in the USA changes employers six times during their first ten years on the labour market. Andersson and Thulin (2008) declare similar results for Sweden. The average mobility for individuals aged 16 to 24 is roughly 25 per cent, while for individuals aged 55 to 64 the average mobility is about five per cent.

56

Where do university graduates go?

beautiful locations such as coastal areas are examples of this (King et al., 1998).

The average age at the time of graduation in Sweden is around 27 years, where also the migration rates for university graduates peak.

The peaks in Figure 1 also relate to the fact that a part of the transaction costs of moving is associated with social networks. The importance of family and friend networks varies with age and this affects the purpose of migration.

From a socio-biological/socio-economic perspective one expects that individuals of fertile age have preferences towards residing close to relatives

(Pleiborn & Strömquist, 1997). Unfortunately, we have in this study no information about in which region university graduates grow up, about whether they are married or not or about whether they have children or not.

3. Data

The empirical analysis is based on data provided by Statistics Sweden on all

Swedish university graduates in 2007.

9

In the year 2008, there were 60 794 individuals holding a university degree from the year 2007 and in the year 2008 as many as 28 119 of these resided in another region than their graduation region. The data provide detailed information on characteristics of individuals and location of education, work, and residence.

According to the definition used here, migration has taken place if a person who graduated from an institution of higher education in 2007 lived in another functional region than the graduation region in 2008. Defining migration in this way means that intra-regional movements are disregarded. This is a reasonable restriction since intra-regional change of residence are not necessarily driven by the same set of factors as inter-regional migration (Rossi, 1980).

There are 81 functional regions in Sweden, referred to as local labour market regions (LLM). These are equivalent to US metropolitan areas and are composed of a number of municipalities (urban areas). The functional regions are characterised by a high intensity of intra-regional commuting flows and are delineated based on the intensity of observed commuting flows between municipalities (Nutek, 1998). The approximated average car travel time distance is 20 to 30 minutes within the LLMs and the car travel time between two locations within a functional region rarely exceeds 50 minutes (Johansson et al.,

2002). In a general way, Figure 2 summarises the choice tree of feasible options for university graduates consisting of three stages, 1, 2 and 3.

9

Data is provided by Statistics Sweden. This is not a public data base and has restricted access.

57

Jönköping International Business School

1. Reside in R

domicile

3a. Stay in R

domicile

2a. Study in R

s

s =domicile

3b. Move to R

r

2b. Study in R

s

s domicile

3c. Stay in R

s

3d. Move to R

r

Figure 2 Choice tree with feasible option for individuals

The first stage (1) represents the domicile region and the second stage with two branches (2a and 2b) represents the graduation region. This first stage can be synonymous with the birth region but this is not necessarily true for all individuals, since families often move between regions before their children start their university studies.

10

In Faggian, et al., (2007b) there is a thorough discussion on how to understand the different types of migration. Repeat

migrants are people leaving domicile region to study and thereafter find a job in a region other than the domicile region and the graduation region. Return

migrants are those studying in a region different from domicile region but move back to the domicile region after graduation. Late migrants are people who attend a higher education in the domicile region and move to another region to find a job. Thus, in the present paper movers can be repeat migrants, return migrants or late migrants. University stayers are people who either lived in the education region before starting their university education or who moved to study and who stay in that region to find a job, which in this paper is referred to as stayers. This is defined in the third stage in Figure 2 with four branches.

Decisions 3b and 3d represent the individuals who decide to leave their graduation region, i.e. movers (shaded in grey). Decisions 3a and 3c concern those staying in their graduation region, i.e. stayers.

Sweden’s functional regions are different in many respects and they can be grouped according to their population growth in the past years. Figure 3 shows

10 Data in the present analysis hold information on birth region for close to 60 per cent of all graduates. Birth region is the region were they were born but there is a large uncertainty in this information given that they might as well have moved in early adolescent years. Data also hold information on where parents reside, which contains valuable information on probability of moving back to where parents reside. Neither this data cover all individuals. However, studies using data from the 70s and 80s regarding all inter-regional labour mobility stress that the return rate is less than 20 per cent (Holmlund, 1984; Westerlund & Wyzan, 1995).

58

Where do university graduates go?

the relation between population growth between 2000 and 2008 and the population size in 2000 in Sweden’s functional regions.

1,1

1,05

Malmö

Stockholm

Gothenburg

1

0,95

0,9

0,85

7 8 9 10 11 12 13 14 15 population in functional region

2000

(ln)

Figure 3 Relation between population growth and population in functional regions

Functional regions with a quota above 1 have experienced positive population growth during this seven-year period and those below 1 have had a decrease in population. The functional regions are in this paper sorted into three groups: (i)

Large regions, which are the three largest functional regions, Stockholm,

Göteborg and Malmö. These three regions alone comprise 40 per cent of the

Swedish population; (ii) Growth regions, which are the 30 regions with growing populations, and (iii) Decline regions, which accordingly are the 48 regions with a population decrease during this eight-year period.

Figure 4 shows the geographic distribution of the three region groups defined above. With the exception of a few regions, the northern parts of

Sweden largely consist of Decline regions. There is also a cluster of Decline regions in the central parts of southern Sweden, where the population density is generally low. The majority of Growth regions are located around the Large regions.

59

Jönköping International Business School

Luleå

Umeå

Stockholm

Large regions

Growth regions

Decline regions

Location of a higher education

Göteborg

Malmö

Figure 4 Map of the three groups: Large regions, Growth regions and Decline regions

There are only three regions in the northern part of Sweden that are defined as

Growth regions. The Growth region in the North-West part is Åre that have experienced a remarkable population growth that differs from the development in most other regions in the Northern parts of Sweden. The two regions in the

North-Eastern part of Sweden with growing populations are regions comprising two large universities, Umeå University and Luleå University of

Technology. The black circles in Figure 4 show where institutions of higher education are located in Sweden. There is a cluster of higher educations in the

Stockholm region but the majority of these are small, such as the University

College of Arts, Crafts and Design, the Military Academy and Beckmans

College of Design. The large institutions of higher education in Stockholm are

Stockholm University, Södertörn University College, Karolinska Institute of

Medicine, The Royal Institute of Technology, and Stockholm School of

Economcis.

60

Where do university graduates go?

3.1 Exposition of data

We begin with a thorough exposition of the data with the purpose to scrutinise whether there are any noticeable patterns. This is especially important, since the output of university graduates in different regions is determined through policy decisions (such as number of students and number of institutions of higher education).

The university graduates can be categorised into four main study areas: pedagogics (teacher training), social sciences, natural sciences and medicine

(incl. nursing). Table 1 presents the number of movers and stayers for each group and the share of women. Less than 50 per cent in all study areas are movers. The percentage in italics in the columns of movers, presents the share of women in the group of movers. Thus, around 76 per cent of all movers in pedagogics are women.

Table 1 University graduates in Sweden, distribution across study areas, share of movers and share of women

Pedagogics, teacher training

Social sciences

Natural sciences

Medicine, nursing

Other

Total

2007

(share of women)

Share of movers

2008 (share of

female movers)

16049 0.44

(0.75) (0.76)

15373 0.49

(0.63) (0.64)

14362 0.43

(0.33) (0.34)

13649 0.47

(0.83) (0.83)

1361 0.69

(0.45) (0.43)

60794 0.46

(0.62) (0.64)

Table 2 illustrates the income of movers and stayers in the three regional types in 2008.

61

Jönköping International Business School

Table 2 Average income (SEK) for movers in 2008 (stayers’ average income in italics and parentheses)*, **

Study area Income of movers to

Large (Stayers)

Income of movers to

Growth (Stayers)

Income of

movers to

Decline (Stayers)

Decline income

/Large income

Pedagogics, teacher training

276 748

(264 762)

261 516

(238 437)

255 520

(230 627)

0.92

(0.87)

Social sciences

Natural sciences

456 103

(447 877)

462 897

(452 561)

363 761

(317 312)

416 707

(354 728)

334 386

(304 321)

387 129

(329 873)

0.73

(0.68)

0.84

(0.73)

Medicine, nursing

314 998

(332 721)

299 358

(288 064)

284 206

(263 848)

0.90

(0.79)

Other

379 832

(276 748)

374 925

(261 516)

333 553

(255 520)

0.88

(0.92)

* The differences in mean are statistically significant.

** If an individual is unemployed and has an income lower than the minimum social welfare, which is 41 652 SEK, he or she has been assigned 41 652 SEK. However, these individuals are only very few in this data.

We see that the average income is highest in large regions and lowest in decline regions for both movers and stayers, irrespective of study area. The final column in Table 2 shows the relation between the average wage earned in a decline region and the average wage earned in the three large regions. The largest percentage difference between a Large and a Decline location for graduates (movers and stayers) can be found in natural and- social sciences.

Graduates residing in a decline location earn around 70 to 80 per cent of what graduates earn in a large region. The corresponding percentage for graduates in pedagogics and medicine is around 90, which shows the more uniform distribution of employment opportunities for this group of graduates as well as the larger urban income premium for graduates from social- and natural sciences.

The decision where to study is not fully endogenous for a student. Even though, there has been a decentralisation of higher educations in Sweden, not all functional regions offer all types of higher education and some functional regions offer no higher education at all. Close to 50 per cent of all graduates have studied in a Large region, approximately 35 per cent in a Growth region and

15 per cent in a Decline region. Table 3 shows the distribution between type of residence regions for graduate stayers and graduate movers.

62

Where do university graduates go?

Table 3 Distribution of graduate stayers, graduate movers and the shares of commuters

in

(share of commuters)

122

(0.44)

083

(0.44)

Live in Growth

(share of commuters)

9 423

(0.28)

10 430

(0.37)

Live in Decline

(share of commuters)

3 130

(0.36)

6 606

(0.30)

The vast majority of all graduates that graduate from Large regions decide to stay in such a region, while the smallest share of stayers is for graduates in

Decline regions. Both cases illustrate the preferences for highly educated individuals to locate in dense regions with diversity in employment opportunities, higher urban wage premiums and more urban amenities. Table 3 also shows the share of commuters, defined as those living in one municipality while working in another municipality. The first row presents the distribution and commuting for stayers and the Large regions have the highest share of commuters. The access to job opportunities is large in these regions and people can choose between many municipalities both in terms of work and residence.

The lowest share of commuters appears for those graduating and residing in a

Growth region. For movers, Large regions also have the largest shares of commuters. The lowest share of graduate movers that also commutes appears for those moving to Decline regions. This may reflect the fewer job opportunities in these regions. The decision to move to a decline region is not a random decision. The option to commute may not exist and once decided to move there, one already has found a job and a residence location.

The geographical distribution of jobs differs between study areas, and one can assume that this is not fully endogenous for individuals. Figure 5 shows the distribution of graduate movers across residence regions and graduate regions for graduates in pedagogics, social sciences, natural sciences and medicine, respectively.

63

1

0,9

0,8

0,7

0,6

0,5

0,4

0,3

0,2

0,1

0

Jönköping International Business School

Residence region

Education region

Decline

Growth

Large

Figure 5 Graduate movers’ choice of residence (left-hand columns) and graduate region

(right-hand columns)

The left-hand side columns in Figure 5 show the distribution of study areas for university graduates across residence regions and reflect the regional labour demand for each study area. Graduates in pedagogics and medicine are largely employed in the public sector and are demanded in all types of regions.

Graduates in social sciences and natural sciences reside in Large- and Growth regions.

11

The set of right-hand side columns in Figure 5 illustrates the restrictions in freedom of choice in terms of study region and are apparent for medicine in particular. There is a limited number of study regions specifically for medicine and very few has graduated from regions categorised as Decline regions. Only eight regions in Sweden have university hospitals and only these regions have all the rights to provide full medical education and the majority of all medical research is conducted in these regions. While legislation restricts the number of graduation regions in medicine, these graduates have opportunities to work and reside in all regions. The vast majority of graduates in medicine is employed in the public sector. Health care in Sweden is the responsibility of the County

Councils who in principle form are regional monopolies in the health care labour market, but the Swedish labour market for graduates in medicine faces a long-term excess demand for labour, which means that a possible monopsony power by employers is balanced by the low labour supply.

11

Table A1 in the Appendix is the contingency Table and Pearson Chi-square for Figure 5.

64

Where do university graduates go?

The other three study areas all have similar distributions in terms of graduation region. This may be an outcome of the two large reforms on decentralisations of higher education in 1977 and 1993 in Sweden.

This section has shown that there are differences between the three largest regions in Sweden (Large), those with a prosperous population development

(Growth) and those with a decreasing population (Decline). Graduates in pedagogics and medicine are similar in many aspects. The demand for both of them are driven by population growth which implies that labour demand is closely associated with policy decisions (size of the public sector) and population size. Graduates in social- and natural sciences differ from the other two, since their labour market is largely determined by market factors, which also is reflected in their incomes. Labour demand for their specific knowledge is determined by national and international demand of products and services produced in their specific industries. What determines labour demand will assumingly affect the residence location of university graduates after graduation.

3.2 Method and analysis

The decision to enter higher education can be thought of as an investment decision by which a person reduces the risk of unemployment and increases career opportunities and standard of living. The choice to move from the region where a person graduated can be explained as a comparison between the expected utility of staying in the graduation region s and the expected utility of moving to another region r.

12

The decision can be described as a combination of a set of possible residence regions with vectors of positive and negative attributes (Y

1

,…….Y

r

) and a set of possible graduation regions with vectors of positive and negative attributes (X

1

,…….X

s

). Graduates seek to maximise the return from their education and the very consequence is that some individuals stay and some individuals decide to move. We are interested in what factors that makes university graduates stay in their graduation region and what factors push them away from their graduation region, where they normally have spent at least 3 years.

The set of factors attracting university graduates into another region are the

pull factors and are aggregated as factors associated with three particular types of regions: Large, Growth, and Decline regions. Attributes related to destination

12

Lee (1966) introduced the probability of moving as a discrete choice between combinations of alternatives of regions and job opportunities. He raises the importance of distinguishing characteristics of region r and characteristics of region s.

65

Jönköping International Business School

regions can be theoretically valuable but may pose empirical and definitional problems. Individuals have chosen the graduation region a priori to the decision to move or stay after graduation (Nakosteen & Zimmer, 1980). Hence, their preferences of location have partly already been revealed. What is of interest is what regional attributes they give up for an alternative location rather than what they can potentially achieve given the current location. The push factors characterising the study region are defined below.

- Labour market specialisation ( ), showing to what extent the

-

-

-

- graduation region is specialised in the type of work tasks performed by the graduate in the new residence region.

13

Employment share ( ), showing the relation between number employed and the regional working population (aged 20-64).

House prices ( ℎ

), measuring the average house price in the central municipality in the functional region.

Service sector ( ), measuring the fraction of employees in the service sector

(NACE 72-74) in region s and is a proxy for urban amenities.

14

Natural amenities ( ), measured as the share of the land area in region s that is recreation areas such as recreational forests, cultural environments and areas distinguished by having environmental qualities.

In addition there are also characteristics of the university graduates themselves that influence inter-regional migration. The decision to move or to stay is expected to primarily be influenced by age ( ), gender ( ), whether or not her or she is an immigrant (not born in Sweden) ( ), whether the individual is self-employed in the residence region ( or she has creative work tasks in the residence region (

), and whether he

) (Florida, 2002c).

13

Specialisation (location quotient):

= Σ

,

Σ

,

, where,

Empl is the number of individuals employed in the same occupation. S is the graduation region.

> 1: The share of employees with this occupation as a share of total employment in region s is larger than this group’s share of Sweden’s total employment, i.e. the group is over-represented in region s.

< 1: The share of employees in this occupation as a share of total employment in region s is smaller than this group’s share of

Sweden’s total employment, i.e. the group is under-represented in region s. Occupations are classified according to Swedish occupational classification, SSYK (3-digit) which is an adaptation of the International

Standard Classification of Occupations, ISCO.

14

The variables measuring the service sector specialisation and the house prices are highly correlated with regional size which means that the latter is not included into the analysis.

66

Where do university graduates go?

The decision where to reside is modelled as a multinomial choice where graduate person i freely chooses between staying in the graduation region (r

stay

) or settling down in one of the regional types described above (r

Large

, r

Growth

or r

Decline

). The probability to choose any of the four possible residential outcomes is expressed in Equation 2 with a vector of individual characteristics of graduate i and a vector with characteristics of graduation region s.

= + ℎ + ℎ

(2)

with i=1,2………,60 794 and r = Stay, Large, Growth, Decline

,where the estimated probabilities are presented in relation to the base outcome

Everything in a multinomial logistic regression model (MNL) is stated relative to the base category. Stay is expressed as the base outcome in the analysis, and consequently one can formulate the estimated probability of a graduate to move to, for instance a Growth region as in Equation 3 (Greene,

2003).

=

(3)

4. Regression results

This section presents the results of the regression analysis of the decision of university graduates of either to Stay in their graduation region or to move and reside in either a Large region, a Growth region or a Decline region.

15

Table 4 presents the risk response ratios (or relative risk ratios), rrr for which staying in the graduation region (Stay) is the base outcome. Relative risk ratios are obtained after running the full multinomial logit model and can be explained as the relative odds (Greene, 2003). The standard interpretation of rrr is: for a unit change in the predictor variable, the rrr of outcome Large relative to the reference group Stay is expected to change by a factor of the respective parameter estimate, given that the remaining variables in the full model are held constant (Greene, 2003). An rrr above 1 indicates that there is an increases (i.e. positive affect) in the chances of residing in that region type compared to the

15 Multinomial logit model is a convenient way to study choice probabilities without any requirements of multivariate integration (Hausman & McFadden, 1984; McFadden, 1973).

67

Jönköping International Business School

base outcome. As an example, in Table 4 we see that the rrr for age is 0.952 which implies that being one year older implies a 5 per cent decreased risk of moving to a Large region and a 0.4 per cent increased risk of moving to a

Growth region. The z-values are in parenthesis and italics below the relative risk ratios for each variable and decision outcome.

Table 4 Multinomial regression results with base outcome (Stay) and control variables, (zvalues in parenthesis and italics)

Individual characteristics

0.952

**

(-25.26)

1.004

**

(2.65)

1.02

**

(11.25)

1.051

*

(1.98)

0.940

*

(-2.46)

0.846

**

(-5.27)

1.080

*

0.743

**

0.596

**

(-9.24)

1.204

*

1.011 0.937

(2.10) (0.13) (-0.57)

1.415

**

0.783

**

0.739

**

(13.39) (-9.25) (-9.01)

0.623

**

0.900

**

0.967

(-14.87) (-3.83) (-1.02)

1.718

**

0.898

**

0.860

**

(-20.22) (-3.68) (-4.01)

0.808

**

0.912

**

0.835

**

(-6.96) (-2.88) (-4.26)

0.833

**

1.123

**

1.180

**

(-5.51) (4.07) (4.87)

Characteristics of graduation region s

(ln)

(ln)

0.506

**

0.807

**

0.728

**

(-17.32) (-5.29) (-6.45)

1.001

(0.13)

1.052

**

(6.32)

0.899

**

(-17.25)

(ln)

(ln)

0.318

**

1.094 0.232

**

(-11.08) (0.83) (-12.87)

0.080

**

0.073

**

0.128

**

(-15.40) (-15.47) (-11.28)

11.834

**

0.567

**

4.300

**

(ln)

Cragg-Uhler R2 = 0.227

(18.00) (-3.96) (8.22)

18.560

**

0.567

**

29.57

**

(8.06)

AIC 130724

BIC -13424

N=60 794

* ** significant at the 0.05 and 0.01 level respectively

All push factors are fractions and log transformed, either by log-transformation or by arcsin- transformation.

‡ The dummy variables are included in the model one at a time. The results presented in the model are for pedagogics but all variables are robust irrespective of using pedagogics, natural sciences, social sciences or medicine.

68

Where do university graduates go?

The usual focus is on coefficients and while rrr cannot be interpreted in an ordinary way, Table 5 further scrutinises the results by presenting the marginal effects for the four possible outcomes (Stay, Large, Growth and Decline). The coefficients in Table 5 are the derivatives for the same control variables as in

Table 4. Since they are derivatives, they describe the effect of the control variable on the margin for each possible outcome, respectively (Greene, 2003).

As an example, the marginal effects for the variable age is positive for Stay,

Large, and Decline but negative for Growth. This implies that the older the graduate is, the lower is the probability to move to a large region.

Table 5 Multinomial regression results- marginal effects, (z-values in parenthesis and italics)

= = = ℎ =

Individual characteristics

0.003

**

-0.007

**

0.002

**

0.002

**

(9.71) (-28.96) (7.84) (16.55)

0.011

*

0.011

**

-0.008

*

-0.014

**

(2.26) (3.43) (-2.23) (-5.50)

0.045

**

0.026

*

-0.037

**

-0.035

**

(6.36) (2.13) (-7.25) (-10.36)

-0.016 0.134

**

-0.003 -0.08

(-0.98) (68.44) (-0.26) (-0.96)

0.008 0.060

**

-0.041

**

-0.026

**

(1.60)

(17.23) (-11.52) (-10.20)

0.053

**

-0.056

**

-0.002 0.005

*

(10.08) (-16.00) (-0.63) (2.04)

-0.036

**

0.085

**

-0.030

**

-0.019

**

(-6.74) (21.35) (-7.83) (-6.91)

0.038

**

-0.023

**

-0.005 -0.010

**

(6.47) (-6.22) (-1.07) (-3.17)

-0.006 -0.029

**

0.02

**

0.015

**

(-1.03) (-7.58) (4.76) (5.15)

Characteristics of graduation region s

(ln)

0.100

**

-0.080

**

-0.007 -0.013

**

(13.30) (-16.22) (-1.20) (-3.31)

(ln)

0.001 4.862e

-4

0.008

**

-0.029

**

(1.05) (0.64) (8.89) (-74.37)

(ln)

0.170

**

-0.134

**

0.072

**

-0.010

**

(8.47) (-10.61) (4.79) (-12.25)

(ln)

0.605

**

-0.230

**

-0.284

**

-0.092

**

(19.19)

(-11.49)

(-12.12)

(-6.46)

(ln)

-0.239

**

0.325

**

-0.183

**

0.096

**

(-9.30) (18.77)

* ** significant at the 0.05 and 0.01 level respectively

(-8.92) (6.69)

69

Jönköping International Business School

4.1 Individual characteristics

The top parts of Table 4 and Table 5 present the influence of the characteristics of the university graduates. For the outcome Large, all individual characteristics are significant. A degree in social sciences has the largest positive effect. The relative risk ratio (rrr) of moving from the graduation region to a Large region is

1.718 which means that the risk of this outcome is nearly 72 per cent. The variable of pedagogics has the largest negative effect on the probability of moving from the graduation region. There is a reduced risk of 37.7 per cent of moving to a Large region.

For the outcome Growth, all individual characteristics are significant except being self-employed. A graduate from medicine an increased risk of around 12 per cent of the probability of moving to a Growth region. Being an immigrant reduces the risk of moving to a Growth region with 25.7 per cent. For the outcome Decline region, all variables are significant except being self-employed and the majority of the variables indicate a reduced risk of moving to this type of regions. The two exceptions are the variables representing age and graduating in medicine with increased relative risk ratios of 2 and 18 per cent, respectively. Now, we turn to a more detailed discussion of the four migration outcomes and the effects of the characteristics of the individual graduates.

Age and gender

Referring to Figure 1, the outcome of variable age is in line with expectations.

It has a rrr of 0.952 for the probability of moving to Large, indicating that higher age reduces the chances of moving to Stockholm, Göteborg or Malmö.

Older persons are reluctant to move to larger regions and this is well established in literature on the social aspects of migration (King et al., 1998).

16

Younger people seek larger (denser) regions to find city life, while elderly people tend to seek a slower pace and beautiful locations, such as coastal areas.

Graduates are inclined to stay in their graduation region for a while, particularly when this is a large region, but are thereafter motivated to move to regions with other attributes when settling down or at least when the children starts compulsory school. This is also supported by the marginal effects for all three residence alternatives, where age is negative in Large but has a positive value in

Growth and also a positive value for Decline regions.

The second interesting result is the one related to gender, showing that graduate men prefer moving to Large regions but not to Growth regions or

16

Inserting the squared age variable generate a change in direction of effect around the age of 42.

70

Where do university graduates go?

Decline regions. Previous studies examine the reasons for co-locations of men and women. Men tend to be more reluctant to move and women tend to outnumber men in large cities. (Costa & Kahn, 2000; Edlund, 2005; Mincer,

1977). The present analysis focuses on people with a higher education only and on the migration decisions of people who are repeat migrants or university stayers. In many cases that these graduates (men and women) have already made the most important decision: to move away from the domicile region. Thus the results only show gender differences in decisions after graduation and not a the overall mobility for men and women over a life time.

Creative, immigrant and self-employed

Having a creative job attract individuals to a Large region. The risk is increased by 41.5 per cent. The risks are reduced substantially for; Growth regions (22 per cent), and Decline regions (26 per cent). The marginal effects in Table 5 for individuals with creative work tasks are not significant for the outcome Stay but for the other three outcomes and the direction of effect changes already between Large and Growth. This most likely reflects the labour market for these kind jobs.

The results show that being an immigrant increases the risk of moving to a

Large region by 8 per cent. Immigrants, even though graduated in Sweden, might have higher barriers to entry outside the large regions. The risk of moving to a Growth region is reduced by 26 per cent and it is reduced by 40 per cent for Decline regions. The marginal effects presented in Table 5 confirm this and the sign shifts from positive to negative between Large regions and Growth regions.

For self-employed, the risk to move to a Large region is increased by 20.4 per cent. With a rrr of 1.011 (i.e., 1.1 per cent), they also have an increased risk of moving to Growth regions. On the contrary, the risk of moving to a Decline region is reduced by 6.3 (1-0.0937) per cent. It seems difficult to find prosperous business opportunities university graduates seeking selfemployments in Decline regions, which could be a result of too large distances to markets and thus a too low market potential.

Study area

The four variables related to study areas show the risks of moving to Large,

Growth and Decline for graduates in pedagogics, natural sciences, social sciences and medicine, respectively. As expected, graduates in medicine have a decreased risk of moving to Large regions but an increased risk of moving to the other

71

Jönköping International Business School

regional types. In terms of labour market and geographical location, they have a higher degree of freedom that graduates from other study areas. Graduates in medicine have very little to lose (in terms of income) whatever region they chose to reside in. Their labour market is to a high degree dependent upon the size of the public sector and the population growth in different regions. The results for graduates in pedagogics are somewhat puzzling since their labour markets are somewhat similar to that of graduates in medicine. Nonetheless, they have a reduced risk of moving to any of three types of regions which suggest that they rather stay in their graduation region. Natural sciences follow the migration patterns of pedagogics in the sense that they rather seem to stay in their graduation region than moving.

Having a degree in social sciences increase the probability of moving to a

Large region with slightly more than 70 per cent. This emphasises that these graduates have a labour market which is highly competitive and these jobs tend to agglomerate in large and dense regions. They rather stay in their graduation region than moving to one of the Growth regions or Decline regions. Thus, the relative risk ratios of the former is 0.898 cent and the rrr of the latter is 0.860.

4.2 Characteristics of graduation region

The lower parts of Table 4 and 5 report the relative risk ratios and z-values of the characteristics of the graduation regions. For the outcome Large regions, all variables are significant except the employment share. The strongest effect is the one indicating the share of natural amenities in the graduation region, though all characteristics of graduation regions have a relatively strong effects in the model. For the outcome Growth regions, all variables are significant. By examining the z-values, the variable measuring the share of services in the graduation region has a strong and negative effect on the probability to move to a Growth region. For the probability of moving to Decline regions, all variables characterising the graduation region are significant and generally different than the other two outcomes. The variable natural amenities is the only variable associated with an increased risk of moving to a Decline region.

Location quotient

A high location quotient indicates, in relative terms, a labour market region with many possible job opportunities for university graduates. The results show that a higher specialisation degree in the graduation region is an attractive characteristic, making graduates willing to stay. The risk of moving to a Large region is reduced by almost 50 per cent while it is reduced by 20 per cent for

72

Where do university graduates go?

Growth regions and 27 per cent for Decline regions. If the region already has a labour market that matches the knowledge one has acquired, and has the ability to absorb it, there are limited reasons for a graduate to move. This variables has relatively strong explanatory power. Putting it differently, if a labour market is quickly saturated, the effect would possibly be that university graduates search other residence locations.

Employment share

The second variable describing the local labour market is a proxy for unemployment, showing the share of the working population (age 20 to 64) with an employment. It is a general proxy for the competition at the actual labour market. The striking result is that the three outcomes differ in direction in terms of effect. A higher regional rate of employment in the graduation region increases the risk of moving to a Large only slightly (0.01 per cent). It also increases the risk of moving to a Growth region by 5.2 per cent. However, it reduces the risk of moving to a Decline region by 10 per cent. The marginal effects in Table 5 are only significant for Growth and Decline regions and have relatively large explanatory powers. If we assume that the employment ratio is also associated with strong competition at the labour market, it seems that a graduate prefers being unemployed in a denser region (Large and Growth) rather than in a less dense region (Decline). The main issue for university graduates is not necessarily unemployment per se, but rather the problem of risking being employed at a position where they cannot apply their university education. A regional labour market with a wide range of employment opportunities is therefore attractive. Instead of having a higher risk of being partially unemployed they can run a higher risk of getting a not perfectly-matched employment in a less diversified market.

House prices

The average price of houses in the central municipality of the graduation region

“pushes” away graduates to Growth regions but not to Large regions or Decline regions. High average house prices in the graduation region reduces the risk of moving to a Large region by almost 70 per cent and reduces the risk of moving to a Decline region by almost 80 per cent. On the other hand, high average house prices in the graduation region pushes graduates into Growth regions with an increased risk of 9.4 per cent. This should most likely be related to what information we have on commuting. I Table 3, one could see that for movers, those residing in a Growth region represent considerably larger share of

73

Jönköping International Business School

commuters compared to those residing in a Decline region. To enter the housing market in the larger cities one needs capital. However, an alternative is to reside in another region which is not as large as Stockholm, Göteborg and Malmö but with a sufficiently large local labour market and/or access to other large labour markets.

17

Services

The variable describing the diversity of services acts as a proxy for urban amenities in the graduation region and strongly reduces the risk of moving. A large share of services in the graduation region reduces the risk of moving to a

Large region by 92 per cent, to a Growth region by 93 per cent and a Decline region by 87 per cent. This indicates that a more diversified service sector, which is also highly correlated with regional size, in the graduation region decrease the likelihood of moving to any other region.

Natural amenities

Natural amenities describe the amenities mostly associated with recreation and tourism. A larger share of natural amenities in the graduation region increases the risk of moving to a Large region, reduces the risk of moving to a Growth region and again, increases the risk of moving to a Decline region. The results indicate that natural amenities have a relatively strong explanatory power and the relative risk ratio is specifically strong for the outcome Large. These very polar results can indicate the large differences between the different types of regions and the fact that those residing in Decline regions have different preferences than those residing in larger regions (this could possibly also be related to age of the individual). Natural amenities in larger (urban) regions are somewhat different from those in a Decline region. In a large, dense region, the marginal utility of an extra acre of accessible nature is most likely higher than in a region where natural amenities are in abundance.

4.3 Where do university graduates go and what are the push factors?

Given the above results, we can now answer the questions stated in the beginning of this paper.

17

The analysis has also been performed with accessibility to wage sums and the results follow these arguments.

74

Where do university graduates go?

Where do university graduates prefer to reside after graduation?

The residence decision vary with individual characteristics of graduates, where the typical person deciding to move to Stockholm, Göteborg or Malmö, instead of staying in the graduation region, is a young man with a degree in social sciences. This man is also likely to have a job in the creative sector. If he is an immigrant plan to be self-employed, the risk increases even further to move to one of the largest regions in Sweden. Similarly, the typical person deciding to move to a Growth region, instead of staying in the graduation region, is a bit older and holds a degree in medicine. Passing a couple of more years, this graduate in medicine can also consider moving to a Decline region.

Are the residence decisions of graduate movers different across study areas?

There are differences related to study areas that should be taken into consideration when interpreting the results of the choice probability model: jobs in pedagogics and medicine are more evenly spatially distributed and these study areas are overrepresented by women. Graduates in medicine and pedagogics are mainly hired in the public sector and are demanded in all regions in Sweden. Contrary, to prior expectations, regions with educations in pedagogics seem to be relatively good of absorbing them. There are fewer regions with a higher education in medicine than there are in pedagogics. Thus, graduates in medicine need to a higher extent move away from their graduation region to find a job. Graduates natural and social sciences are largely attracted to larger and growing regions, showing the more market driven structure of their labour markets.

How do push factors, in this case the characteristics of the graduation region, affect the residence decision for graduate movers?

A graduation region characterised by a high employment share and a large supply of natural amenities pushes university graduates away to Large regions.

Contrary, a graduation region with a large specialisation in the relevant jobs for university graduates, a high share of urban amenities and high house prices, push university graduates away from graduation regions but not to Large regions. A Growth region is an attractive option when the graduation region is characterised by a high employment share and high average house prices. The only regional characteristic, pushing graduates away from the graduation region into Decline regions, is natural amenities.

75

Jönköping International Business School

5. Conclusions

The present paper analysed the location choice of Swedish university graduates and how regional characteristics of their graduation region influence their decisions to stay in their graduation region or to move. The educational system, the structure of the institutions of higher education and the structure of the labour market are a priori important to the analysis.

The analysis began by looking at the distribution of stayers and movers across graduation regions and residence regions and the distribution across study areas. From this initial part of the analysis, one can see that graduates in medicine, with a near relation to the public sector, have no particular preferences with respect to the residence region. Graduates in social sciences seem to choose larger regions in favor of smaller regions. This is further reflected in the average incomes, where the largest differences appear in these latter two educational types. A teacher or doctor residing in Decline region earn around 90 per cent of the salary of a teacher or doctor in Stockholm, Göteborg or Malmö. The corresponding figure is around 70 per cent for graduates in social- and natural sciences. Consequently, it seems that the large-region income premium is the largest for graduates in social-and natural sciences. This follows prior expectations in the sense that these are more dependent on market accessibility, a diversified labour market and fluctuations in labour demand.

The second part of the analysis studied the impact of individual- and regional characteristics on the residence decision after graduation.

Demographic variables such as age, gender, creative occupation and selfemployment all follow prior expectations. The results show that the Decline regions have the hardest time to attract people with a higher education. A wide range of employment opportunities are attractive for highly educated individuals. Instead of having a higher risk of being partially unemployed they can run a higher risk of getting a not-perfectly matched employment in a less diversified market (Decline). Another striking result is associated with graduates’ difficulties to enter the housing market and preferences towards residing in

Growth regions but not in Decline regions.

One can conclude with an important policy related question, associated with

Decline regions and their difficulties to retain the locally produced knowledge but also their difficulties to attract graduates from other regions. Is it defensible to decentralise higher education’s into smaller regions when these regions have only a limited ability to absorb the students who graduate in these regions? This

76

Where do university graduates go?

have a bearing on the costs of decentralisation and whether these costs are higher than the benefits? Thus, the analysis leaves much room for further research and primarily in the direction of a deepening analysis of differences between types of regions and their absorptive capacities.

77

Jönköping International Business School

References

Abramovsky, L., Harrison, R., & Simpson, H. (2007). University Research and the

Location of Business R&D. The Economic Journal, 117(519), C114-C141.

Acemoglu, D. (1996). A Microfoundation for Social Incerasing Returns in Human

Capital Accumulation. Quarterly Journal of Economics, 111(3), 779-804.

Almeida, P., & Kogut, B. (1999). Localization of Engineers in Regional Networks.

Management Science, 45(7), 905-917.

Andersson, M., & Thulin, P. (2008). Globalisering, Arbetskraftens Rörlighet och

Produktivitet: Globaliseringsrådet, Underlagsrapport 23.

Anselin, L., Varga, A., & Acz, Z. (1997). Local Geographic Spillovers between

University Research and High Technology Innovations. Journal of Urban Economics,

42(3), 422-448.

Becker, G. S. (1964). Human Capital: A Theoretical and Empirical Analysis with Special

Reference to Education. Chicago, IL: Chicago University Press.

Ciccone, A., & Hall, R. E. (1996). Productivity and the Density of Economic Activity.

American Economic Review, 86(1), 54-70.

Corcoran, J., Faggian, A., & McCann, P. (2010). Human Capital in Remote and Rural

Australia: The Role of Graduate Migration. Growth and Change, 41(2), 192-220.

Costa, D. L., & Kahn, M. E. (2000). Power Couples: Changes in the Locational Choice of the College Educated, 1940-1990. Quarterly Journal of Economics, 115(4), 1287-1315.

Day, K. M. (1992). Interprovincial migration and local public goods. The Canadian

Journal of Economics/Revue canadienne d'economique, 25(1), 123-144.

Detang-Dessendre, C., & Molho, I. (2000). Residence Spells and Migration: A

Comparison for Men and Women. Urban Studies 37(2), 247-260.

Duranton, G., & Puga, D. (2004). Micro-foundations of Urban Agglomeration

Economies. In J. V. Henderson & J.-F. Thisse (Eds.), Handbook of Regional and Urban

Economics. Vol. 4: Cities and Geography (pp. 2063-2117). Amsterdam: Elsevier.

Eckel, C., & Neary, J. P. (2010). Multi-Product Firms and Flexible Manufacturing in the

Global Economy. Review of Economic Studies, 77(1), 188-217.

Edlund, L. (2005). Sex and the City. Scandinavian Journal of Economics, 107(1), 25-44.

Eliasson, K., Lindgren, U., & Westerlund, O. (2003). Geographical Labour Mobility:

Migration or Commuting? Regional Studies, 37(8), 827-837.

Faggian, A., McCann, P., & Sheppard, S. (2006). An analysis of ethnic differences in

UK graduate migration behaviour. Annals of Regional Science, 40(2), 461-471.

Faggian, A., McCann, P., & Sheppard, S. (2007a). Human Capital, Higher Education and Graduate Migration: An Analysis of Scottish and Welsh Students. Urban Studies,

44(13), 2511-2528.

Faggian, A., McCann, P., & Sheppard, S. (2007b). Some evidence that women are more mobile than men: gender differences in U.K. graduate migration behavior. Journal of

Regional Science, 47(3), 517-539.

Ferguson, M., Ali, K., Olfert, M. R., & Partridge, M. (2007). Voting with Their Feet:

Jobs versus Amenities. Growth and Change, 38(1), 77-110.

Florida, R. (2002a). The Economic Geography of Talent. Annals of the Association of

American Geographers, 92(4), 743 - 755.

78

Where do university graduates go?

Florida, R. (2002b). The Economic Geography of Talent. Annals of the Association of

American Geographers, 92(4), 743-755.

Florida, R. (2002c). The Rise of the Creative Class: and how it's transforming work, leisure,

community & everyday fife. New York, NY: Basic Books.

Gabriel, S. A., Shack-Marquez, J., & Wascher, W. L. (1992). Regional house-price dispersion and interregional migration. Journal of Housing Economics, 2(3), 235-256.

Glaeser, E. L. (1998). Are Cities Dying? Journal of Economic Perspectives, 12(2), 139-160.

Glaeser, E. L., & Gottlieb, J. D. (2006). Urban Resurgence and the Consumer City.

Urban Studies, 43(8), 1275-1299.

Glaeser, E. L., Kolko, J., & Saiz, A. (2001). Consumer city. Journal of Economic Geography,

1(1), 27-50.

Glaeser, E. L., & Mare, D. C. (2001). Cities and Skills. Journal of Labor Economics, 19(2),

316-342.

Gottlieb, P. D., & Joseph, G. (2006). College-to-work migration of technology graduates and holders of doctorates within the United States. Journal of Regional

Science, 46(4), 627-659.

Graves, P. E. (1983). Migration with a composite amenity: the role of rents. Journal of

Regional Science, 23(4), 541-546.

Graves, P. E., & Linneman, P. D. (1979). Household migration: Theoretical and empirical results. Journal of Urban Economics, 6(3), 383-404.

Greene, W. H. (2003). Econometric Analysis (5th ed.). Upper Saddle River, NJ: Prentice

Hall.

Haque, N. U., & Kim, S. J. (1994). Human Capital Flight: Impact of Migration on Income and

Growth. IMF Working Papers. Washington DC: International Monetary Fund.

Hausman, J., & McFadden, D. (1984). Specification Tests for the Multinomial Logit

Model. Econometrica, 52(5), 1219-1240.

Helsley, R. W., & Strange, W. C. (1990). Matching and agglomeration economies in a system of cities. Regional Science and Urban Economics, 20(2), 189-212.

Herzog, H. W., & Schlottmann, A. M. (1986). What Can Be Learned from the Recent

Migrants? Growth and Change, 17(1), 37-50.

Holmlund, B. (1984). Labour Mobility: Studies of Labor Turnover and Migration in the

Swedish Labor Market. In I. Utredningsinstitut (Ed.). Stockholm.

Jacobs, J. (1961). The Death and Life of Great American Cities. New York, NY: Random

House.

Jacobs, J. (1969). The Economy of Cities. New York, NY: Random House.

Johansson, B., Klaesson, J., & Olsson, M. (2002). Time distances and labor market integration. Papers in Regional Science, 81(3), 305-327.

King, R., Warnes, A. M., & Williams, A. M. (1998). International retirement migration in Europe. International Journal of Population Geography, 4(2), 91-111.

Krieger, H. (2008). Migration and mobility culture: an analysis of mobility intentions. In

J. Alber, T. Fahey & C. Saraceno (Eds.), Handbook of Quality of Life in the Enlarged

European Union (pp. 355-384). New York, NY: Routledge.

Lee, E. S. (1966). The Theory of Migration. Demography, 3(1), 47-57.

Lloyd, R., & Clark, T. N. (2001). The city as an entertainment machine. In F. K.

Gotham (Ed.), Research in Urban Sociology (Vol. 6: Critical Perspectives on urban

Redevelopment, pp. 357-378). Oxford, UK: JAI/Elsevier Science.

Lucas Jr., R. E. (1988). On the mechanics of economic development. Journal of Monetary

Economics, 22(1), 3-42.

Marshall, A. (1890/1920). Principles of Economics. London, UK: Macmillan.

79

Jönköping International Business School

McCann, P. (2001). Urban and Regional Economics. Oxford, UK: Oxford University Press.

McFadden, D. (1973). Conditional logit analysis of qualitative choice behaviour. In P.

Zarembka (Ed.), Frontiers in Econometrics (pp. 105-142). New York, NY: Academic

Press.

Mellander, C., Florida, R., & Stolarick, K. (2011). Here to Stay—The Effects of

Community Satisfaction on the Decision to Stay. Spatial Economic Analysis, 6(1), 5-

24.

Mincer, J. (1977). Family Migration Decisions. NBER Working Paper No.199. Cambridge,

MA: National Bureau of Economic Research.

Nakosteen, R. A., & Zimmer, M. (1980). Migration and Income: The Question of Self-

Selection. Southern Economic Journal, 46(3), 840-851.

Niedomysl, T. (2004). Tourism and interregional migration in Sweden: an explorative approach. Population, Space and Place, 11(3), 187-204.

Nutek. (1998). Småföretag och regioner i Sverige 1998: med ett tillväxtperspektiv för hela landet [in Swedish]. Report B 1998:10. Stockholm: Närings- och teknikutvecklingsverket: Nutek.

OECD. (2006). Competitive cities in the global economy. In Organisation for Economic

Cooperation and Development (Ed.). Paris.

Partridge, M. D., Ali, K., & Olfert, M. R. (2010). Rural-to-Urban Commuting: Three

Degrees of Integration. Growth and Change, 41(2), 303-335.

Pleiborn, M., & Strömquist, U. (1997). Nio gånger i livet. Förstudie om flyttningar inom

[in Swedish]. Stockholms län. Stockholm: Regionplane- och trafikkontoret.

Rossi, P. H. (1980). Why Families Move. Beverly Hills, CA: Sage Publications.

Schwartz, A. (1973). Interpreting the Effect of Distance on Migration. Journal of Political

Economy, 81(5), 1153-1169.

Schwartz, A. (1976). Migration, Age and Education. Journal of Political Economy, 84(4),

701-720.

Sjaastad, L. A. (1961). Income and migration in the United States. Chicago, IL: University of

Chicago.

Sjaastad, L. A. (1962). The Costs and Returns of Human Migration. Journal of Political

Economy, 70(5, Part 2), 80-93.

Statistics Sweden. (2009). The labour market for the highly educated-jobs suited to education three years after graduation Official statistics of Sweden: Statistics Sweden,

Population and Welfare/Education and jobs.

Tiebout, C. M. (1956). A Pure Theory of Local Expenditures. The Journal of Political

Economy, 64(5), 416-424.

Topel, R. H. T., & Ward, M. P. (1988). Job Mobility and the Careers of Young Men:

NBER Working Paper No.2649.

Westerlund, O., & Wyzan, M. L. (1995). Household Migration and the Local Public

Sector: Evidence from Sweden, 1981–1984. Regional Studies, 29(2), 145-157.

Whisler, R. L., Waldorf, B. S., Mulligan, G. F., & Plane, D. A. (2008). Quality of Life and the Migration of the College-Educated: A Life-Course Approach. Growth and

Change, 39(1), 58-94.

Winters, J. V. (2011). Human Capital and Population Growth in Nonmetropolitan U.S.

Counties: The Importance of College Student Migration. Economic Development

Quarterly, 25(4), 353-365.

80

Where do university graduates go?

Appendix

Table A1 Contingency table of graduate movers: type of graduate region and type of residence region

Graduate region Large

Growth Decline

Large

Growth

Decline

1 902

6 545

2 636

3 655

5 240

1 535

Pearson Chi-Square 1 303 (sig. 0.000)

1 051

4 120

1 435

81

Paper 2

Export dynamics and product relocation

Lina Bjerke

.

2 …… .……….

83

Export dynamics and product relocation

Lina Bjerke

ABSTRACT

This paper examines whether there are specific regional attributes that can explain why some regions are more efficient in renewing their export base with export products from other regions in Sweden? It identifies the export products that has been completely relocated between regions in Sweden during the period 1997 to 2003 for 38 industrial sectors. This relocation is explained by variables describing the regional characteristics, variables describing the regional export support system and the size of regional demand in the entry region. The findings indicate that while within-sector knowledge is important, the surrounding knowledge affects the absorptive capacity in a negative way. Also, the knowledge effect varies when controlling for industry heterogeneity. The export support system has the overall largest effect on attracting relocated exports from other regions.

Keywords: complete export product relocation, regional attractiveness, manufacturing sector differences, product life cycles

JEL classification codes: R11, R30, O33

84

Export dynamics and product relocation

Lina Bjerke

1. Introduction

A region’s composition of export products is dynamic and over time old export products are replaced by new export products or new export product varieties.

This paper examines regions that are good at attracting new export products from other regions within a country. In fact, the new wave of globalisation with reduced transportation costs and lowered barriers for international trade has had the consequence that exports from some regions’ have been outcompeted by new products and new actors. Other regions have been able to renew their exports and maintain or even improve their export position. Such renewal can partly be explained by knowledge inflows and absorptive capacity. One type of knowledge inflows can come via international trade. However, while international trade can be one source to attain new knowledge, inter-regional knowledge flows are possibly even more important. The common language, the common institutional and juridical system and logistic networks are based on social settings that are prominent features in knowledge diffusion (Boschma,

2004; Breschi & Lissoni, 2001; Polanyi, 1966).

Previous research related to product relocation demonstrates regional differences in the ability to absorb new export products (c.f. Maurseth &

Verspagen, 1999). Following Johansson and Karlsson (1991) there are three main factors describing the optimal conditions for innovation and imitation activities; (i) relevant competence for development work; (ii) information about customer preferences and willingness to pay for various product characteristics and (iii) information about new technical solutions.

This line of thinking has it origin in the product life cycle theory (Norton &

Rees, 1979; Vernon, 1966).

1

The optimal geographic location of production is

1 Theories on product life cycles are extensions of spatial theories on sector dynamics as they were pioneered by Marshall (1920) Kuznets (1960), Burns (1934) and Schumpeter (1939). Processes of agglomeration and decentralisation of economic activities within nations has inspired a development of lead-lag models that have been applied to Swedish and Norwegian data (Forslund & Johansson, 1995).

85

Jönköping International Business School

assumed to vary over time which means that some regions can specialise in knowledge intensive activities such as product and process innovations, while others specialise in later stages of the product life cycle. The early stage of product development is knowledge-intensive, but the knowledge-intensity is sharply reduced at the stage of standardisation and maturation.

2

Building on this and following Hayter (1997), there are strong arguments to believe the following: when a product disappears from a region’s basket of export products and enters into the scope of another region’s exports, there are economic forces pushing these products out from one region and pulling them into another region. These two opposed factors can be associated with each other but can also be looked upon as independent of each other.

Thus, the present paper examines the number of completely relocated export between regions in Sweden between 1997 and 2003 and what are the pull factors in the entry region. Also, are there noticeable differences across sectors in the manufacturing industry? An export product is defined as a combination of a firm identity code and an 8-digit product identity code. A complete relocation implies that the export product completely leaves the export base in an exit region (region of first entry) and enters the export base of an entry region.

3

In fact, one should distinguish between partial relocation, where old and new units are linked together, and complete relocation with physical movement, i.e. discontinuity in space (see e.g Abernathy & Utterback, 1978;

Brouwer et al., 2004; Dumais et al., 2002; Schmenner, 1980; van Dijk &

Pellenbarg, 2000).

This paper diverges from previous literature in the respect that it studies complete relocation of exports on the detailed level of sectors in the manufacturing industry. The results of the regression analysis indicate that the regional within-sector knowledge is a significant factor for a high absorption of new export products. The impact of within-sector knowledge is even further strengthened since the results also show that non-sector related knowledge has a negative effect on absorptive capacity.

2. Regional export base renewal

To stay competitive, there are some general strategies to secure export renewal:

(i) improve quality of existing export products, (ii) introduce completely new

2

Johansson and Karlsson (2003) give an explicit description of the diffusion process between metropolitan

(knowledge intensive) regions and non-metropolitan regions.

3

The product can therefore have been relocated earlier but we are unable to trace that.

86

Export dynamics and product relocation

products and/or (iii) attract export products from other locations (Johansson,

1993). Absorbing new products from other locations is a type of knowledge diffusion, and the magnitude of this is dependent on the rate and speed of product development but also on the absorptive capacity of the entry region.

Thus, export renewal is associated with the local economic milieu, where externalities are generated from regional R&D investments, product and process development and innovations. Relocation can therefore be processes of agglomeration and decentralisation of economic activities.

2.1 Location, relocation and entry barriers of trade

Much of product development and innovative activities take place in large and dense regions with a high knowledge intensity. In smaller countries, such as

Sweden, smaller (low-density) regions are to a substantial degree linked to the larger regions through various networks, which support export activities. The national infrastructure and transport networks are often organised with large and diversified regions as the central hubs. These regions are characterised by agglomeration of knowledge, R&D and innovation, but they also often comprise the necessary institutional and physical infrastructure. The advantages of agglomeration and urban specialisation can be traced back to Marshall’s

(1890/1920) type of localisation economies, Ohlin’s (1933) specialisation of urban regions in production of traded goods and services and Jacobs’s (1969) urbanisation economies.

Diversified regions are characterised by internal localisation economies and can act as nursery cities for new products (Duranton & Puga, 2001; Markusen,

1985). Externalities generated from agglomeration of labour, capital, knowledge and trade (export and import) play a central role for many business networks.

Large companies normally have their head offices and major R&D facilities in these regions, while their production and distribution facilities are located in smaller regions, at home or abroad. Thus, when best practice production has been established, being located in a diverse environment becomes less important and often too costly. Products relocate to specialised regions where large-scale production is more suitable (Brouwer et al., 2004; Brouwer et al.,

1999; Erickson, 1976; Ewers & Wettman, 1980; Martin et al., 1979; Oakey et al., 1980).

If relocation is a result of product maturation ceteris paribus, it should be reflected in prices. In accordance with the literature on product differentiation, products within the same sector with similar unit values are categorised as horizontally differentiated whereas products with different unit values are

87

Jönköping International Business School

vertically differentiated. Horizontally differentiated products are of similar quality and price (Krugman, 1980), whereas vertically differentiated products differ in quality and price (Abd-el-Rahman, 1991; Aiginger, 1997; Falvey &

Kierzkowski, 1987; Flam & Helpman, 1987).

An often emphasised perspective on export dynamics is the costs of entering a market and opening up for trade (Bernard & Bradford, 1999;

Bernard et al., 2003; Bernard et al., 2007; Clerides et al., 1998; Das et al., 2007).

Initiating trade comprises a risk component but while there are entry costs opening up for trade, even very small firms open trade links with small volumes. There is a transitory period of searching and learning. When this is over, the trade link can remain and even grow or it can be closed down and then entry costs become sunk costs. This is relevant to relocation of exports since exporting from the entry region can entail one out of two options. First, exports from the new location can rely on trade links already established in the exit region. This can be the case when it is an export product relocation within the same firm. It can also be a relocation between two locations that are characterised by strong trading networks and spillover effects. The second option is to establish a new trade link at the new location, gathering new information about global demand, language, culture and trading regulations in the country of the trading partner. Thus this creates a new sunk cost. The literature shows that most firms serve only one export destination and the number of firms with multiple export destination markets declines rapidly with the number of destinations (Eaton et al., 2005). If there are sunk costs related to trade, one can assume that the regional absorptive capacity is predominantly determined by accessible knowledge and the region’s export support systems.

2.2 Multi-product firms and export base renewal

The majority of firms exports more than one product. Actually, multi-product firms dominate domestic markets but also international trade (Bernard et al.,

2011; Goldberg et al., 2010). With US data, Bernard et al. (2006) demonstrate that more than 60 per cent of multi-product firms alter their product base (fivedigit SIC products) every five years and that this is a relocation of economic activities within firms towards more productive uses of resources. Nearly 60 per cent of the export firms in Sweden exported more than one product in

2003, and 22 per cent of the firms exported more than ten products.

Economies of scope within exports arise with a common input base and appear when costs of joint exports are lower than costs of separate exports such that

(Panzar & Willig, 1981; Teece, 1980),

88

Export dynamics and product relocation

, < , 0 + 0,

(1)

,where

,

is the export costs associated with the product k and product

z. There is economies of scope whenever the joint exports of k and z is cheaper than exporting them separately.

The mere existence of export diversification does not exclude export specialisation. Regional externalities create incentives for firms to agglomerate in space, and firm-specific knowledge is important to create new varieties of potential exports. The degree of firm specialisation can be illustrated by the number of export products. In such a case, specialisation occurs at a regional level while diversification is firm-specific (Andersson and Johansson, 1984).

4

In a standard transformation function with a vector of export products y and a vector of production inputs x, the efficient combination for a multiproduct exporting firm using numerous inputs can formally be expressed as,

, = 0

(2)

If p is the vector of net unit transport costs of exports and w is the vector of input prices one can express a generalised profit function as,

, ≡

,

s.t

, = 0 (3)

In a competitive setting, a profit-maximising exporting firm has few incentives to fix output or input at certain levels in the long run. Consequently, if innovation or production and transportation possibilities change over time, firms have incentives to adjust their exports.

5

If export prices are fixed ceteris

paribus, such adjustments of export conditions can imply that the optimal location for a firm also changes since input-prices can be closely associated with regions. Thus, agglomeration and localised knowledge spillovers imply that one region can be optimal for several firms within the same sector and the optimal location differs between sectors rather than between specific export products.

Some exports are easier to relocate than others due to fixed costs and sunk costs of location. Changing export location will only occur when the entry region brings at least as much profit as the exit region when also these costs are

4

Appendix A1 presents the cost function and economies of scope building on Andersson and Johansson

(1984).

5

This substitution effect was first introduced by Moses (1958). It is a decision to locate at the highest tangency condition between site-specific budget constraints i.e. envelope budget constraints, and the isoquant of that particular firm (McCann, 2001).

89

Jönköping International Business School

taken into consideration. As a consequence, products within sectors with high fixed costs are most likely less mobile than those otherwise. Due to sector differences, the analysis in the present paper is therefore executed at the sector level where each sector comprises a wide variety of export products and all products are assumed to have similar preferences for regional export support systems.

3. Hypotheses and data

Summarising the above discussion, larger and denser regions generate knowledge externalities and can be looked upon as breeding places for product development and innovation. Knowledge and skills are not only important in product development but also in diffusion and absorption of export products.

Hypothesis H1: To absorb relocated export products, a region needs capacity in terms of

knowledge and skills related to the type of relocated export product.

If a relocated product is less capital-intensive and more standardised in the entry region than in the exit region, the potential increase in export (production) volume reduces the per-unit costs of production. A specialised infrastructural network in the entry region facilitates trade but also enables location in peripheral regions. With such infrastructural network, firms have access to global demand, global supply but also to inter-regional knowledge flows.

6

In addition, there are sunk costs associated with opening trade links and costs of maintaining these trade links and these costs are lower the better the infrastructure networks.

Hypothesis H2: A sophisticated export support system in terms of infrastructure and

networks in a region increases the probability that new export products will enter.

The empirical analysis rests on data provided by Statistics Sweden (SCB).

7

The first set of data concerns exports and covers the seven-year period 1997-2003.

The data are tabulated across products, destination and origin markets, and functional regions in Sweden.

6 National borders are likely to alter distance sensitivity in a discrete manner (Portes & Rey, 2005).

7

Data is provided by Statistics Sweden. This is not a public data base and has restricted access.

90

Export dynamics and product relocation

An export product is defined as a combination of a firm identity code and an 8-digit product identity code. Products are classified according to the 8-digit

CN system, and export values and volumes are provided for each firm and product combination in each region.

8

The data cover the manufacturing industry in Sweden divided into 38 sectors. One can follow information on values (SEK), volumes (KG) and geographic location over time so that spatial relocation of export products can be identified. Relocation is a complete relocation defined as a relocation between an exit region and an entry region.

The product completely leaves the export base in the exit region and it is also new for the entry region. The sample extraction is illustrated in detail in Figure

A1 in Appendix A2.

The geographical unit is functional regions in Sweden and they are 81 in number and are equivalent to local labour markets, LLM (Nutek, 1998). Local labour markets are identified by their high intra-regional flow of commuters and are delineated based on the intensity of observed commuting flows between municipalities. The approximated average travel distance is 20-30 minutes within the LLMs, and the travel time between two locations in a region rarely exceeds 50 minutes (Johansson et al., 2002). Travel intensity declines sharply at the borders of LLMs. In this sense, it can be argued that the LLMs are arenas for face-to-face interaction (Andersson & Karlsson, 2006).

3.1 Methodology

The ambition is to estimate the influence of the entry region attributes on the number of incoming export products into a sector and region from other regions. To structure a symmetric matrix of regions and sectors, keeping data at a disaggregated level there are an excess amount of regions and sectors having no relocated products. There are statistical ways to handle overdispersion and excess amounts of zeros. A zero-inflated negative binomial model with a random term can reflect the unexplained intra-subject differences included in the model (Fritsch & Falck, 2002; Moghimbeigi et al., 2008). Observations with zero values can result from two kinds of regimes. First, the probability of a positive number of relocated export products in region and sector is zero where a zero observation cannot be a result of a stochastic Poisson process. Second, observations with zero values are regarded as outcomes of a stochastic Poisson process with a positive probability of having relocated products in the region

8 CN is the Combined Nomenclature based on the Harmonized Commodity Description and Coding System of the Customs Cooperation Council.

91

Jönköping International Business School

and sector. A zero-inflated negative binomial model excludes observations with zero values that cannot be accounted for as results of a Poisson process. This can be accounted for in a zero-inflated Poisson (ZIP) model (Cameron &

Trivedi, 1986). Let X denote the explanatory variables. The expected number of completely relocated export products is related to X through a zero-inflated

Poisson model with probability as:

, where

(3)

=

(4)

, where is a

∙ ∗ 1

vector and X is a

∙ ∗ + 1 data matrix and is a

+ 1 ∗ 1 vector where is the number of explanatory variables in the model.

Table 1 presents the dependent variable but also describes the explanatory variables in X. There is one set of explanatory variables describing the characteristics of the regional economic milieu. There is another set of variables describing the regional export support system and a final set of variables describing local and regional domestic markets. Due to a skewed distribution, some variables are log transformed in the regression model.

92

Export dynamics and product relocation

Table 1 Variable descriptions: Regional and sector characteristics explaining export relocation*

Variable Description Definition Effect

Dependent variable

Relocation

r,j

Number of incoming completely relocated export products into region r and sector j in 2003

Regional economic milieu

Knowledge intensity

r,j

Surrounding knowledge

r

Manufacturing

r

Regional service

r

Percentage share of workforce in region r and sector j with university education of 3 years or longer in 1997

Percentage share of workforce in region r with university education of 3 years or longer in all other sectors but sector j in 1997

Share of employees in manufacturing industry in region r in 1997 as a fraction of total employment in region r

Share of regional employment

in NACE 65-74 as a fraction of total employment in 1997.

Minimum efficient export scale of region r in sector

j in 1997

Knowledge spillover effects of sector-related knowledge

Knowledge spillover effects from surrounding economic milieu

Specialisation in manufacturing sector

Supply of regional subsidiary services

+

+

+/-

+

MEES

r,j

Industrial scale economies in terms of scope of export products

+

Export support system

Trade links

r,j

Trade infrastructure

r

(ln)

Distance to metropolitan

r

Sum of export and import links, i.e. countries for sector j in region r in 1997

Region r’s accessibility to ports in Sweden and northern Europe in 1997

Travel time between region

r and the Stockholm metropolitan region in 1997

Established trade networks +

Export and import transport possibilities

Distance to highlevel networks and knowledge

+

+

Domestic demand

Regional size

r

(ln)

∆ Regional demand

r

Market access

r

Employment size of region r in

1997

Growth of gross regional product in r between 1997 and 2003 (fixed prices)

Total accessibility to wage sums in region r in 1997

Effects of agglomeration and urbanisation

Growth in potential regional demand as complement to global demand

Potential regional market access

* Table A1 in Appendix A3 presents the bivariate correlations for all variables. Table A2 in

Appendix A3 presents all 38 sectors in the manufacturing industry.

+

+

+

93

Jönköping International Business School

Table 1 presents alternative ways to approximate trade and related infrastructure. Distance (in travel time by car) to the Stockholm metropolitan region is a way to illustrate the role of this region in processes of export renewal in Sweden. The Stockholm region comprises the capital city Stockholm and has a number of favourable characteristics in terms of product development and innovation. Being located far away from this region, which has a large inflow of international and inter-regional imports, can be a disadvantage in terms of export renewal.

An alternative measure of an export support system is to count the number of trade links and a final way is to measure the accessibility to national and international ports. The geographic location of Sweden implies a high dependence on maritime transportation for delivering exports, and expecially for sectors with large export volumes.

9

More than 90 % of Sweden’s trade

(imports and exports) are transported by cargo ships and ferries

(Sjöfartsverket, 2006).

To measure the extent of scale economies in the exporting sectors, the minimum efficient export scale MEES is included. This relies on the common measure of minimum efficient scale (MES), developed and applied by Comanor and Wilson (1967) and further used by Audretsch (1995). In the present paper, the minimum efficient export scale (MEES) measures the average number of export products for the largest exporting firms accounting for the 75 th percentile in the region and the sector. Those average numbers are thereafter divided by the total number of export products in the region and the sector.

Regional knowledge intensity is measured using sector-specific as well as non-sector-specific measures. Regional size is a way to measure agglomeration and urbanisation economies but this does not necessarily correspond to successful regional economic growth. Economic growth signifies increasing customer demand. Figure 1 illustrates the correlation between economic growth and population in Swedish functional regions for the period 1997 to

2003. There is not a clear relation between the two variables and therefore both are included in the model.

9 Accessibility is measured as an exponentially decreasing function of distance. For a presentation of a full model see Johansson et al., (2003).

94

Export dynamics and product relocation

12

Karlshamn

11

10

9

8

15

14

13

Stockholm

Göteborg

Malmö

Strömstad

Jokkmokk

7

6

0 -0,1 0,1 0,2 0,3 0,4 0,5 0,6

Percentage growth of gross regional product 1997-2003 (fixed in 1997 prices)

Figure 1 Correlation between GRP growth (in fixed prices) and population (data source is

Statistics Sweden)

3.2 Descriptives

Table 2 illustrates the descriptive statistics for all variables and presents the number of observations (81 regions and 38 manufacturing sectors), minimum value, maximum value, mean and standard deviations. The full correlation matrix is presented in Table A1 in Appendix A3

Table 2 Descriptive statistics for dependent variable and explanatory variables (all variables as non-log)

*

N

Relocation 3 078 0 155

Knowledge intensity

Surrounding knowledge

Regional service

MEES

3 078

3 078

3 078

3 078

0

0

Manufacturing 3

0.020

0

1

0.182

0.163

1

Trade links

Trade infrastructure

Distance to metropolitan 3 078

Regional size

∆ Local demand

3 078 0

3 078 0.068

3 078

3 078

7.212

1079

-0.089

1.885

0.027

0.036

0.059

0.254

241 31.599

46519.660 1 849.839

767.878

957 961

0.491

296.174

48 650.400

0.222

9.348

0.077

0.031

0.029

0.311

38.735

6 107.003

176.286

117 305.632

0.103

Market access 3 078 1256.258 8 971 217 390 386.923 1 088 604

The model is restricted in the analysis due to multicollinearity between some of the regressors

In Table 2 it is interesting to see that some industries in some regions have no reported trade in the starting year 1997 (sum of import and export links in

95

Jönköping International Business School

region r and sector j). The variable measuring travelling time to the Stockholm region is highly skewed which illustrates the geography of Sweden with long distances. Accessibility to ports in Sweden and northern Europe also illustrates the relatively low accessibility to ports for exporting products that some regions have.

Table 3 is a contingency split sample table. Rows represent export unit values (value in Swedish Crowns (SEK) per volume in kilograms (KG)) of product k in 1997 distributed among quartiles: high, medium high, medium low and low unit value. Columns represent export value growth of k between 1997 and 2003. A majority (2 746) of relocated products show a negative export value growth during the time period 1997 to 2003.

Table 3 Contingency table of number of relocated export products in each category. Rows: export unit value in 1997. Columns: Export value growth between 1997 and 2003

Export value growth between 1997 and 2003**

Unit value 1997* High

(>1.62)

Medium high

(1.61-0.81)

Medium low

(0.80-0)

Neg. Total

Medium high (390.60) 217 229

Medium low (119.68) 232 261

Total 1 038 1 008 1 008 2 748 5 802

Pearson Chi-square 78.42 (sig. 0.000)

Likelihood ratio 75.27 (sig. 0.000)

Linear-by-linear association 30,93 (sig. 0.000)

*Distributed among quartiles, break points in italics.

**

− − 1

+ − 1

2

, distributed among percentiles 33 and 66.

Break points in italics. Negative values in a separate column.

Moving exports from one location to another can imply fixed costs that export continuers do not experience. The majority of products had a medium high

(750) or high (702) unit value before relocation. This brief look supports the product life cycle theory. In the later phases of the product life cycle, products tend to have increased exports and face increasing price competition. To remain active on the export market, firms need to reduce costs and/or renew exports and carry out development of production and distribution processes.

This is often related to the timing of the product relocation, which means that geographic movements of export products to new locations are parallel to a reduced export unit value. Nearly 50 per cent (2 748 out of 5 802) of the relocated exports had a negative unit value growth between 1997 and 2003. The second largest group is the one with the 360 products that had a low unit value in 1997 that also has experienced a high export value growth since relocation.

96

Export dynamics and product relocation

By only looking at Table 3 one cannot distinguish between a value decrease or a volume increase. However, either of these changes reflect an export adjustment. In a more detailed way, Table 4 illustrates the distribution of relocated products into four groups: (i) positive value growth and positive volume growth, (ii) positive value growth and negative volume growth, (iii) positive volume growth and negative value growth, and (iv) negative value growth and negative volume growth.

Table 4 Number of relocated products with positive and negative value growth and positive and negative volume growth.

Positive value growth

Negative value growth

Positive volume growth

2 558 346

Negative volume growth

467 2 358

In Table 4 one can see that there are only few products for which the change in value and volume has been in opposite directions. The majority of products have either had a positive growth for both value and volume or a negative growth for both of them.

Furthermore, Table 5 demonstrates the average export value growth of relocated export products in each sector in the manufacturing industry and the average export value growth of non-relocated exports (continuing exports) in the same sectors.

97

Jönköping International Business School

Table 5 Number of relocated export products into sectors and average export value growth of categories of exports: relocated and continuing, highest growth of the two

categories in italics and negative values in bold

Sector No.

*

Wood products 133

Relocated exports

0.5

Continuing exports

8.4

Petroleum products and nuclear fuel

Basic metals

Pesticides and agro-chemical products

Paints, varnishes

Soaps, detergents, toilet preparations

Other chemicals

Man-made fibres

Rubber and plastic products

Non-metallic products

28

273

11

76

85

119

4

507

150

56

8.4

40.5

10.2

27.0

12.6

90.7

3.6

8.1

32.7

13

29.5

32.6

26.6

7.0

-20.1

13.5

6.9

Fabricated metal products

Energy machinery

Non-specific purpose machinery

Agricultural and forestry machinery

Machine tools

Special purpose machinery

Weapons and ammunition

Domestic appliances

Office machinery and computers

Electric motors, generators and transformers

Electrical distribution, control, wire and cable

Accumulators, batteries

Lightening equipment

Other electrical equipment

Electronic components

541

311

257

21

103

209

2

88

174

108

251

20

63

98

113

Television and radio receivers, audio-visual electronics 85

Medical equipment 58

Measuring instruments

Optical instruments

Watches, clocks

Motor vehicles

Other transport equipment

220

37

15

220

49

21.4

2.4

-10.9

-0.5

82.1

Furniture, consumer goods 349

Σ 5 802

14.8

*Number of completely relocated export products between 1997 and 2003

-35.6

10.7

12.6

-20.9

-2.4

6.8

34.3

0.4

-8

-0.2

67.1

0.8

8.2

-40.7

-14

4.9

7.1

18.3

Negative growth is more common for relocated exports than for exports continuing in the same region. Negative growth in both categories of exports only appears in two sectors, signal transmission and telecommunications and watches

and clocks. It is possible that such growth differences can be the result of discontinuity in space rather than a phenomenon of business cycle fluctuations or changes in customer demand. If average export value growth in this period were more closely related to economic business cycles, one would assume that

14.1

8.0

13.2

18.8

-3.4

9.5

4.8

8.3

1.4

-19.8

22.1

30.2

16.1

13.8

5.6

11.3

30

8.3

9.9

44.8

45.9

-0.4

5.3

16.2

98

Export dynamics and product relocation

products have similar development irrespective of geographic location. Such an alteration in demand would cause reduced exports at all locations.

The most apparent obstacles of relocation are immobile factors of production. Sectors with a large share of immobile factors of production, characterised by constant returns to scale and high land intensity, are less likely to relocate. Examples of these can be petroleum products that are dependent on specific locations determined by the limited supply of suitable locations. A trapped resource can also be knowledge, for instance close collaborations with universities, knowledge intensive business services (KIBS), specific suppliers and customers. As a consequence, knowledge-based sectors may have a lower relocation probability.

Sectors can also be characterised by substantial internal or external scale economies. External economies can be localisation economies and urbanisation economies arising from agglomeration forces. They can also be characterised by very large firms with distinctive first-mover advantages. It can be sectors that are dependent on natural resources such as firms in the petroleum and miningindustry. These have high barriers to entry and are less mobile in space.

10

For some sectors in Table 4, there are particularly large differences between the two groups relocated exports and continuing exports. These are petroleum products and

nuclear fuel, pesticides and agro-chemical products, basic metals, agriculture and forestry

machinery, medical equipment and other transport equipment. These sectors are, to varying extent, associated with trapped resources and therefore also large sunk costs of relocation. Hence, these export products have very few alternative locations. These sectors also have had relatively high unit value growth, which can indicate that they only relocate when they can reap advantages at the new location and they do so in a successful way. Product life cycle theory suggests that, ceteris paribus as a product matures, volume increases and unit value decreases. Table 4 does not clearly support this hypothesis. It appears that negative growth is slightly more common for relocated exports.

Figure 2 illustrates the geographic distribution of the entry of relocated export products.

A darker shade of grey signifies a larger number of absorbed relocated exports.

10

Nelson and Winter (1982) present an ecological perspective in which sector dynamics appear as results of imitation and adoption. Pavitt (1984) suggests disaggregate arrangements with supplier-dominated, production-intensive and science-based sector classifications. Sellenthin and Hommen (2002) present a

Swedish classification system using CIS data and classify according to innovation intensity and R&D expenditures.

99

Jönköping International Business School

Entry

0-5

6-29

30-91

92-390

391-1889

Stockholm

Göteborg

Malmö

Figure 2 Number of relocated products: entering products.

Figure 3 presents the geographic distribution of exit of the same set of products as in Figure 2. A comparison of the two maps high-lights two important observations. Numerous regions and industries in Sweden have a high export product exit and high export product entry. A number of regions have prominent roles in Swedish trade with a rapid product renewal, which is reflected in both maps. Stockholm, Göteborg and Malmö are large imports nodes where new knowledge enter. This knowledge is thereafter diffused to other regions in Sweden. Thus, the second interesting observation is that it appears that the exit of export products is slightly more spatially concentrated its largest concentration to functional regions around the Stockholm region suggesting that the Stockholm region acts as an incubator of new products.

100

Export dynamics and product relocation

Exit

0-10

11-31

32-139

140-571

572-846

Stockholm

Göteborg

Malmö

Figure 3 Number of relocated products: exiting products.

4. Results

Table 6 presents the results from the estimations of the zero-inflated Poisson model with the number of incoming (relocated) export products into a region and a sector as dependent variable. Due to multicollinearity between regressors, they are included step wise. Table 6 shows model variations. Models I -III are estimations without controlling for sector heterogeniety. Models IV -VI are identical to models I-III with the only exception that sector dummy variables are introduced to detect any systematic differences which are attributable to industrial sectors and which cannot be explained by the control variables.

101

Jönköping International Business School

Table 6 Zero-inflated Poisson regression, dependent variable: Number of relocated products ‡ , N=3 078, z-values in parentheses and * p<0.05 ** p<0.01

Regressors I

Knowledge intensity

r

Surrounding knowledge

r

0.52

**

(3.35)

1.86

*

(2.26)

II

0.44

0.71

**

(2.81)

(1.00)

III

0.64

**

(4.16)

4.74

**

(8.28)

IV

2.24

**

-2.55

*

(2.65)

(11.25)

V VI

2.06

**

(10.51)

-0.89

(1.11)

2.21

**

-5.43

**

(8.46)

(11.11)

Manufacturing

r

1.71

**

(3.85)

1.60

**

(4.34)

0.72

**

(2.14)

3.30

**

(7.07)

1.35

**

(3.55)

0.33

(0.98)

Regional service

r

MEES

r j

-0.03

(-0.35)

**

0.32

(3.34)

**

0.69

**

(3.07)

0.58

**

(2.74)

0.18

**

(2.81)

2.23

**

(9.81)

1.49

**

(7.01)

0.01

**

(16.56)

1.39e

-5**

(7.54)

0.74

(3.32)

**

Trade links

r j

Trade infrastructure

r

Distance to metropolitan

Regional size

r

r

(ln)

(ln)

∆ Regional demand

r

Market access

r

0.01

**

(37.10)

8.87e

-6**

(4.85)

-8.14e

(-3.75)

-4**

0.01

**

(36.98)

9.84e

-4**

(5.63)

-9.84e

(-4.74)

-4**

0.01

**

(40.93)

1.36e

-6**

(9.24)

-1.13e

(-5.50)

-3**

0.01

**

(14.10)

7.70e

-6**

(4.08)

-6.47e

(-2.89)

-4** -1.38e

-4**

(-6.25)

0.01

**

(26.35)

1.64e

-6**

(10.53)

-1.32e

(-6.15)

-3**

0.25

**

(6.01)

0.41

(1.33)

0.49

(1.52)

2.68

**

(9.41)

0.63

**

(12.60)

0.79

**

(2.51)

Constant -2.98

(-5.30)

**

CONTROLL FOR

SECTOR

HETEROGENIETY

#

-3.51

(-8.71)

**

-2.58

(-4.83)

**

-6.46

NO NO NO YES

(-10.56)

**

LR chi2 8446

Prob>chi2 0.000

8499

0.000

8413

0.000

9712

0.000

0.66

*

(2.01)

0.46

**

(14.45)

-5.94

(-13.52)

**

-1.33

(-5.68)

**

YES YES

9759 9524

0.000 0.000

‡ The regressors in the zero and the count component are identical except for controlling for sector heterogeneity. Sector dummies are then excluded and the regressors in the zero and count components are only overlapping. In the inflated model, there are 60 per cent zeroes.

Overall regional trade intensity (sum of export value and import value as a share of the total gross regional product) has also been tested for but with non-significant results.

#

The excluded sector is furniture, consumer goods

102

Export dynamics and product relocation

4.1 Regional absorptive capacity

This section discusses the empirical findings related to the regional economic milieu starting by scrutinising the results concerning sector-related and nonsector-related regional knowledge intensity.

Sector specific knowledge

A large share of employed people with the ability to absorb and use new knowledge would presumably enhance the possibility to attract new export products. Sector-related knowledge measures the share of labour with a threeyear or longer university education in region r in sector j. Starting with models I to III, with no control for sector heterogeneity, within-sector knowledge is positive and significant across all model variations.

This suggests that regional sector-specific knowledge has a significant impact on the number of absorbed new export products that are relocated from their region of first entry (exit region). Absorbing new export products is a way to renew the regional export base, and in order to do so, the right skills and competences must be regionally available to adapt the export products to new conditions, and possibly new production techniques.

Here, an export product is defined as a combination of a firm identity code and an 8-digit product code, which implies that the firm has started exports from another location for this specific product. Thus, an economic environment with high knowledge intensity is still a very important feature for the relocation of export products. Controlling for sector heterogeneity gives much higher coefficients and z-values illustrating that the effect varies substantially across industries.

Surrounding knowledge

The variable representing the surrounding knowledge, i.e., outside each specific sector in the region, changes signs from Models I and III to models IV and VI and turns from significantly positive to significantly negative when sector dummies are used. This is reasonable if we assume that these relocated exports products are in a phase of standardisation where there is a relatively low need for other knowledge except from the sector specific knowledge. Likelihood ratios increase significantly when introducing sector differences suggesting higher explanatory power. We also see that the results on surrounding knowledge intensity are not robust across Models IV to VI. Looking at

103

Jönköping International Business School

correlations in Table A1 in Appendix A3, we see that surrounding knowledge intensity and regional employment size has a correlation coefficient of 70 per cent which possibly creates a problem for a number of the variables describing the regional economic milieu.

Hence, the number of relocated products into sector j in region r is negatively affected by high knowledge intensity in all other sectors in this region, when sector dummies are included. When export products are relocated the firm in question seeks a location with appropriate skills and competences given the type of product. The negative effect of non-sector-related knowledge can illustrate the fact that the these products are already standardised and established at the international market. The exporting firm knows the best practice technology and the product is in a later stage of the product cycle. It is not heavily dependent on new knowledge but is more dependent on sectorrelated knowledge that can guarantee an effective and cost-efficient production.

Specialisation in manufacturing

Having regional specialisation in manufacturing is positive for the relocation of export products and this result is robust across all models in Table 5. This is not a surprising result: export products are relocated to regions with a history in manufacturing. Locating in a region with a specialisation in manufacturing facilitates the search for labour with the appropriate skills, competences and working experience but possibly also the search for capital and other industryrelevant inputs. A regional specialisation in services has a positive effect for the relocation of export products but is only significant when controlling for sector heterogeneity.

Industry economy of scope in exports

The final variable describing the regional economic milieu is the minimum efficient export scale of the sector in the entry region in 1997. This measures the average number of export products for the largest exporting firms accounting for the 75 th

percentile in the region and in the actual sector. Those average numbers are then divided by the total number of exports in the sector in the region to achieve a variable reflecting the size of the actual scale economies. Looking at the results in Table 6, this proxy of export scale economies suggests that regions with a high entry rate of new export products have a relatively high concentration of sectors with a high minimum efficient

104

Export dynamics and product relocation

export scale. This supports the idea that firms are mostly multi-product exporters, and it seems likely that the sectors and regions with a relatively larger number of export products are those that are successful in attracting and absorbing new export products.

4.2 Regional export support system

Trade links

The number of open trade links is chosen to capture the established export networks of a region from which new export products can gain advantage. This variable is positive, highly significant and robust throughout the alternative models. The coefficients are relatively small but the z-values are high suggesting that this effect is strong in a comparison with the other explanatory variables.

Some of the potential fixed costs related to establishing new trade links can be avoided by using already established export networks in the region. The expenses related to relocation can thereby be reduced. Similar results appear for regional accessibility to ports. Easy access to trade infrastructure has a positive effect on regional export renewal. The second variable describes one aspect of the export support system and shows the accessibility (based upon car travel time) to the major domestic ports but also to dominant ports in Northern

Europe. It is also positive throughout the models and robust across models, though the coefficients are rather small. Table 4 demonstrated that the majority of the relocated export products were in sectors with voluminous exports such as metal products, machinery and rubber and plastic products. Larger volumes are more expensive to transport ceteris paribus and therefore this is a reasonable result.

Distance to the major metropolitan region in Sweden

The variable distance to the major metropolitan region in Sweden measures the geographic distance (based on car travel time) to the largest (in terms of population) functional region in Sweden. It is negative in all models suggesting that a geographical location farther away from Stockholm has a negative impact on the entry of new export products. This explanatory variable regional size aims at capturing the effects of agglomeration and urbanisation economies and is positive when included in Model I and Model IV. The z-value is high, specifically when controlling for sector heterogeneity.

105

Jönköping International Business School

Accessible demand

Now we turn to the variables describing accessible domestic demand. Starting with general growth of regional demand, measured as the growth of gross regional product between 1997 and 2003. The results show that entry regions with a prosperous economic development are particularly good at renewing their export base with new export products from other regions. One cannot distinguish between cause and effect but the result anyhow supports the idea that local and global demands might be complements to each other.

Market access measures the total access (within and external to region) to wage sum and this variable also has a positive impact on the entry of new export products.

11

Sector heterogeneity

The intriguing results associated with sector heterogeneity require further analysis. To derive information on sector heterogeneity, Table 7 presents the results when a restricted model is estimated with only the dependent variable and sector dummies. The first column lists the sectors with positive and significant effects on the number of relocated export products. They are ordered in size with respect to their z-values, starting with the highest z-value.

The second column shows the sectors with a significant negative coefficient, and the final column presents sectors turning out as insignificant. The column of insignificant sectors comprises agricultural and forestry products, which are spatially tied with a strong path dependency in location. The fact that motor

vehicles and electrical distribution, control, wire and cable end up in this column is rather surprising. According to Table 5, these sectors are also the two sectors with most export product relocation. Looking at the order of sectors in the lefthand column, we can see that they are all export oriented and have relatively high frequencies of export relocation.

11 These variables that are highly correlated with size are estimated separately to avoid problems concerning multicollinearity.

106

Export dynamics and product relocation

Table 7 Coefficients of zero inflated Poisson with sectors only*

-Metals

-Non-metallic products

-Other electrical equipment

-Man-made fibres

-Fabricated metal products

-Petroleum products and

nuclear fuel

-Other transport equipment

-Electric motors,

generators and transformers

-Paper

-Wood products

-Energy machinery

-Medical equipment

-Domestic appliances

-Signal transmission and

telecommunication

-Other chemicals

-Special purpose machinery

-Optical instruments

-Basic metals

-Motor vehicles

-Measuring instruments

-Agricultural and forestry

machinery

-Electrical distribution,

control, wire and cable

-Paints, varnishes

-Non-specific purpose

machinery

-Publishing and printing

-Weapons and ammunition

-Watches, clocks

-Soaps, detergents, toilet preparations

-Electronic components

-Pharmaceuticals

-Pesticides and agro-chemical

products

-Television and radio receivers, audio-visual electronics

-Office machinery and computers

-Machine tools

-Rubber and plastic products

-Lightening equipment

-Accumulators, batteries

-Rubber and plastic products

-Lightening equipment

-Accumulators, batteries

*Each column list sectors in order of size with respect to their z-values

LR chi2=1050.02 prob>chi2=0.000; the excluded sector is furniture, consumer goods

5. Conclusions and further research

This paper analyses how regional characteristics affect the regional capacity to absorb relocated export products, defined as export products that have left one region (exit region) and entered the export base of another region. To do so, the number of completely relocated products was counted for the 38 sectors in the manufacturing sector in all 81 functional regions in Sweden. The number of relocated export products was explained with variables describing (i) the

107

Jönköping International Business School

regional absorptive capacity, (ii) the regional export support system and (iii) the domestic demand.

Starting with the findings regarding factors related to regional absorptive capacity, knowledge intensity in the own sector positively affects the likelihood of a high rate of regional export base renewal through relocation of export products. The effect is further strengthened when introducing sector dummies to detect systematic differences which are attributable to sectors. Sector dummies are included to capture the effects not explained by other control variable, and the effect of within-sector knowledge increases considerably in magnitude when controlling for sector heterogeneity. With sector dummies the influence of surrounding knowledge is negative, suggesting that what actually matters for relocating an export product is the knowledge that can be related to the sector of the actual product. The product and its production processes has assumingly already become standardised and thus the product is less reliant on general knowledge spillovers generated from the surrounding regional economic milieu.

The findings also indicate that the number of trade links (export and import) has the strongest effect on the success of attracting new export products. The export support system is important to establish relocated export products in a region. By looking at completely relocated products, we actually examine the second entry of products. The significance of an export support system is strengthened even further by the positive sign for accessibility to ports and the negative effect of a location farther away from the Stockholm metropolitan region. It is not a very striking finding that sectors that are traditionally associated with a path-dependent location pattern show significantly different results than other sectors. A number of sectors behave as basic sectors but cannot be associated with similar location-specific characteristics. This needs to be further disentangled. Also, these sector differences should be further related to regional differences.

The demand variables contribute to explaining relocation of export products to a varying extent, and total accessibility to inter-regional and intraregional potential customers turn out as strongly affecting the number of relocated products coming into the entry region. The growth of the regional demand in the entry region is also important to entry of relocated export products. The same is true for regional size measured as the number of employees in the region.

If firms, to a great extent renew their export bases with the help of other regions’ exports, it is from a regional perspective important to create favourable

108

Export dynamics and product relocation

economic environments and regional business milieus. Considering previous research, relocation is to great extent a wish to move production to a location where production and distribution costs are lower. In a world where international trade increases and transportation costs falls, an alternative to inter-regional export product relocation is a relocation is across national borders.

There is still much room for advancements and further analysis, especially of sector differences in speed and rate of export base renewal and relocation patterns.

109

Jönköping International Business School

References

Abd-el-Rahman, K. (1991). Firms’ competitive and national comparative advantages as joint determinants of trade composition. Review of World Economics, 127(1), 83-97.

Abernathy, W. J., & Utterback, J. M. (1978). Patterns of Industrial Innovation.

Technology Review, 80(7), 40-47.

Aiginger, K. (1997). The use of unit values to discriminate between price and quality competition. Cambridge Journal of Economics, 21(5), 571-592.

Andersson, M., & Karlsson, C. (2006). Regional Innovation Systems in Small and

Medium-Sized Regions: A Critical Review and Assessment. In B. Johansson, C.

Karlsson & R. Stough (Eds.), The Emerging Digital Economy: Entrepreneurship, Clusters,

and Policy (pp. 55-82). Berlin: Springer-Verlag.

Andersson, Å. E., & Johansson, B. (1984). Knowledge Intensity and Product Cycles in

Metropolitan Regions IIASA Working Paper 84-13. Laxenburg, Austria: International

Institute for Applied Systems Analysis.

Audretsch, D. B. (1995). Innovation and Industry Evolution. Cambridge, MA: MIT Press.

Bernard, A. B., & Bradford, J. J. (1999). Exceptional exporter performance: cause, effect, or both? Journal of International Economics, 47(1), 1-25.

Bernard, A. B., Jensen, J. B., Kortum, S., & Eaton, J. (2003). Plants and Productivity in

International Trade. American Economic Review, 93(4), 1268-1290.

Bernard, A. B., Jensen, J. B., Redding, S. J., & Schott, P. K. (2007). Firms in

International Trade. Journal of Economic Perspectives, 21(3), 105-130.

Bernard, A. B., Redding, S. J., & Schott, P. K. (2006). Multi-Product Firms and Product

Switching. NBER Working Paper Series No.12293. Working Paper. National Bureau of

Economic Research. Cambridge, MA.

Bernard, A. B., Redding, S. J., & Schott, P. K. (2011). Multiproduct Firms and Trade

Liberalization. Quarterly Journal of Economics, 126(3), 1271-1318.

Boschma, R. A. (2004). Competitiveness of Regions from an Evolutionary Perspective

Regional Studies, 38(9), 1001-1014.

Breschi, S., & Lissoni, F. (2001). Knowledge Spillovers and Local Innovation Systems:

A Critical Survey. Industrial and Corporate Change, 10(4), 975-1005.

Brouwer, A. E., Mariotti, I., & van Ommeren, J. N. (2004). The firm relocation decision: An empirical investigation. Annals of Regional Science, 38(2), 335-347.

Brouwer, E., Budil-Nadvornikova, H., & Kleinknecht, A. (1999). Are Urban

Agglomerations a Better Breeding Place for Product Innovation? An Analysis of

New Product Announcements. Regional Studies, 33(6), 541 - 549.

Burns, A. F. (1934). Production trends in the United States since 1870. New York, NY:

National Bureau of Economic Research.

Cameron, A. C., & Trivedi, P. K. (1986). Econometric Models Based of Count Data:

Comparisons and Applications of Some Estimators and Tests. Journal of Applied

Econometrics, 1(1), 29-53.

Comanor, W. S., & Wilson, T. A. (1967). Advertising Market Structure and

Performance. Review of Economics and Statistics, 49(4), 423-440.

Das, S., Roberts, M. J., & Tybout, J. R. (2007). Market Entry Costs, Producer

Heterogeneity, and Export Dynamics. Econometrica, 75(3), 837-873.

Dumais, G., Ellison, G., & Glaeser, E. L. (2002). Geographic Concentration as a

Dynamic Process. Review of Economics and Statistics, 84(2), 193-204.

110

Export dynamics and product relocation

Duranton, G., & Puga, D. (2001). Nursery Cities: Urban Diversity, Process Innovation, and the Life Cycle of Products. American Economic Review, 91(5), 1454-1477.

Eaton, J., Kortum, S., & Kramarz, F. (2005). An Anatomy of International Trade:

Evidence from French Firms Working Paper. New York University: Department of

Economics.

Erickson, R. A. (1976). The filtering-down process: Industrial location in a nonmetropolitan area. The Professional Geographer, 28(3), 254-260.

Ewers, H.-J., & Wettman, R. W. (1980). Innovation-oriented regional policy. Regional

Studies, 14(3), 161-179.

Falvey, R. E., & Kierzkowski, H. (1987). Product Quality, Intra-Industry Trade and

(Im)perfect Competition. In H. Kierzkowski (Ed.), Protection and Competition in

International Trade: Essays in Honor of W.M. Corden. New York, NY: Basil Blackwell.

Flam, H., & Helpman, E. (1987). Vertical Product Differentiation and North-South

Trade. American Economic Review, 77(5), 810-822.

Forslund, U., & Johansson, B. (1995). The Mälardalen: A Leading Region in

Scandinavia and Europe? In P. C. Cheshire & I. R. Gordon (Eds.), Territorial

Competition in an Integrating Europe (pp. 3-27). Aldershot, UK.

Fritsch, M., & Falck, O. (2002). New Firm Formation by Industry over Space and Time: A

Multi-Level Analysis Freiberg Working Paper # 11 2002. Faculty of Economics and

Business Administration. Freiberg University of Mining and Technology.

Goldberg, P. K., Khandelwal, A. K., Pavcnik, N., & Topalova, P. (2010). Imported

Intermediate Inputs and Domestic Product Growth: Evidence from India. The

Quarterly Journal of Economics, 125(4), 1727-1767.

Hayter, R. (1997). The Dynamics of Industrial Location: The Factory, the Firm and the Production

System. New York, NY: Wiley.

Jacobs, J. (1969). The Economy of Cities. New York, NY: Random House.

Johansson, B. (1993). Ekonomisk dynamik i Europa [in Swedish]. Malmö: Liber-Hermods i samarbete med Institutet för framtidsstudier.

Johansson, B., & Karlsson, C. (1991). Från brukssamhällets exportnät till kunskapssamhällets

importnät [in Swedish]. Karlstad: Länsstyrelsen i Värmlands län.

Johansson, B., & Karlsson, C. (2003). Växande branscher-om Stockholmsregionens samspel med

övriga landet [in Swedish]. Storstadspolitik 8:2003. Stockholms läns landsting:

Regionplane- och trafikkontoret.

Johansson, B., Klaesson, J., & Olsson, M. (2002). Time distances and labor market integration. Papers in Regional Science, 81(3), 305-327.

Johansson, B., Klaesson, J., & Olsson, M. (2003). Commuters’ non-linear response to time distances. Journal of Geographical Systems, 5(3), 315-329.

Krugman, P. (1980). Scale Economies, Product Differentiation, and the Pattern of

Trade. American Economic Review, 70(5), 950-959.

Kuznets, S. (1960). Population Change and Aggregate Output. With Comments by R.E

Quandt and M.Freidman. In J. C. Ansley (Ed.), Demographic and Economic Change in

Developed Countries (pp. 340-367). Princeton, NJ: Princeton University Press.

Markusen, J. R. (1985). Profit Cycles, Oligopoly, and Regional Development. Cambridge, MA:

MIT Press.

Marshall, A. (1890/1920). Principles of Economics. London, UK: Macmillan.

Marshall, A. (1920). Industry and trade- a study of industrial technique and business organization;

and of their influences on the conditions of various classes and nations (3. ed.). London:

Macmillan.

111

Jönköping International Business School

Martin, F., Swan, N., Banks, I., Barker, G., & Beaudry, R. (1979). The Interregional

Diffusion of Innovations in Canada. Ottawa: Economic Council of Canada.

Maurseth, P. B., & Verspagen, B. (1999). Europe: one or several systems of innovation?

An analysis based on patent citations. In J. Fagerberg, P. Guerrieri & B. Verspagen

(Eds.), The Economic Challenge for Europe: Adapting to Innovation-Based Growth (pp. 149-

174). Cheltenham, UK and Northampton, MA: Edward Elgar.

McCann, P. (2001). Urban and Regional Economics. Oxford, UK: Oxford University Press.

Moghimbeigi, A., Eshraghian, M. R., Mohammad, K., & McArdle, B. (2008). Multilevel zero-inflated negative binomial regression modeling for over-dispersed count data with extra zeros. Journal of Applied Statistics, 35(10), 1193-1202.

Moses, L. N. (1958). Location and the Theory of Production. Quarterly Journal of

Economics, 72(2), 259-272.

Nelson, R. R., & Winter, S. G. (1982). An Evolutionary Theory of Economic Change

Cambridge, MA Harvard University Press.

Norton, R. D., & Rees, J. (1979). The product cycle and the spatial decentralization of

American manufacturing. Regional Studies, 13(2), 141-151.

Nutek. (1998). Småföretag och regioner i Sverige 1998: med ett tillväxtperspektiv för hela landet [in Swedish]. Report B 1998:10. Stockholm: Närings- och teknikutvecklingsverket: Nutek.

Oakey, R. P., Thwaites, A. T., & Nash, P. A. (1980). The regional distribution of innovative manufacturing establishments in Britain. Regional Studies, 14(3), 235-253.

Ohlin, B. (1933). Interregional and International Trade. Cambridge, MA: Harvard University

Press.

Panzar, J. C., & Willig, R. D. (1981). Economies of Scope. American Economic Review,

71(2), 268-272.

Pavitt, K. (1984). Sectoral patterns of technical change: Towards a taxonomy and a theory. Research Policy, 13(6), 343-373.

Polanyi, M. (1966). The Tacit Dimension. London, UK: Routledge & Kegan Paul.

Portes, R., & Rey, H. (2005). The determinants of cross-border equity flows. Journal of

International Economics, 65(2), 269-296.

Schmenner, R. W. (1980). Choosing New Industrial Capacity: On-Site Expansion,

Branching, and Relocation. Quarterly Journal of Economics, 95(1), 103-119.

Schumpeter, J. (1939). Business Cycles: A Theoretical, Historical and Statistical Analysis of the

Capitalist Process. New York, NY: McGraw Hill.

Sellenthin, M. O., & Hommen, L. (2002). How innovative is Swedish industry? A factor and cluster analysis of CIS II. International Review of Applied Economics, 16(3), 319-331.

Sjöfartsverket. (2006). Sjöfartens utveckling 2006- Sjöfartverkets sektorrapport [in

Swedish].Norrköping: Sjöfartsverket.

Teece, D. J. (1980). Economies of scope and the scope of the enterprise. Journal of

Economic Behavior & Organization, 1(3), 223-247. van Dijk, J., & Pellenbarg, P. H. (2000). Firm relocation decisions in The Netherlands:

An ordered logit approach. Papers in Regional Science, 79(2), 191-219.

Vernon, R. (1966). International Investment and International Trade in the Product

Cycle. Quarterly Journal of Economics, 80(2), 190-207.

112

Export dynamics and product relocation

Appendix A1

Following Andersson and Johansson (1984), one can assume that the cost function has the following simplified form for product k produced in region r :

= +

, where and is a regional specific process associated with labour input a and capital input b. As the scale of production increases, production is standardised and the rate of b decreases. That is, knowledge intensity of production decreases over time such that

< <. . . . <

and these gradual movements would follow

⁄ > 0

and

⁄ < 0

along the product cycle. The dynamics of production can be summarised in two related conditions. First, in order for product k to relocate it must be that the b -coefficient decreases at a faster rate than the a coefficient. This means a standardisation process with a substitution of capital with labour. Second, the entry region must have a lower relative price of standardised inputs (a) and knowledge-intensive inputs (b) compared to the exit region.

113

Jönköping International Business School

Appendix A2

The details on the sample extraction is illustrated in Figure A1. Stages 1 to 7 illustrate the location possibilities of export products, and the sum of products in Stage 7 are the number of completely relocated products. Export product k

(combination firm identity code and product identity code) belongs to sector j and is completely relocated if at time t it is exported from region r and at time t-

τ is exported from region m where r≠m. At time t region m’s export of product

k is zero and at time t-τ region r’s export of product k is zero. Consequently, region r is the entry region and region m is the exit region. Export products located in the same region at t and t-τ are export continuers.

All exported products ∑

= ∑

in all regions

at time t (1)

Product k exists in region r at time t-τ (2)

New product, k exists in region

r at time t but not at time t-τ

(3)

Product k does not exist in region m at time t-τ, where

r≠m (4)

Product k exists in region

m at time t-τ, where r≠m (5)

Product k exists in region m at time t, where r≠m (6)

Product k does not exist in region

m at time t, where r≠m (7)

Figure A1 Sample extraction from original data

114

12

13

14

15

16

17

18

19

20

21

22

23

24

Table A 2 Sectors in manufacturing industry

7

8

9

10

11

3

4

5

6

Sector Description

1

2

Wood products

Paper

Publishing and printing

Petroleum products, nuclear fuel

Basic metals

Pesticides, agro-chemical products

Paints, varnishes

Pharmaceuticals

Soaps, detergents, toilet preparations

Other chemicals

Man-made fibres

Rubber and plastic products

Non-metallic products

Metals

Fabricated metal products

Energy machinery

Non-specific purpose machinery

Agricultural and forestry machinery

Machine tools

Special purpose machinery

Weapons and ammunition

Domestic appliances

Office machinery and computers

Electric motors, generators and transformers

35

36

37

38

25

26

27

28

29

Electrical distribution, control, wire and cable

Accumulators, batteries

Lightening equipment

Other electrical equipment

Electronic components

30 Signal transmission, telecommunication

31 Television

32 audio-visual electronics

Medical equipment

33

34

Measuring instruments

Optical instruments

Watches, clocks

Motor vehicles

Other transport equipment

Furniture, consumer goods

116

Paper 3

Imports, knowledge flows and renewal of regional exports

Martin Andersson, Lina Bjerke & Charlie Karlsson

.

3 …… .……….

117

Imports, knowledge flows and renewal of regional exports

Martin Andersson, Lina Bjerke & Charlie Karlsson

ABSTRACT

We examine the role of regional import flows for renewal of regional industries.

The hypothesis is that imports stimulate renewal of local industries by being vehicles for technology diffusion and means by which local firms can exploit advantages of global specialisation. We find robust and positive relationships between high-quality imports and renewal of regional exports, where the latter are measured by the introduction of novel export products of local firms.

Connectedness to international markets via import networks appears as a stimulus for the renewal of regional exports.

Keywords: imports, exports, industry renewal, innovation, networks, knowledge spillovers

JEL classification codes: R10, R11, R12, D85, F10

118

Imports, knowledge flows and renewal of regional exports

Martin Andersson, Lina Bjerke & Charlie Karlsson

1. Introduction

In a dynamic economy with strong competition at a global scale, firms need to continuously renew and upgrade products and services to retain their competitive position. An extensive literature argues that supply-side characteristics of the locality in which firms are located play an important role for firms’ capacity to do so (Feldman, 1999; Porter, 1998).

1

The local scale is emphasised as a source of knowledge and ideas through spillover phenomena and ‘local buzz’, localised institutions and systems features, as well as regionally embedded inter-firm cooperation structures (Cooke, 2001; Feldman, 2000;

Gertler, 1995; Lawson & Lorenz, 1999). The broader literature on agglomeration and localisation externalities, such as Glaeser et al (1992) and

Rosenthal and Strange (2001), shares this focus on place-specific characteristics and resources. Several scholars have yet emphasised a need to appreciate and integrate the importance of extra-regional and international linkages (Bathelt et al., 2004; Shin et al., 2006). There is for example increasing evidence that firms combine local and global sources in their product renewal and innovation processes (Asheim & Isaksen, 2002; Moodysson et al., 2008; Simmie, 2003,

2004; Trippl, 2011). Extra-regional and global linkages can of course take many forms such as trade networks, FDI and strategic alliances and the literature seldom specifies what types of extra-regional linkages are important.

This paper concurs with latter strand of literature and focuses on the role of extra-regional linkages in the form of import networks for the renewal of local manufacturing industries. Import networks involving different source countries throughout the world constitute pertinent linkages bearing on regional

1

As an example, Porter (1998, p.78) stated that ‘the enduring competitive advantages in a global economy lie increasingly in local things – knowledge, relationships, motivations – that distant rivals cannot match’. In a similar vein, Maskell and Malmberg (1999, p. 172) argued that “the formation of the world market .... increases the importance of heterogeneous, localised capabilities for building firm-specific competencies”.

119

Jönköping International Business School

upgrading and renewal processes. These linkages are means by which local firms can exploit global specialisation advantages and are vehicles for knowledge flows and inflow of novelties to a region (Coe & Helpman, 1995;

Lööf & Andersson, 2010).

The idea of a relationship between imports and renewal and upgrading of local industries is certainly not a new one. Studies of the historical role of cities as ‘trading centres’ typically regard imported products as sources of ideas for new products and emphasise the innovation advantages of large-city regions with a diverse supply of products from all over the world (Braudel, 1949/1966;

Heckscher, 1931; Jacobs, 1969, 1984). Modern analyses in this vein include

Simmie (2003) and Simmie and Sennett (1999), who show that metropolitan regions acting as ‘international trading gateways’ constitute important milieus for innovative firms, in particular for their ability to offer access to codified and tacit knowledge originating from multiple sources and spatial scales. Another strand of literature focuses on imports as vehicles for international technology diffusion (Coe & Helpman, 1995; Keller, 2004; Madsen, 2008). This literature takes R&D-based models of endogenous growth as their starting point, according to which local firms can access foreign R&D by importing capital goods.

2

This perspective emphasises high-quality import flows with high knowledge and technology contents that can bring about learning effects and embodied technical change (Lööf & Andersson, 2010). High-quality imports to a region may thus reflect linkages for the inflow of new knowledge and technology, stimulating product upgrading and renewal.

Despite an extensive literature, the empirical evidence on the role of highquality import flows for renewal of regional industries is still scant. The literature on international technology diffusion often disregards the regional perspective and frames analyses at the level of nations or sectors. At the same time, most of the studies on the role of global sources of knowledge for regional industrial renewal and innovation do not consider imports, but instead build on survey-based information on firms’ knowledge sourcing activities.

There is to our knowledge no study that specifically links high-quality import flows to regional industrial renewal and upgrading. A study that does bear on import networks is Boschma and Iammarino (2009), who examine the role of related and unrelated variety on regional growth through the lens of trade data for Italian provinces. The focus of this paper is yet different and fills a gap in the literature.

2

Each capital good is assumed to be based on a unique design, developed in an R&D process. By importing capital goods, then, a firm can indirectly access technologies developed in foreign countries (cf. Rivera-Batiz

& Romer, 1991).

120

Imports, knowledge flows and renewal of regional exports

We make use of Swedish longitudinal audited register data on exports and imports tabulated across regions and sectors to assess the role of regional highquality imports for renewal of regions’ manufacturing export sectors. Export renewal of a sector and a region is measured by the appearance of novel export products of different qualities over a time interval of seven years. New highquality export products are assumed to reflect renewal and upgrading of regional export sectors.

3

The distinction between imports and novel export products of different qualities is based on established methods developed in the literature on horisontal and vertical product differentiation in international trade (Abd-el-Rahman, 1991; Greenaway et al., 1995; Torstensson, 1991).

The empirical analysis is designed to test whether regional high-quality imports can explain subsequent renewal and upgrading of export sectors. We set up an empirical model in which initial regional characteristics explain subsequent introduction of novel export products, focusing on the influence of high-quality import flows. We control for several other characteristics that may influence renewal and upgrading of regional exports, such as human capital, measures of agglomeration economies, general openness of the region, distance to metropolitan regions, local presence of advanced business services and industry composition. To examine potential spillover effects across sectors in a region, we also include high-quality imports and human capital in other sectors as regressors.

We find a significant and positive influence of high-quality imports on the introduction of new high-quality export products that is robust to the inclusion of several control variables motivated by previous literature. Our results also suggest import ‘spillover effects’ in the sense that high-quality imports in other sectors have a consistent positive impact on a sector’s export renewal. We also find that the general openness of a regional economy (as measured by export and imports as fraction of regional GDP) has a positive influence. These findings may be interpreted as evidence that regional ‘international trading gateways’ indeed stimulate industrial upgrading and renewal their local sectors

(cf. Simmie, 2003; Simmie & Sennett, 1999).

3

A merit of this measure is that we can consistently identify novel products made in different regions and sold on international markets. Conceptually, the measure comes close to Schumpeter’s (1934) suggestion that both product design and identification of destination markets are parts of an innovation.

121

Jönköping International Business School

2. Imports, knowledge flows and export renewal-conceptual framework

The growth of the export base is a main driver of jobs, incomes and welfare in many regions and countries; one of the sources of such growth is continuous renewal and upgrading of the export supply. Knowledge and technology are essential inputs in such processes, and there are several arguments that import networks are important sources in this context. The R&D activities in a given region for instance, only make up a small fraction of the total volume of R&D investments in the world economy. Imports constitute a means to access the global knowledge stock (Dollar et al., 1988; Grossman & Helpman, 1991a,

1991b, 1994; Keller, 2004; Marin, 1995). Knowledge and technology development in other regions of the world is an important generator of changes in products and production processes, even in larger agglomerations. This suggests that imports of high-quality products may play an important role in the renewal of industries in different regions.

Conceptually, there are several perspectives on the role of imports for industrial renewal at the regional level. One is that imports are means to access inputs in the production process of regional sectors and thus a means to renew the regional capital stock. Another perspective is provided by studies in economic history (Braudel, 1949/1966; Heckscher, 1931; Jacobs, 1969, 1984), which regard imported products as a source of ideas for new products and emphasise the advantages of large city regions with a diverse supply of products from all over the world. Both types of perspectives point to a link between imports and renewal at the regional level. We discuss these perspectives below.

Starting with imports and their relation to production processes, there are clear arguments that imports as inputs in production processes are positively associated with static gains in, for example, productivity. Imports are a means to exploiting global specialisation and employing inputs developed at the forefront of knowledge and technology. Imports can, for instance, imply lower costs for capital and material inputs, access to inputs of higher quality as well as an enlargement of the variety of inputs in the production process (cf. Amiti &

Konings, 2007; Broda & Weinstein, 2006; Halpern et al., 2006).

Imports of intermediate goods to the production process are however not only associated with static gains, but can also like final goods generate dynamic effects in the form of ideas for novel products or improvements of existing products. Such dynamic effects are often labelled ‘learning effects’ in the literature (cf. Acharya & Keller, 2008) and focus on imports of high quality

122

Imports, knowledge flows and renewal of regional exports

capital goods with high knowledge and technology content. The learning effect is an important feature of the literature on international technology diffusion, surveyed in Keller (2004). The conceptual framework of this literature is rooted in R&D-based models of growth and trade, in which technology and knowledge are embodied in differentiated intermediate goods. Moreover, imports of high-technology inputs by firms in a region may sharpen the innovation activity of local (potential) suppliers attempting to develop substitute goods.

Imports of high-quality goods can generate learning effects that influence industrial renewal in several ways. For instance, the technology embodied in imported goods can make the production of novel products possible and stimulate ideas for improvements and expansion of existing product lines.

From a regional perspective, it is also possible for high-quality import flows to a subset of firms in a region to generate spillover effects to other firms. Such spillover effects may take several forms. The knowledge and technology accumulated by the importing firm may spill over to other firms in the region through, for instance, labour mobility of engineers (cf. Almeida & Kogut, 1997) and other technical personnel, which can stimulate knowledge accumulation and renewal activities in other firms. The knowledge imported from abroad can thus spill over between firms in the region through various knowledge flows.

These knowledge flows can in turn stimulate innovation (Glaeser, 1994;

Glaeser & Saiz, 2004). In a similar fashion, the local presence of firms that import of high-quality goods may induce other firms to start importing through demonstration effects.

Some authors have also argued that throughout history, imported products have been a source of the initiation of new product cycles in regions. The

Swedish economist and economic historian Eli Heckscher claimed for example that imports stimulate local production that substitutes imports (Johansson,

1993). Similar ideas are present in the works by Jane Jacobs (1969, 1984). She maintained that imported products have throughout history been a source of new ideas and initiation of new product cycles in city regions.

4

Imported new products can conceptually of course generate strong incentives for imitation or development of complements, since their presence for sale implies that they have passed two types of tests. The first is that it has been proved that there are technical solutions for the new product that work and the second is that the

4

In Jacobs’s theoretical scheme, innovation and renewal are stimulated, in particularly in large city regions, by richness and diversity in import flows. The driving force for economic renewal is the profits that firms can capture by substituting import flows with their own production and sales, i.e. through import substitution.

123

Jönköping International Business School

import flow verifies that there are customers, which means that there is a regional market.

In summary, there are strong arguments in different strands of literature that import networks constitute an important type of extra-regional linkages for upgrading and renewal of regional sectors. In particular, high-quality products to a region can amplify the regional knowledge stock and stimulate renewal and upgrading of the regional export base. Empirical evidence scant but the subsequent section presents an empirical test of the role of high-quality import flows for the renewal of regional exports using Swedish trade data tabulated on sectors and regions.

3. Data and construction of variables

The analysis rest on two main sources of data, both of which are maintained by

Statistics Sweden (SCB).

5

The first is data on exports and imports over a sevenyear period 1997-2003 tabulated on municipalities, firms and products in

Sweden. Export and import products are classified according to the CN 8-digit classification system.

6

The data cover the manufacturing industry, which is aggregated into 38 sectors (see Table A1 in the Appendix).

7

These data allow for an analysis of how the composition of regional export flows change over time, such that novel export products can be identified. The information on both volumes (kg) and values (SEK) in fine product categories makes it possible to apply methods to discriminate between low- and high-quality imports and exports.

The second set of data consists of information on employment of labour with different levels of education in each sector and functional region in

Sweden. These data are used to derive control variables such as measures of the level of education and industry structure. We also make use of a matrix with travel time distances by car between regions in Sweden to compute variables reflecting time distances to metro regions.

The unit of analysis is 81 functional regions (Local Labour Market, LLM), each consisting of a number of municipalities connected by intense commuting flows (Nutek, 1998). They are characterised by a high intensity of intra-regional commuting flows and are delineated based on the intensity of observed

5

This is not a public data base and has restricted access.

6

CN is the European Community’s Combined Nomenclature based on the Harmonised Commodity

Description and Coding System of the Customs Cooperation Council.

7

Exports and imports of services are not included.

124

Imports, knowledge flows and renewal of regional exports

commuting flows between municipalities. The commuting intensity declines sharply at the borders of the LLMs. In this sense, it can be argued that the functional regions are arenas for face-to-face interaction (Andersson &

Karlsson, 2006).

3.1 Measurement of high-quality imports and export renewal

Our variables of main focus are high-quality imports and novel high-quality export products. To classify import and novel export products to different quality levels we follow the literature on horizontal and vertical product differentiation. Horizontally differentiated products are assumed to be of similar quality, whereas vertically differentiated products are of different quality

(Falvey & Kierzkowski, 1987; Flam & Helpman, 1987).

Empirical operationalisations of product qualities in international trade contexts are based on the assumption that differences in quality are reflected in differences in average unit values of trade flows in narrowly defined sectors (see e.g. Abd-el-Rahman, 1991; Aiginger, 1997). Products in a given sector with similar unit values are assumed to be horizontally differentiated whereas products with different unit values are classified as vertically differentiated.

Higher unit values are assumed to reflect higher quality, which is consistent with the observation that quality is reflected in consumers’ willingness to pay.

Following this logic, we identify high-quality import products by comparing the import value per volume unit (kilogram) of each 8-digit product with the median value of the import value per volume unit (kilogram) for each 8-digit product at the national level. A given import product is classified as high-quality if its value per kilogram is higher than the national median import value per kilogram in the 3-digit industry to which it belongs. An import product is defined as a unique combination of a product and a firm identity code. Firms are thus assumed to export differentiated products. The product code refers to an 8-digit CN classification code. Since each firm can be spatially identified and classified into sectors, the aggregation of high-quality and total imports across regions and sectors is straightforward.

Products with a value per volume unit above the median are classified as high-quality, whereas products below the median are classified as low-quality imports. Let i denote an import product and let S

HQ

=

{1,…,n} be a set of high-quality import products and S

HQ

is a subset of all import products at the 8-

125

Jönköping International Business School

digit level

r is given by:

. The fraction of high-quality imports in sector s in region

,

=

,

,

(1)

, where the numerator sums the values of all high quality imports of 8-digit products included in the 3-digit sector, s, for each firm f in region r.

Turning to the identification of new export products, i.e. export renewal, we use the same strategy to identify new export products as Andersson and

Johansson (2008). An export product refers to a unique combination of a product and firm identity code. The number of new export varieties in a given sector and region is based on the appearance of new combinations of a products and firm identity codes. To make this measure formally precise, let

denote the export value of export product i in sector s and region r in period t. If denotes the set of new export products between period (t) and period + , all elements i in this set satisfy the following function:

= : = 0⋀ + > 0

(2)

In the subsequent analysis, we have (t) = 1997 and + = 2003. The number of elements in N across regions within this time interval is the number of new export products, which is assumed to reflect export renewal. Highquality export products are defined in line with Equation 2:

= : = 0⋀ + > 0

(3)

A limitation of export data is that they do not provide information on whether a variety has been sold on domestic markets before being exported. In principle, there are four alternatives for the introduction of a new product: (1) introduction on domestic markets and finally exports with a time lag, (2) immediate sales on both domestic and foreign markets, (3) introduction on export markets only and (4) only domestic sales. In the empirical analysis, we focus on new export products that emerge over a seven-year period. Even though (1) applies and there is a time lag from domestic market to export, as long as the time lag is roughly constant, the number of new export products

126

Imports, knowledge flows and renewal of regional exports

will reflect export innovations. Due to a lack of data, we have to accept that trade between functional regions inside Sweden at the 8-digit code cannot be observed. It is only when a supply flow passes a country border that it can be observed. Since Sweden’s domestic market is relatively small and a relatively large fraction of firms in manufacturing sectors are active on international markets, we consider this a minor problem in the present analysis. As in the case of import products, since each export product is based on a firm identity code and each firm has a spatial identifier, the new export products can readily be connected to specific regions.

While Andersson and Johansson examine the frequency of new export products, we focus in this paper on novel high-quality exports. The fraction of new high-quality exports is identified in a way similar to high-quality imports.

We compare the value per volume unit (kilogram) of each novel export product with the median value per kilogram of novel exports at the national level in the

3-digit industry to which the product belongs. In this way, we create a subset

of novel high-quality export products. Thus, from Equations 2 and 3 we define the share of new high-quality export products as a fraction of total new exports in sector s and region r in the following way:

=

(4)

, which will be used as the dependent variable in the empirical analysis. Highquality imports are assumed to be an external knowledge source and a vehicle for knowledge and technology flows. We expect that the knowledge and technology content of the novel export products is higher in regions with a larger fraction of high-quality imports. The next section presents the distribution of import diversity, high-quality imports and novel high-quality export products across functional regions in Sweden.

3.2 High-quality imports and export renewal across Swedish regions

Figure 1 illustrates the novel high-quality exports as a fraction of all novel export products in Sweden’s functional regions in the year 2003. The different shades of grey represent the percentage shares of all new exports in a region that are recognised as new high-quality export products. The map contains the

127

Jönköping International Business School

289 municipalities in Sweden: all municipalities in a functional region are assigned the same value. The large concentration of high-quality exports in some functional regions in the northern parts of Sweden can be explained by high-technology exports such as particular vehicles and electronic components.

Figure 2 illustrates the percentage shares of all imports in Sweden’s regions that are recognised as high-quality imports. High-quality imports are concentrated to the area around the metropolitan region of Stockholm, though we also observe a number of regions located in the northern part of the country with a large fraction of high-quality imports. Stockholm is an important node in the

Swedish economy with large flows of imports and exports as well as a large share of employees with a university degree. Figures 1 and 2 indicate that highquality imports are more spatially concentrated than new high-quality exports.

This supports the proposition that a limited number of regions function as import nodes, supporting other regions with high-quality imports (cf.

Andersson and Johansson, 2000).

% share of new high quality exports

0.032-0.221

0.221-0.324

0.324-0.434

0.434-0.552

0.552-0.807

Figure 1 Regional shares of new high-quality exports in total new exports in the year 2003.

128

Imports, knowledge flows and renewal of regional exports

% share of high quality imports

0.006-0.246

0.246-0.296

0.296-0.379

0.379-0.438

0.438-0.942

Figure 2 Regional shares of high-quality imports in total imports in the year 2003.

4. Empirical strategy, model and results

We wish to estimate the influence of high-quality imports of a sector and region on the introduction of novel high-quality exports, where both measures are expressed as fractions. This means that our dependent variable is fractional and by construction bounded within the

[0,1] interval. A common strategy in a context like this is to apply a log-odds transformation of the fractional dependent variable P, such that ∗

= ⁄ 1 −

. In this case, P* is assumed to be linearly related to the explanatory variables, and the model is estimated with ordinary least squares (OLS). This transformation yields predictions that lie within the

[0,1] interval, but as discussed by Papke and Wooldridge (1996)

129

Jönköping International Business School

and Wooldridge (2002, p.662 ), it has two basic problems. First, it does not allow P to take the extreme values 0 or 1. Second, the estimated probability

cannot be recovered without additional distributional assumptions. In our case, there are several observations for which the fractional dependent variable, takes the value 0 or 1. Consequently, there are regions were sectors have zero new-high quality exports or regions with sectors with only new high-quality exports.

We employ the Fractional Logit Model (FLM), which can account for observations for which the fraction is 0 or 1 and is more flexible than an OLS model on log-odds transformed variables (Papke & Wooldridge, 1996). The

FLM applies a quasi-maximum likelihood estimation procedure. Let Χ denote explanatory variables. The expected fraction of new high-quality exports is assumed to be related to Χ through a logistic function:

∣ =

(5)

, which is estimated with the Bernoulli log likelihood function (see Wooldridge,

2002, p.460 )

,

= ∣ Χ + 1 − 1 − ∣ Χ

(6)

Table 1 describes the explanatory variables in Χ . In the Appendix, Table A1 shows the 38 sectors in the manufacturing industry, Table A2 presents the pairwise correlation for all variables, and Table 3A displays the descriptive statistics.

130

Imports, knowledge flows and renewal of regional exports

Table 1 Explanatory variables reflecting regional characteristics in the regression analysis

Variable Description Justification

(expected outcome)

High-quality exports in sector sr (t+τ)

High-quality imports in sectorsr (t)

High-quality imports in sectors other than sr (t)

Knowledge intensity in sector sr

Overall regional trade intensity

r (t) (log)***

Knowledge intensity in sectors other than sr (t)

Size of region

r (t) (log)

Fraction of total export value in 2003 generated by novel export products with a value per volume unit above the median in Sweden (see Equation 3)

Fraction of total import value in

1997 generated by import products with a value per volume unit above the median in Sweden (see Equation 1)

Fraction of total import value in

1997 generated by import products with a value per volume unit above the median (average for other sectors)

Percentage share of workforce in sector s with a long (> 3 years) university education in sector

s in the region in 1997

The sum of export value and import value as a share of the total gross regional product 1997 in the region

Average percentage share of workforce in sectors other than

s with a long (> 3 years) university education in the region in 1997

Total employment in the

region in 1997

Regressand

Vehicle for knowledge

And technology flows from abroad to the sector (+)

Spillover effects across sectors (+)

Knowledge and absorptive capacity of the workforce in each sector and region (+)

Openness in region indicating opportunities of knowledge inflows and knowledge sharing (+)

Spillover effects, human capital externalities (+)

Distance from region r to nearest metropolitan region (log)

Export dummy sector sr (t)

Local presence of knowledge- intensive business services r (t)

Firm size sr (t)

Average wage sr

(t)

Sector dummy

Time distance by car to the

closest metropolitan region

(either Stockholm, Göteborg or Malmö)

Dummy that equals 1 if the region has any type of exports in

sector s 1997, 0 otherwise

Employment in NACE 72-74 as a fraction of total employment in region r in 1997

Average number of employees

in firms 1997 in sector s in the region

Average wage per employee in firms 1998 in sector s in the region

Dummy that equals 1 if observation belongs to sector s,

0 otherwise

Agglomeration and

Urbanisation phenomena in each region (+)

Proximity to diversity of sectors and richness of knowledge sources (-)

Sector-specific export experience in each sector in the region (+)

Local presence of ancillary service sectors as incubators for exports (+)

The effect of sector size composition (-)

Indicator of the average productivity of workers in sector s in region r (+)

Control for heterogeneity across sectors.

131

Jönköping International Business School

Though the explanatory variable of main interest is the fraction of high-quality imports we control for a number of other regional characteristics that could influence novel high-quality export products. To allow for spillover effects we include the fraction of regional high-quality imports in sectors other than s. The average knowledge intensity of the employees in the sector in question as well as that in all other sectors are assumed to be two other control variables.

Average knowledge intensity is approximated by the fraction of employees with a long university education (at least three years). These variables are assumed to reflect the absorptive capacity of the workers in the region and also the potential for human capital spillovers (cf. Glaeser, 1994; Rauch, 1993).

Development of high-quality export products may also be assumed to be associated with product development processes requiring employees with long university education.

Furthermore, the model includes a variable reflecting the general openness of the regional economy. This is measured by the sum of the region’s export and import values as a fraction of regional gross domestic product. The general openness of the regions is intended to reflect overall connectedness to international markets, which may affect the inflow of knowledge and information to the regional economy. Arguments developed by, among others,

Simmie (2002, 2003) suggest that regions acting as ‘international trading gateways’ are favourable environments for renewal and upgrading processes.

8

Regional size in terms of employees is intended to capture general agglomeration phenomena, in particular the potential influence of external economies of scale associated with the size of the regional economy

(Andersson & Lööf, 2011) Distance to the closest metropolitan region (either

Stockholm, Göteborg or Malmö) accounts for proximity to different knowledge sources and diversity in sectors. The regional employment fraction in knowledge-intensive business services (defined as sectors 72-74 according to the two-digit NACE classification system) reflects the local presence of ancillary services that may stimulate regional exports (cf. Macpherson, 2008).

We further include a dummy variable that is 1 if the region has exports in sector s, reflecting export experience in the sector. We assume that regions with export experience in a sector are more prone to develop new export products in that sector. Firm size and average wages are two final control variables.

Average wages are assumed to reflect the overall productivity of workers in the sector and region. As regards firm size, there are two alternative interpretations.

8

As alternatives to openness we have also tested for general trade diversity in terms of the number of source and destination countries. These alternative indicators have no impact on the results and show overall lower significance than the openness measure described in the table.

132

Imports, knowledge flows and renewal of regional exports

One is that small firms are more flexible and innovative (Acs & Audretsch,

1988), such that regions in which a sector is dominated by smaller firms are better apt to generate high-quality export products.

9

An alternative view is that the local presence of large firms acting as ‘gatekeepers’ is an important stimulus to the local economy (Morrison, 2008). The variable average firm size may capture both effects, making the expected influence ambiguous. The variables finally also include sector dummies to account for heterogeneity across sectors.

4.1 Results

This section presents the results of the analyses of the influence of high-quality imports on the generation of new high-quality export products across sectors in

Swedish regions. Table 2 presents marginal effects based on FLM estimates, whereas Table 3 presents Tobit model estimates where the dependent variable is modified. As is evident from both tables, the main results are also robust to various alternative specifications of the main model.

The estimates reported in Table 2 support the main presumption of the paper: high-quality imports stimulate the generation of new high-quality export products. The results also show that high-quality imports in other sectors as well as overall regional openness have a positive influence. This supports the idea that regional trade networks foster product renewal and upgrading processes, where the positive effect of high-quality imports is not confined to the sectors responsible for the imports.

Most of the control variables have the expected sign. The average knowledge intensity of the employees in the sector has a positive influence on high-quality new export products. Average knowledge intensity in other sectors is yet not significant, suggesting limited spillover effect across sectors. Regional size is positive and significant in all specifications. This is consistent with our expectations and provides support for the assumption that agglomeration phenomena stimulate renewal and upgrading of regional export sectors. The estimated marginal effect of distance to the nearest metropolitan region is positive but insignificant. Conditional on all other control variables, it is thus not of importance for explaining novel high quality export products. Export experience in each specific sector is positively associated with introducing novel export products of high-quality. This may be interpreted as evidence that established exports of a sector and region have accumulated export experience

9

Smaller firms may also need a sharper focus on novel exports rather than exports in general in order to keep a market position.

133

Jönköping International Business School

and knowledge, which stimulates the development of new high-quality export products.

Average firm size in the sector and region has a negative effect on novel high-quality exports renewal. One interpretation of this result is that, everything else being equal, small firms need to be relatively more innovative than their larger counterparts in order to stay viable (Acs & Audretsch, 1988).

Furthermore, average wages of a sector and regions as well as the local presence of knowledge-intensive business services do not have any significant influence on export renewal in terms of novel high-quality exports.

134

Imports, knowledge flows and renewal of regional exports

Table 2 Quasi-maximum likelihood estimates of explanatory variables in 1 on the fraction of new high-quality exports. (fractional logit model).Dependent variable: High-quality exports in sector s (t)

Variable 1 4 5 6

High-quality imports in sector s (t)

High-quality imports in sectors other than s

(t)

Overall trade intensity/

Knowledge intensity in sector s

Average knowledge intensity in sectors other than s (t)

Size of region (t) (log)

All control variables

0.119*

(5.45)

0.158*

(3.73)

0.028*

(3.81)

0.133*

(3.24)

-0.016

(-0.05)

Excluding average knowledge intensity in other sectors

0.119*

(5.45)

0.158*

(3.77)

0.028*

(3.79)

0.133*

(3.24)

-

Excluding wage sum

0.119*

(7.45)

0.157*

(3.75)

0.029*

(3.78)

Excluding distance to metropolitan region

0.119*

(5.41)

0.168*

(4.08)

0.025*

(3.48)

Excluding local presence of KI business services

0.118*

(5.39)

0.163*

(3.98)

0.027*

(3.80)

Not controlling for sector heterogeneity

0.151*

(7.14)

0.206*

(5.07)

0.024*

(3.30)

0.132*

(3.22)

0.132*

(3.21)

0.128*

(3.10)

0.170*

(4.07)

- - - -

Distance to closest metropolitan region

(log)

Export dummy sector

s (t)

Local presence of knowledge-intensive business services (t)

Firm size (t)

0.053*

(6.00)

0.015

(1.51)

0.198*

(9.20)

-0.408

(-1.21)

0.053*

(6.25)

0.015

(1.65)

0.198*

(9.26)

-0.417

(-1.64)

0.053*

(6.25)

0.015

(1.65)

0.198*

(9.25)

-0.418

(-1.64)

0.046*

(6.56)

0.198*

(9.27)

-0.4439

(-1.74)

0.039*

(6.90)

0.200*

(9.49)

0.027*

(4.73)

0.246*

(14.78)

- -

Wage sum (t)

Sector dummies

-2.85

-4 *

(-2.79)

-8.97e

-9

(-0.45)

-2.85

-4 *

(-2.79)

-8.96e

-9

(-0.45)

-2.87

-4 *

(-2.80)

-2.86

-4 *

(-2.77)

-2.82

-4 *

(-2.75)

-2.80

-4 *

(-3.09)

- - - -

Included Included Included Included Included Excluded

# obs 3 078 3 078 3 078 3 078 3 078 3 078

Note: The table reports marginal effects after GLM. The dependent variable is new high-quality exports as a fraction of the total value of new exports between 1997 and 2003. See Equations 3 and 4 in Section 3.1 for definitions. All explanatory variables were measured in 1997, z-values in parentheses and * p<0.01

135

Jönköping International Business School

As a test of the robustness and sensitivity of the results in Table 2, we also perform estimations with an alternative dependent variable. We estimate a

Tobit model with the average log of the average export volume unit value of all new export products in a sector and region. The estimations in Table 2 are based on high-quality novel export products as a fraction of all new products, whereas the Tobit estimates show the effects of the explanatory variables on the average export price (value per kilogram) of all new export products in a sector and region. The idea behind this additional estimation is to test whether the results are sensitive to pre-specified quality levels. When employing the alternative dependent variable, we do not rely on a pre-defined threshold for high quality, though higher average export prices of the novel products are still assumed to reflect higher quality. The reason for employing the Tobit estimator in the presence of the alternative dependent variable is that the average export prices of a sector and region are left-censored with 0 as their minimum value.

10

The results of this undertaking are presented in Table 3. Our main results are unaffected, i.e. the Tobit estimates also support the hypothesis that highquality imports play a significant role for high-quality exports. High-quality imports in other sectors as well as general openness are also positive and significant. We also find a positive and significant impact of within-sector knowledge intensity, but as before, the knowledge intensity of employees in other sectors has no statistically significant influence. Regional size and export experience are positive and significant while distance to the closest metropolitan area is insignificant. Average wage, average firm size and the local presence of knowledge-intensive business services are all insignificant. The main difference between Tables 2 and 3 is that average firm size is insignificant in the latter table.

In summary, the empirical analysis lends support to the argument that extraregional linkages in the form of import networks of high-quality products from foreign source countries are important for the renewal and upgrading of local export sectors, as measured by the introduction of novel high-quality export products. We also find evidence of spillover effects in the sense that highquality imports in other two-digit sectors in a region have a positive influence on export renewal and upgrading of a sector in a region. Furthermore, the general openness of a region to international trade has a consistent positive effect on high-quality new export products. These results are robust to the inclusion of several control variables.

10

Average export price is zero for those sectors and regions that do not introduce novel export products.

136

Imports, knowledge flows and renewal of regional exports

Table 3 Tobit estimates. Dependent variable: Average export unit value (log),

*** p<0.01, ** p<0.05, *p<0.1

High-quality imports in sector s (t)

All control variables

0.987***

(6.53)

Excluding average knowledge intensity in other sectors

0.983***

(6.51)

Excluding wage sums

0.982***

(6.49)

4

Excluding distance to metropolitan region

0.954***

(6.53)

5

Excluding local presence of

KI business services

0.956***

(6.53)

6

Not controlling

for sector heterogeneity

1.512***

(8.95)

1.245***

(4.46)

1.263***

(3.77)

1.287***

(4.63)

1.247***

(4.54)

1.235***

(4.50)

1.245***

(4.01)

High-quality imports in sectors other than s (t)

Overall trade

Knowledge intensity in sector

s

Average knowledge intensity in sectors other than s (t)

Size of region (t)

(log)

Distance to closest metropolitan region

(log)

Export dummy sector s (t)

0.203***

(4.37)

0.930***

(3.33)

-1.751

(-0.73)

0.206***

(3.33)

-0.083

(-1.19)

2.681***

(17.53)

0.046***

(4.32)

0.921***

(3.30)

- - -

0.193***

(3.25)

-0.067

(-1.01)

2.269***

(17.70)

0.202***

(4.38)

0.953***

(3.43)

0.194***

(3.27)

-0.067

(-1.00)

2.700***

(17.74)

0.214***

(4.79)

0.953***

(3.42)

-

0.227***

(4.61)

2.700***

(17.73)

-

0.229***

(5.35)

0.927***

(3.34)

-

0.198***

(4.61)

2.709***

(17.83)

0.184***

(3.45)

1.911***

(6.14)

-2.474

(-0.90)

0.140**

(2.02)

-0.0957

(-1.18)

3.062***

18.73)

-1.549

(-0.64)

Local presence of knowledgeintensive business services (t)

Firm size (t)

Wage sum (t)

-1.309

(-0.62)

-5.743e

(-1.06)

-4

2.450e

-7

(1.47)

-2.20

(-1.27)

-5.671e

(-1.05)

-4

2.460e

-7

(1.48)

-2.175

(-1.25)

-2.061

(-1.19)

-5.456e

(-1.01)

-4

- -

-5.547e

-4

(-1.03)

-

-5.431e

(-1.01)

-4 -1.200e

(-1.84)

-3 *

4.360e

(2.29)

-7 **

Sector dummies

# obs

YES

3 078

YES

3 078

YES

3 078

YES

3 078

YES

3 078

NO

3 078

137

Jönköping International Business School

5. Conclusions and future research

Analyses of regional growth and renewal more often than not emphasise local supply-side characteristics such as density of skilled workers, R&D resources and market size. While several recent papers have argued that extra-regional and international linkages are important for renewal and upgrading of regional sectors, there is little evidence on what types of linkages matter. This paper has emphasised connectedness to international markets in terms of import networks as an important form of extra-regional linkages conducive to renewal and upgrading of regional export sectors. Drawing on the literature emphasising imports as a vehicle for inflow of knowledge and technology from abroad and as a stimulus for new ideas, it has focused on high-quality import flows as a determinant of export renewal as evidenced by novel high-quality export products of sectors in different regions.

The empirical analysis supports the hypothesis that imports of high-value products are important for the renewal of exports of sectors at the regional level. We find that the quality of a region’s imports has a significant influence on the export renewal of sectors and regions, where renewal is approximated by novel high-quality export products. This result is robust and remains after controlling for several other regional characteristics, such as distance to metropolitan regions, local pool of skilled workers, regional size, local presence of ancillary service providers and firm size. We also find evidence of a regional spillover effect in the sense that novel high-quality export products in a sector and region are positively influenced by the high-quality imports of other sectors in the same region. Furthermore, the general openness of a region to international trade has a consistent positive effect on renewal and upgrading of regional export sectors.

A general implication of the paper is that conceptualisations and models of regional growth and industrial renewal need to include ‘the global’ as well as

‘the local’ and their interplay. Trade networks appear to be an important regional asset in this context, which act as a conduit for an inflow of new knowledge and information. The results lend overall support to the idea that regions acting as ‘international trading gateways’ are indeed conducive for processes pertaining to renewal and upgrading of regional sectors (cf. Simmie,

2003). In particular, regions that combine import linkages to the global economy with a strong local milieu appear as markedly favourable regional environments for upgrading and renewal processes. We find, for example, evidence that the introduction of novel high-quality export products is

138

Imports, knowledge flows and renewal of regional exports

positively influenced by both regional size (a standard proxy for agglomeration phenomena and external scale economies), knowledge intensity of the local workforce in the sector and import networks and general openness. Our empirical results are thus consistent with the conceptual frameworks developed by authors such as Bathelt et al (2004) and Simmie (2003, 2004) who emphasise that the simultaneous presence of ‘local buzz’ and ‘global pipelines’ is a key for innovative regions with strong export sectors.

139

Jönköping International Business School

References

Abd-el-Rahman, K. (1991). Firms’ competitive and national comparative advantages as joint determinants of trade composition. Review of World Economics, 127(1), 83-97.

Acharya, R. C., & Keller, W. (2008). Estimating the Productivity Selection and

Technology Spillover Effects of Imports NBER Working Paper No. 14079 (Vol. No.

14079). Cambridge, MA: National Bureau of Economic Research.

Acs, Z. J., & Audretsch, D. B. (1988). Innovation in Large and Small Firms: An

Empirical Analysis. American Economic Review, 78(4), 678-690.

Aiginger, K. (1997). The use of unit values to discriminate between price and quality competition. Cambridge Journal of Economics, 21(5), 571-592.

Almeida, P., & Kogut, B. (1997). The Exploration of Technological Diversity and

Geographic Localization in Innovation: Start-Up Firms in the Semiconductor

Industry. Small Business Economics, 9(1), 21-31.

Amiti, M., & Konings, J. (2007). Trade Liberalization, Intermediate Inputs, and

Productivity: Evidence from Indonesia. The American Economic Review, 97(5), 1611-

1638.

Andersson, M., & Johansson, B. (2008). Innovation Ideas and Regional Characteristics:

Product Innovations and Export Entrepreneurship by Firms in Swedish Regions.

Growth and Change, 39(2), 193-224.

Andersson, Å. E., & Johansson, B. (2000). The Stockholm region in the global economy. In Å. E. Andersson & D. Andersson (Eds.), Gateways to the Global

Economy. Cheltenham, UK & Northampton, MA: Edward Elgar.

Andersson, M., & Karlsson, C. (2006). Regional Innovation Systems in Small and

Medium-Sized Regions. In B. Johansson, C. Karlsson & R. Stough (Eds.), The

Emerging Digital Economy: Entrepreneurship, Clusters, and Policy (pp. 55-82). Berlin:

Springer-Verlag.

Andersson, M., & Lööf, H. (2011). Agglomeration and productivity: evidence from firm-level data. The Annals of Regional Science, 46(3), 601-620.

Asheim, B. T., & Isaksen, A. (2002). Regional Innovation Systems: The Integration of

Local ‘Sticky’ and Global ‘Ubiquitous’ Knowledge. Journal of Technology Transfer,

27(1), 77-86.

Bathelt, H., Malmberg, A., & Maskell, P. (2004). Clusters and knowledge: local buzz, global pipelines and the process of knowledge creation. Progress in Human Geography,

28(1), 31-56.

Boschma, R., & Iammarino, S. (2009). Related Variety, Trade Linkages, and Regional

Growth in Italy. Economic Geography, 85(3), 289-311.

Braudel, F. (1949/1966). The Mediterranean and the Mediterranean World in the Age of Philip

II. New York, NY: Harper & Row

Broda, C., & Weinstein, D. E. (2006). Globalization and the Gains from Variety.

Quarterly Journal of Economics, 121(2), 541-585.

Coe, D. T., & Helpman, E. (1995). International R&D spillovers. European Economic

Review, 39(5), 859-887.

Cooke, P. (2001). Regional Innovation Systems, Clusters, and the Knowledge

Economy. Industrial and Corporate Change, 10(4), 945-974.

Dollar, D., Wolff, E. N., & Baumol, W. J. (1988). The Factor-Price Equalization Model and Industry Labour Productivity: An Empirical Test across Countries. In R. C.

140

Imports, knowledge flows and renewal of regional exports

Feenstra (Ed.), Empirical Methods for International Trade (pp. 23-47). Cambridge, MA:

MIT Press.

Falvey, R. E., & Kierzkowski, H. (1987). Product Quality, Intra-Industry Trade and

(Im)perfect Competition. In H. Kierzkowski (Ed.), Protection and Competition in

International Trade: Essays in Honor of W.M. Corden (pp. 143-161). New York, NY:

Basil Blackwell.

Feldman, M. P. (1999). The New Economics of Innovation, Spillovers and

Agglomeration: A Review of Empirical Studies. Economics of Innovation and New

Technology, 8(1-2), 5-25.

Feldman, M. P. (2000). Location and Innovation: The New Economic Geography of

Innovation, Spillovers and Agglomeration. In G. L. Clark, M. P. Feldman & M. S.

Gertler (Eds.), The Oxford Handbook of Economic Geography (pp. 373-394). Oxford,

UK: Oxford University Press.

Flam, H., & Helpman, E. (1987). Vertical Product Differentiation and North-South

Trade. American Economic Review, 77(5), 810-822.

Gertler, M. (1995). "Being there": proximity, organization, and culture in the development and adoption of advanced manufacturing technologies. Economic

Geography, 71(1), 1-26.

Glaeser, E. (1994). Cities, Information, and Economic Growth. Cityscape: A Journal of

Policy Development and Research, 1(1), 9-47.

Glaeser, E. L., Kallal, H. D., Scheinkman, J. A., & Shleifer, A. (1992). Growth in Cities.

Journal of Political Economy, 100(6), 1126-1152.

Glaeser, E. L., & Saiz, A. (2004). The Rise of the Skilled City [with Comments by G.T

Burtless & W.S Strange]. Brookings-Wharton Papers on Urban Affairs, 5, 47-105.

Greenaway, D., Hine, R., & Milner, C. (1995). Vertical and Horizontal Intra-Industry

Trade: A Cross Industry Analysis for the United Kingdom. The Economic Journal,

105(433), 1505-1518.

Grossman, G. M., & Helpman, E. (1991a). Innovation and Growth in the Global Economy.

Cambridge, MA: MIT Press.

Grossman, G. M., & Helpman, E. (1991b). Trade, knowledge spillovers, and growth.

European Economic Review, 35(2-3), 517-526.

Grossman, G. M., & Helpman, E. (1994). Technology and Trade NBER Working Paper

No.4926. Cambridge, MA: National Bureau of Economic Research.

Halpern, L., Koren, M., & Szeidl, A. (2006). Imports and Productivity. Mimeograph. New

York, NY: Federal Reserve Bank of New York.

Heckscher, E. F. (1931). Merkantilismen [in Swedish] Stockholm: Norstedt & Söner.

Jacobs, J. (1969). The Economy of Cities. New York, NY: Random House.

Jacobs, J. (1984). Cities and the Wealth of Nations: Principles of Economic Life. New York,

NY: Random House.

Johansson, B. (1993). Ekonomisk dynamik i Europa: nätverk för handel, kunskapsimport och

innovationer [in Swedish]. Malmö: Liber-Hermods i samarbete med Institutet för framtidsstudier.

Keller, W. (2004). International Technology Diffusion. Journal of Economic Literature,

42(3), 752-782.

Lawson, C., & Lorenz, E. (1999). Collective Learning, Tacit Knowledge and Regional

Innovative Capacity. Regional Studies, 33(4), 305-317.

Lööf, H., & Andersson, M. (2010). Imports, Productivity and Origin Markets: The Role of Knowledge-intensive Economies. World Economy, 33(3), 458-481.

141

Jönköping International Business School

Macpherson, A. (2008). Producer service linkages and industrial innovation: results of a

Twelve-Year Tracking Study of New York State Manufacturers. Growth and Change,

39(1), 1-23.

Madsen, J. B. (2008). Economic Growth, TFP Convergence and the World Export of

Ideas: A Century of Evidence. Scandinavian Journal of Economics, 110(1), 145-167.

Marin, D. (1995). Learning and Dynamic Comparative Advantage: Lessons from

Austria's Post-war Pattern of Growth for Eastern Europe CEPR Discussion Paper

No. 1116. London, UK: Centre for Economic Policy Research.

Maskell, P., & Malmberg, A. (1999). Localised learning and industrial competitiveness.

Cambridge Journal of Economics, 23(2), 167-185.

Moodysson, J., Coenen, L., & Asheim, B. (2008). Explaining spatial patterns of innovation: analytical and synthetic modes of knowledge creation in the Medicon

Valley life-science cluster. Environment and Planning A, 40(5), 1040-1056.

Morrison, A. (2008). Gatekeepers of Knowledge within Industrial Districts: Who They Are,

How They Interact. Regional Studies, 42(6), 817-835.

Nutek. (1998). Småföretag och regioner i Sverige 1998: med ett tillväxtperspektiv för hela landet.

Report B 1998:10. [in Swedish]. Stockholm: Närings- och teknikutvecklingsverket:

Nutek.

Papke, L. E., & Wooldridge, J. M. (1996). Econometric methods for fractional response variables with an application to 401(k) plan participation rates. Journal of Applied

Econometrics, 11(6), 619-632.

Porter, M. E. (1998). Clusters and the New Economics of Competition. Harvard Business

Review, November-December 1998, 77-90.

Rauch, J. E. (1993). Does History Matter Only When It Matters Little? The Case of

City-Industry Location. Quarterly Journal of Economics, 108(3), 843-867.

Rivera-Batiz, L. A., & Romer, P. M. (1991). International trade with endogenous technological change. European Economic Review, 35(4), 971-1001.

Rosenthal, S. S., & Strange, W. C. (2001). The Determinants of Agglomeration. Journal

of Urban Economics, 50(2), 191-229.

Schumpeter, J. A. (1934). The Theory of Economic Development: An Inquiry into Profits, Capital

Credit, Interest, and the Business Cycle. Cambridge, MA: Harvard University Press.

Shin, M. E., Agnew, J., Breau, S., & Richardson, P. (2006). Place and the Geography of

Italian Export Performance. European Urban and Regional Studies, 13(3), 195-208.

Simmie, J. (2002). Trading Places: Competitive Cities in the Global Economy. European

Planning Studies, 10(2), 201-214.

Simmie, J. (2003). Innovation and Urban Regions as National and International Nodes for the Transfer and Sharing of Knowledge. Regional Studies, 37(6-7), 607-620.

Simmie, J. (2004). Innovation and Clustering in the Globalised International Economy.

Urban Studies, 41(5-6), 1095-1112.

Simmie, J., & Sennett, J. (1999). Innovative clusters: global or local linkages? National

Institute Economic Review, 170(1), 87-98.

Torstensson, J. (1991). Quality differentiation and factor proportions in international trade: An empirical test of the Swedish case. Review of World Economics, 127(1), 183-

194.

Trippl, M. (2011). Regional innovation systems and knowledge-sourcing activities in traditional industries-evidence from the Vienna food sector. Environment and Planning

A, 43(7), 1599-1616.

Wooldridge, J. M. (2002). Econometric Analysis of Cross Section and Panel Data. Cambridge,

MA: MIT Press.

142

Appendix

Imports, knowledge flows and renewal of regional exports

32

33

34

35

36

37

38

26

27

28

29

30

31

Table A1 Industries in manufacturing sectors

11

12

13

14

15

16

7

8

9

10

17

18

19

20

21

22

23

24

Sector Description

1

2

3

4

5

6

Wood products

Paper

Publishing and printing

Petroleum products, nuclear fuel

Basic metals

Pesticides, agro-chemical products

Paints, varnishes

Pharmaceuticals

Soaps, detergents, toilet preparations

Other chemicals

Man-made fibres

Rubber and plastic products

Non-metallic products

Metals

Fabricated metal products

Energy machinery

Non-specific purpose machinery

Agricultural and forestry machinery

Machine tools

Special purpose machinery

Weapons and ammunition

Domestic appliances

Office machinery and computers

Electric motors, generators and transformers cable

Accumulators, batteries

Lightening equipment

Other electrical equipment

Electronic components

Signal transmission, telecommunication

Television and radio receivers, audiovisualelectronics

Medical equipment

Measuring instruments

Optical instruments

Watches, clocks

Motor vehicles

Other transport equipment

Furniture, consumer goods

143

Imports, knowledge flows and renewal of regional exports

Table A3 Summary statistics

Variable

Mean Median Std.dev Min Max

High-quality new exports in sector s (t)

0.308 0.137 0.353 0 0.999

0.287 0.179 0.307 0 0.999 High-quality imports in sector s (t)

High-quality imports in sectors other than s (t)

0.237 0.237 0.191 0 0.960

Overall trade intensity/

Openness in region (t)

(log)

Knowledge intensity in sector s

Average knowledge intensity in sectors other than s (t)

Size of region (t) (log)

0.090 0

10.708 10.656

0.158 0 1

0.037 0.029 0.031 0 0.192

1.298 8.000 14.469

Distance to closest metropolitan region (log)

Export dummy sector s

(t)

0.828 1 0.377 0 1

Local presence of knowledge-intensive business services (t)

Firm size (t) 18.501 0 77.092 0 1.26e

7

Wage sum (t) 96668.09 94045.63 244705.30 0 2490

145

Paper 4

Knowledge-intensive business services, creative labour inflow and firm productivity

Lina Bjerke

.

4 …… .……….

147

Knowledge-intensive business services, creative labour inflow and firm productivity

Lina Bjerke

ABSTRACT

Two of the most striking changes associated with post-industrialism are the increasing share of jobs in the service sector and the knowledge-biased technological change. These structural shifts have restructured the labour market in such a way that a substantial number of firms in the service sector are now highly dependent on knowledge and, in particular, the inflow of new knowledge to be competitive. This paper analyses the role of creative labour inflow for the firm productivity in knowledge-intensive business services

(KIBS), where productivity is measure as value added per employee. Using data for firms and individuals in Sweden, the results support previous empirical studies but at the same time sheds new light on the relation between creative labour and KIBS firms. More specifically, creative labour inflow is positive for all firms but the effects are larger for firms in the KIBS sector. The nature and characteristics of KIBS firms imply specific capacities to attract creative labour and by doing so, they also absorb the new knowledge these employees embody.

Keywords: creative labour, labour inflow, labour mobility, KIBS, firm productivity, regional business milieu

JEL classification codes: J60, J24, R10, L25, L80

148

Knowledge-intensive business services, creative labour inflow and firm productivity

Lina Bjerke

1. Introduction

The post-industrial era brought a number of fundamental changes affecting the labour market. One such structural change is knowledge-biased technological change, which has implied a rapid increase in demand for skilled workers, in particular in the service sector, skilled workers who are able to provide comprehensive and customised interpretations of random and uncertain data, information and technology (Lundvall, 1998; O'Mahony et al., 2008). A second structural change is the steady growth of the number of employees in the service sector, including both standardised services and knowledge services

(EMCC, 2005). A third and striking change is the development and rapid growth of information and communication technologies (ICT), shaping the way information and knowledge are handled, stored, diffused and absorbed. These economic changes have promoted fast growth in knowledge-intensive business services (KIBS) in most developed economies.

1

In Sweden, firms in KIBS have experienced rapid growth after Sweden started to recover after the economic downturn at the beginning of the 1990s. Between the years 1990 and 2010, firms in the KIBS sector increased their number of employees (workplaces) by

119 (103) per cent compared to the 17 (21) per cent growth for all firms in

Sweden. The exceptional growth of the KIBS sector has spurred great interest

1

Firms are KIBS if they are classified into codes 72-74 of the Swedish Standard Industrial Classification, SNI.

This classification follows the EU standard NACE. For a detailed description see Table A1 in the Appendix.

Occupations are classified according to the Swedish standard for occupational classification, SSYK which is an adaptation of the International Standard Classification of Occupations, ISCO. SSYK and ISCO are based on two main concepts: (i) kind of work performed, defined as a set of tasks or duties designed to be executed by one person, and (ii) skill, defined as the skill level, the degree of complexity of constituent tasks, and skill

specialisation, the field of knowledge required for competent performance of the constituent tasks. For a detailed description see Table A2 in the Appendix. Data provided by Statistics Sweden. This is not a public data base and has restricted access.

149

Jönköping International Business School

in this sector. Furthermore, the firms in this sector have, beside their remarkable growth, a number of special characteristics worth studying.

Knowledge-intensive business service firms are particularly good at supporting innovation in client firms as facilitators, carriers and sources of innovation (den Hertog, 2000; Miles et al., 1995). Czarnitzki and Spielkamp

(2003) call them “bridges for innovation” and describe them as the link between manufacturing industries and business-related services. Firms in the

KIBS sector compete mainly with knowledge and work tasks that to a high extent are associated with the development, absorption and diffusion of knowledge. Thus, employees are to a high extent hired for their ability to solve complex problems and their capacity to absorb and develop knowledge and further apply it in collaboration with a client firm (Miles, 1999).

2

Another feature of KIBS firms is that they have a large share of employees with consultancy and creative occupations, who are particularly skilled at absorbing new knowledge from a variety of knowledge sources at home and abroad (Mellander, 2009). The empirical analysis in this paper uses micro data covering all firms in Sweden for the period 2001 to 2008 and it highlights how the inflow of labour with creative occupations affects the productivity of KIBS firms. The results show that creative labour inflow has a positive effect on firm productivity that is larger for KIBS firms than for non-KIBS firms. This analysis differs from that in the previous literature in a number of aspects.

Primarily, firms in the KIBS sector are considered as unique with respect to dependence on their creative labour inflow and have been given specific focus in the analysis. Secondly, the influence of the regional economic milieu is given special considerations based on the argument it has a supplementary role to play for firm productivity besides the creative labour input.

Creative individuals are attracted to regions with a larger demand for their capability and experience, and we can expect knowledge spillovers from the surrounding milieu to play an important role for what these creative individuals can achieve. The approach in this paper is in line with the paper by Boschma et al., (2009), who examine the effects of skill portfolios and labour inflow on plant productivity. That study used Swedish data with an primarily focused on firm-related characteristics and labour inflows; it shows that labour inflow per se does not have a positive effect on firm productivity but that productivity depends on the type of skills that flow into the firm and whether the incoming labour originates from the same geographical region or comes from another region.

2 Firms in KIBS also play a role in innovations, business start-ups, and growth of regions (Czarnitzki &

Spielkamp, 2003; Miles, 2003; Miles et al., 1995; Muller & Zenker, 2001).

150

Knowledge-intensive business services, creative labour inflow and firm productivity

2. Creative labour inflow and KIBS productivity

Firm productivity can be defined as efficiency in production, i.e., how much output is obtained from a given set of inputs. This is a topic of particular interest and it has persistently documented that there are substantial differences across firms even within a very narrowly defined set of industries

(Hsieh & Klenow, 2009; Syverson, 2004, 2011). While, the productivity concept is relatively straightforward, productivity measures are not. Should it be aggregated into a single output measure or should it be measured as a quality adjusted input measure (e.g. wageS measuring marginal product of labour)? To handle intermediate inputs one normally uses output and subtracts intermediates and then deals with value added (Syverson, 2011).

3

Knowledgeintensive business services can be representatives for the knowledge economy where knowledge is both an input and an output (Gallouj, 2002; Miles, 2001) and their primary valued added consist of the identification, creation, accumulation and dissemination of knowledge with the purpose of creating customised service solutions and innovations (Bettencourt et al., 2002). In accordance with studies in labour economics, worker’s knowledge (human capital) can be one factor explaining part of the productivity differences between firms (Abowd et al., 2005; Fox & Smeets, 2011).

In terms of knowledge input, educational attainment has dominated as the explanatory factor of firm productivity for a long time (Glaeser & Saiz, 2004).

Building on alternative streams of academic research one can argue that education measures indicate the underlying capability of employees but say nothing about how much of this capability that is converted into productive work. The conceptual framework for analysing real productivity effects could instead be based on labour creativity and the idea that creativity can be separated from education. Creativity is the ability to adopt and react to constant changes in the milieu and involves the skills to transform these changes into something that improves the productivity of firms (Amabile, 1983; Karlsson,

2011; Runco, 2004; Smith et al., 1984). The creative class is an alternative measure for human capital (Glaeser, 2005), based on occupations and specifically occupations in science, engineering, arts, culture, entertainment and the

3

It is acknowledged that it can be difficult to account traditional productivity measures for services since much of their output and innovations are intangible (Leiponen, 2005). In the regression analysis of this paper, we also use alternative measures to measure performance (wage per employee, wage per employee in sector in regions, and sales per employee). The effects of the control variables are robust for all performance measures.

However, this need further research.

151

Jönköping International Business School

knowledge-based professions of management, finance, law healthcare and education (Florida, 2002a, 2002b, 2002c). By focusing on an occupational-based creative class measure, one can analyse what employees actually do at work and better capture the effect of the absorptive and innovative capacity.

4

There is an overlap between creative people and people with higher education but the two are not identical. An early attempt to show this while connecting occupations, educations and industries was presented by Andersson (1985). There are fundamental differences between industries and occupations: industries are categorised according to their final products with no consideration on the occupational distributions within industries. A relatively small part of literature has followed up on this and by doing so, one can better examine what employees do with their knowledge and what is converted into labour productivity and innovation (Florida, 2002c; Florida et al., 2008; Marlet & van

Woerkens, 2004; Mellander, 2009; Mellander & Florida, 2007).

2.1 Creative labour and knowledge diffusion

A number of studies constitute the empirical foundation of national labour mobility and knowledge diffusion. Saxenian’s (1994) work on Route 128 in the

USA, the study by Glaeser et al., (1992) on 170 US cities, and the study on engineers in the semiconductor industry by Angel (1989) all present evidences that individuals are carriers of knowledge. Workers acquires knowledge at work and when they leave a job for a new one, they do not leave their knowledge behind. A range of studies emphasise the impact of mobility of skilled labour and the way workers diffuse embodied knowledge (Almeida & Kogut, 1999;

Andersson & Thulin, 2008; Braunerhjelm, 2007; Cooper, 2001; Maskell &

Malmberg, 1999; Power & Lundmark, 2004).

The link between embodied knowledge flows and firm productivity has not yet fully been explored (Dahl & Sorenson, 2007). While incoming knowledgeintensive labour adds new knowledge to the existing knowledge base, high rates of labour inflow generate better firm productivity if, and only if, firms have the capacity to absorb and apply this new knowledge. Not all knowledge is difficult to share but when knowledge is tacit, uncertain and complex ,skilled workers with education, competence and experience are the only ones able to handle

4

Studies criticise the empirical validity of theory of the creative class. The majority of them deal with the question whether creative individuals are the engines of economic growth; Hoyman and Faricy (2009) find no evidence for this. Also, Markusen (2006) disapproves of the index structure; she argues that the creative class is highly heterogeneous and argues that occupations are largely based on educational attainment and the additional grouping of workers into a creative class is insignificant when controlling for education.

152

Knowledge-intensive business services, creative labour inflow and firm productivity

such complex knowledge sharing (Almeida & Kogut, 1999; Cooper, 2001;

Maskell & Malmberg, 1999; Power & Lundmark, 2004). In a study using

Swedish data, Boschma et al., (2009) scrutinize the impact of skilled labour inflow on plant productivity and find that labour mobility per se does not affect firm productivity. They differentiate across types of incoming labour and primarily show that incoming knowledge can have a negative or positive effect for firm productivity, depending on degree of similarities to existing knowledge.

It could be that incoming knowledge adds nothing or very little to existing knowledge base, i.e. lock-in effects (Nooteboom et al., 2007; Noteboom, 2000).

2.2 Regional economic milieu

The emphasis on labour inflows for firm productivity is not to deny the importance of the regional economic milieu. Agglomeration of economic actors generates spillover effects and externalities (Jacobs, 1969; Krugman, 1991;

Marshall, 1890/1920). Thus, knowledge spillovers and related externalities tend to be more frequent in denser regions. Therefore, these regions are potentially good locations for firms that are highly dependent upon new knowledge.

Knowledge intensive business services are characterised by strong supplier-user interactions. This is a feature that can be associated with agglomeration and geographic proximity (Johansson & Quigley, 2004). It is also likely that KIBS firms locate in dense regions due to demand-side factors. The complex work tasks performed by KIBS firm employees most often require direct face-to-face contacts with clients. These firms are confronted with the problem of finding solutions to customers’ by recombining existing knowledge and complementing it with new knowledge inputs. Thus, they are largely engaged in consultancy work tasks, which require close contacts with clients (Muller & Zenker, 2001).

Larger markets allow potentially fixed costs to be distributed over more clients and the lower transportation costs in larger and denser markets allow firms to cut the unit costs of their services. The effect of reduced transportation costs depends on the distance sensitivity of the services supplied and to some extent one can assume that the ICT development has an effect on some knowledgeintensive services but certainly not on all (Andersson & Beckmann, 2009).

There is a vast body of research illustrating the role of KIBS in innovation and growth of regions (Czarnitzki & Spielkamp, 2003; Marshall et al., 1987;

Miles, 2003; Miles et al., 1995; Muller & Zenker, 2001).

5

However, there is not

5

Knowledge spillovers in agglomeration economies are accounted for in detail by Duranton and Puga (2004),

Glaeser (2008), Fujita and Thisse (2002) and Johansson and Quigley (2004). Closely related to, but contrary to

Jacobs’s diversity, is the debate on competition externalities (Porter, 1990).

153

Jönköping International Business School

much research showing how agglomeration and regional knowledge spillovers affect the productivity of KIBS firms.

Summarising the discussion above, knowledge intensive business services

(KIBS) have particular characteristics and are dependent upon their ability to absorb new knowledge and to incorporate it into their existing knowledge base to serve other customers. Creative labour inflow has a positive effect on firm productivity in the KIBS sector and that inflow of creative occupations have a larger impact on firm productivity in the KIBS sector than on firm productivity in the non-KIBS sector. This follows since people with creative occupations can be play an important role for adapting and applying new complex knowledge to increase productivity. In addition to this, the nature of work tasks performed by labour in KIBS firms incorporates a spatial perspective, and we assume that regional agglomeration positively affects firm productivity in

KIBS firms

3. Method, data and model formulation

This paper uses longitudinal data from Statistics Sweden (SCB) that includes all

Swedish firms for the period between 2001 to 2008.

6

The data comprise detailed information about firms and employees and allow linking the two together. For two main reasons, the analysis only considers all singleestablishment firms. The first is that multi-establishment firms could cause inaccuracy in the productivity measures since accounting data can be rearranged across establishments within a firm. The second reason is the assumed higher dependence of single-establishment firms’s on location-specific characteristics.

So, by using only firms with one establishment, the data are homogenised in such a way that the data for this specific sub-populations can be analysed in a satisfying way. The share of drop-outs from data is around 2 per cent. For obvious reasons, we only include firms established before 2001 that have been in business throughout the whole period. Individuals without a specific occupation or a specific work-place and firms that cannot be linked to a specific industry are also excluded from the analysis. As a consequence, the share of drop-outs by doing this is 6 per cent for the firms and 6 per cent for the employees.

We are interested in firm productivity and choose to define it as value added per employee. This is a valid proxy of firm productivity due to the close

6

Data is provided by Statistics Sweden. This is not a public data base and has restricted access.

154

Knowledge-intensive business services, creative labour inflow and firm productivity

association with real industrial output (Kuznets, 1960). The analysis follows two alternative strategies in an attempt to capture the impact of the incoming creative labour in KIBS and non-KIBS firms, respectively. Thus data processing follows two main strategies. The first is to highlight occupational categories and to distinguish between creative labour inflows and general labour inflows. This is an attempt to see the effects of incoming labour embodying new knowledge and with a high capacity to work with complex problem solution. The second strategy is to split the sample into KIBS and non-KIBS firms. A large part of the activities in KIBS firms is consultancy activities. By defining the KIBS sector using industry codes (NACE 72-74), the data reveal the main production activities of the firms.

Equation 1 illustrates the estimated model where

,

is the value added per employee in firm j at time t. The amount of new creative labour in firm j between time t and t-1 is represented by

,

,

.

In a similar way, is the total number of incoming labour. The model also has a vector of firm attributes

,

,

and a vector of regional attributes

,

.

,

= + ∆

,

+ ∆

,

+

,

+

,

+

,

(1)

, where t = 2002,…....,2008

Labour inflow and creative labour inflow are defined as incoming labour into firms between two years. As a first step, all individuals changing jobs between two years (e.g. 2007 and 2008) are distinguished. The distribution of these individuals across all firms in Sweden in 2008 is thereafter discerned. That means that one can see where these individual worked before, what type of work tasks they performed at their previous employment and whether this employment was in the same region or in a different region. The significance of incoming labour with higher education is also controlled for in the analysis.

For firms to adapt to new conditions, they need the ability to renew their production base and to adjust their production processes and supply of products. Incoming labour can have similar, related or unrelated knowledge, given what already exists in each firm. In terms of geographical distances, labour inflows can be described in a number of ways. Some individuals come from the same functional region (Local labour market region, LLM), and are defined as intra-regional movers. Other individuals come from other functional regions and are defined as inter-regional movers. Intra-regional and inter-regional labour flows can complement or substitute current knowledge, in the same

155

Jönköping International Business School

manner as between-firm mobility (Boschma, Eriksson, & Lindgren, 2009).

However, even using a very fine industry classification, one can expect to find a variety of occupations that are more or less related to each other. Some are quite similar, others are closely or distantly related, while some might be completely unrelated.

Against this background, it is possible to divide the new work tasks of incoming individuals into similar, related, and unrelated in relation to their earlier work tasks. Individuals are classified as belonging to a similar inflow if they have the same occupation (3-digit SSYK) in their new job (e.g. in 2008) as in their previous job (e.g. in 2007). This implies that they move from one firm to another firm but with very similar work tasks. Moving individuals are defined as belonging to the group related inflow if they keep the same 1-digit SSYK code but are not classified as belonging to the group similar inflow. Thus, they have one type of work task in their employment in 2007 and when they changed firms they also changed work tasks, but changed them into something related to what they did in their previous job. Unrelated inflows are all other inflows.

This differentiation allows us to classify incoming labour and analyse how different inflows affect firm productivity.

As a consequence, the present empirical analysis uses six control variables to assess differentiated impacts of labour inflows with respect to their occupations at their previous job and the location of their previous job. Three of these are for intra-regional inflows, intra-regional similar, intra-regional related and intra-regional

unrelated and three for inter-regional inflows, inter-regional similar, inter-regional

related and inter-regional unrelated. By using occupations rather than industry codes, we hopefully have a more accurate description of what type of knowledge incoming labour brings to the new employment and to what degree they contribute to the productivity of their new employer.

Firm age and firm size are included to control for firm characteristics besides the characteristics of the incoming labour flows. Firms in the KIBS sector are oriented towards customised supply, close relations with customers and fast production adjustments. One may assume that age can encumber such fast adjustments, thereby causing inertia (Marshall, 1890/1920). However, for firms with a more standardised production, the opposite has been shown to be true. Larger firms perform better than smaller firms and this is very much related to market power (Shepherd, 1986).

Regional variables control for characteristics in the regional economic milieu that may affect firm productivity. The regional share of producer services (non-

KIBS), the regional share of highly educated labour, and the regional

156

Knowledge-intensive business services, creative labour inflow and firm productivity

distribution of occupations are incorporated into the model to control for possible regional externalities. The 81 Swedish functional regions used in this study are equivalent to local labour market regions (LLMs) (Nutek, 1998). The

LLMs are characterised by a high intensity of intra-regional commuting flows and are delineated based on the intensity of observed commuting flows within and between municipalities (urban regions). The approximated average car travel time distance is 20 to 30 minutes within the LLMs, and the car travel time between two locations in an LLM rarely exceeds 50 minutes (Johansson et al., 2002). Travel intensity declines sharply at the borders of the LLMs, and one can argue that the LLMs are arenas for a high intensity of face-to-face interactions (Andersson & Karlsson, 2006). Table 1 presents the explanatory variables with a motivation for their inclusion and their expected effect on firm productivity.

157

Jönköping International Business School

Table 1 Variables: Firm productivity j, inflow characteristics, characteristics of region r, and characteristics of firm j

Variable name Definition Motivation (expected outcome)

Firm productivity

Inflow characteristics

t,j

∆ All labour inflow t,j

Value added per employee in firm j Regressand, controlling for industry variations

Amount of incoming labour into firm j

General knowledge inflow (-)

∆ Creative inflow t,j

∆ Education inflow

Intra-regional similar occupation

t,j t,j

Amount of incoming creative labour into firm j

Amount of incoming labour with ≥ 3 years of university education into firm j

Amount of inflow from same 3-digit SSYK into firm j

Creative knowledge inflow as a vehicle for knowledge absorption and firm productivity (+)

Control for overlapping effects with creative labour inflow (+)

Intra-regional related occupation

t,j

Intra-regional

Inter-regional similar occupation

t,j

unrelated occupation

t,j

Amount of inflow from same

1-digit SSYK (except similar) into firm j

Amount of inflow from all other 3-digit SSYK into firm j

Amount of inflow from same 3-digit SSYK into firm j

Labour inflow with recognised knowledge, which does not add to existing knowledge base (-)

Labour inflow bringing complementary

Knowledge which adds to existing knowledge base (+)

Labour inflow bringing knowledge, neither recognised nor complementary to the existing knowledge base (-)

Labour inflow bringing no new knowledge and not adding to existing knowledge (-)

Inter-regional related occupation

t,j

Inter-regional unrelated occupation

t,j

Amount of inflow from same 1-digit SSYK

(except similar) into firm j

Amount of inflow from all other 3-digit SSYK into firm j

Labour inflow bringing complementary knowledge and adding to existing knowledge (+)

Labour inflow bringing knowledge which is neither recognised nor complementary to existing knowledge base (-)

Firm characteristics

Age

t,j

Time in business (+/-)

Firm size

t,j

Regional economic milieu

Age of firm j at t

i

: 1: if > 20 years,

0: otherwise

Number of employees in firm j Internal economies of scale (+)

Proxy for regional agglomeration and urbanisation (+)

Regional presence of subsidiary services, spillover effects (+)

Regional size

t,r

Supply intensity

t,r

Total number of employees in region r

Regional employment in

NACE 65-67, 80-85, 90-93 as a fraction of total employment region r

Knowledge intensity

t,r

Share of regional population with ≥3 years of university

Creative intensity

t,r

education region r

Share of regional population with a creative occupation region r

General spillover effects, human capital externalities (+)

Spillover effects

158

Knowledge-intensive business services, creative labour inflow and firm productivity

The descriptive statistics for firm productivity and the explanatory variables for

KIBS firms and non-KIBS firms are displayed in Tables 2 and 3, respectively.

The descriptive analysis shows quite clearly that there are clear differences between KIBS firms and non-KIBS firms. Firms in the KIBS sector are on average smaller; the largest firm in the non-KIBS sector is about three and a half times larger than the largest firm in the KIBS sector. However, KIBS firms have a higher firm productivity. Non-KIBS firms have higher average labour inflows but lower average inflows of creative and high-skilled labour, which is in line with previous findings that knowledge-intensive business services largely involve creative work tasks and rely on high-skilled labour. Non-KIBS firms are also on average larger. Considering the regional economic milieu, there are also differences between the two categories worth mentioning. Looking at overall mean of regional size, KIBS firms tend to be located in larger regions with an even larger share of subsidiary supply services and knowledge labour. They are also located in regions with a larger share of creative labour (compared with non-KIBS firms) but, surprisingly, the difference between the two is not extremely large in this respect. This can indicate that firms in the KIBS sector with a large share of creative labour are dependent on the regional economic milieu but not necessarily reliant on being close to firms in the same sector.

Table 2 KIBS descriptive, N= 125 258, n (number of KIBS firms)= 17 894, T (years)=7 †

mean

Overall std.dev between

Std.dev within

Overall minimum

Overall maximum

Firm productivity 574.22 1984.64 1779.52 878.76 0.50 395

Inflow characteristics

All inflow

Creative inflow

Education inflow

Firm characteristics

Firm age (binary)

Firm size

Regional economic

milieu

Regional size

Supply intensity

1.14 6.94 4.56 5.24 0 876

0.75 5.59 3.50 4.37 0 828

0.41 3.15 2.13 2.32 0 393

0.09

4.24

0.28

14.79

*

14.53

*

2.77

0

1

1

935

53 872 132 284 122 274 57 641 1 109 1 076 752

0.83 0.05 0.05 0.01 0.70 0.98

Knowledge intensity 0.15 0.07 0.06 0.01 0.01 0.24

Regional creativity 0.32 0.09 0.09 0.01 0.11 0.43

Between: variation of means of each firm across time

Within: variation of deviation from respective mean to each firm

159

Jönköping International Business School

Table 3 Non-KIBS descriptive, N= 670 943, n (number of non-KIBS firms) = 95 849, T

(years)=7 †

Firm productivity mean

Overall std.dev

Std.dev between

Std.dev within

Overall minimum

Overall maximum

453.13 591.06 498.72 317.21 2.14 58

Inflow characteristics

All inflow

Creative inflow

Education inflow

Firm characteristics

Firm age (binary)

Firm size

Regional milieu

Regional size

1.61 11.03 6.62 8.82 0

0.34 3.79 2.09 3.16 0

3

1

0.15 3.00 2.13 2.11 0 1

0.24

6.35

25 364

0.43

27.07

*

26.83

*

3.60

119 760 112 374 41 415

0

1

1 109

1

3 345

1 076 752

Supply intensity 0.74 0.26 0.26 0.01 0.02 0.98

Knowledge intensity 0.12 0.06 0.06 0.01 0.01 0.24

Regional creativity 0.28 0.08 0.08 0.01 0.11 0.43

Between: variation of means of each firm across time,

Within: variation of deviation from respective mean to each firm

4. Empirical results

Table 4 displays the results from the pooled OLS estimations with clusterrobust standard errors where log values are used to reduce the effects of skewed distributions. The first three columns examine the effects of creative labour inflow into KIBS firms and the final three columns consider only non-

KIBS firms. The central variables in the analysis are the variables reflecting labour mobility and, specifically, the one measuring the creative labour inflows.

160

Knowledge-intensive business services, creative labour inflow and firm productivity

Table 4 Pooled OLS estimations with cluster robust standard errors, dependent variable

Firm productivity

t,j

(z-values in italics and parentheses)

OLS

7 years and 17 894 firms

Pooled OLS (year 2002 to year 2008)

7 years and 95 848 firms

Non-

KIBS

Non-

KIBS

Non-

KIBS

Inflow characteristics

∆ All labour inflow

t,j

∆ Education inflow

t,j

-0.016***

(-6.40)

∆ Creative inflow

t,j

0.018***

(9.58)

-0.011***

(-7.45)

0.015***

(8.93)

-0.010***

(-9.94)

0.026***

(7.41)

-0.014***

(-8.27)

0.011***

(6.48)

-0.003***

(-4.14)

0.009***

(4.77)

-0.002***

(-5.67)

0.005**

(2.28)

Intra-regional similar occupation

t,j

Intra-regional related occupation

t,j

Intra-regional unrelated occupation

t,j

Inter-regional similar occupation

t,j

Inter-regional related occupation

t,j

Inter-regional unrelated occupation

t,j

Firm characteristics

-0.049***

(-3.67)

0.056***

(4.41)

-0.020***

(-4.30)

-0.066*

(-1.60)

0.085**

(2.11)

-0.024***

(-5.11)

-0.025***

(-5.43)

0.038***

(7.94)

-0.010***

(-5.09)

-0.079***

(-4.61)

0.082***

(4.68)

-0.028***

(-8.03)

0.271***

(6.11)

0.155***

(62.71)

0.026***

(5.99)

0.164***

(50.17)

0.027***

(6.16)

0.146***

(84.41)

Age

t,j (>20 years=1; 0 otherwise)

Firm size

t,j

(ln)

Regional milieu

-0.027*

(-1.78)

0.109***

(24.13)

-0.028*

(-1.83)

0.115***

(23.13)

-0.024*

(-1.61)

0.101***

(22.34)

Regional size

t,r

(ln)

Supply

Intensity

t,r

(ln)

Knowledge

Intensity

t,r

(ln)

0.044***

(14.28)

0.05***

(14.48)

0.66***

(13.09)

Creative region

t,r

(ln) 0.289***

(11.88)

(Constant)

F-value

Prob>F

5.50

(159.14)

172.64

(0.000)

R

2

0.044***

(14.32)

(14.50)

(13.13)

(11.96)

5.56

111.51

(0.000)

0.046***

0.659***

0.291***

(156.16)

0.297 0.188 0.115

0.044***

(14.08)

0.046***

(14.26)

0.647***

(12.85)

0.284***

(11.64)

5.58

(157.62)

178.82

(0.000)

No. of obs 125 258 125 258 125 258

0.027***

(6.11)

0.05***

(37.77)

0.282***

(12.13)

0.093***

(8.18)

5.67

(159.18)

1097.43

(0.000)

0.352

670 943

0.018***

(13.68)

0.048***

(37.43)

0.280***

(12.03)

0.095***

(8.31)

5.67

(158.19)

1020.25

(0.000)

0.256

670 943

0.019***

(13.98)

0.048***

(38.17)

0.285***

(12.21)

0.095***

(8.27)

5.68

(156.64)

1626.11

(0.000)

0.185

670 943

*0.1, **0.05, ***0.001

Regional variables are included separately due to multicollinearity but all coefficients are robust. Coefficients in the table are the ones estimated with Regional size.

‡ Yearly least square estimations generate coefficients with only minor variations over time supporting a pooled estimation technique. The estimation has also been performed excluding the year 2008 to control for effects of the financial crisis, but estimations generate robust results. A fixed effect model has also been estimated for all firms and the seven years with robust results. Modelling it, controlling for time effects, general labour inflow becomes insignificant but remaining variables are robust. See Table A3 in the Appendix.

161

Jönköping International Business School

The results in Table 4 show that firm productivity is influenced by what work tasks the incoming labour performs in the new employment and how these work tasks are related to previous occupations. However, the effects vary between intra-regional and inter-regional inflows. F-statistics and their probabilities are displayed at the bottom of Table 4 and all probabilities have a very low value, indicating that at least some of the parameters are non-zero and validate that the model equation is fitting the data. The R

2 values are at of acceptable levels between 0.185 and 0.352 for non-KIBS firms and between

0.115 and 0.297 for KIBS firms. The first model for both firm categories has the largest explanatory power. One must here bear in mind that the number of observations for non-KIBS firms is much larger than for KIBS firms. Below follows a detailed discussion of the results concerning how the different explanatory factors that affect firm productivity starting with inflow variables.

7

4.1 Labour inflow

Table 4 presents estimated coefficients and standard scores, (z-values) to deal with the difficulties in interpreting variables with different scaling. Labour inflow has been categorised into total labour inflow, creative labour inflow and labour inflow with a three-year or longer education. The former two were estimated together while the latter was estimated as a substitute variable to creative labour inflow. Making such a distinction gives us an indication whether creative labour inflows have stronger influence than general labour inflow or inflow of educated labour. Tables A4 and A5 in the Appendix present the correlation matrices for KIBS and non-KIBS firms, respectively. There one can see that incoming creative labour into KIBS firms largely correlates with incoming labour with higher education. This pattern does not appear in non-

KIBS firms, where the correlation coefficients between creative labour inflow and high education labour inflow is only 0.055.

The labour inflow is further sub-divided to examine whether there are any significant differences in effects of labour coming from the same region, intra-

regional inflow and labour coming from another region, inter-regional inflow. To examine possible differences in absorptive capacities at the firm level, these intra-regional and inter-regional labour inflows are also split into individuals performing similar or related work tasks at their old and new jobs and individuals doing work tasks at their new job, which are completely unrelated to their

7 Two alternative measures of firm performance has also been applied: (i) sales per employee, and (ii) wage in industry and region per employee. No major discrepancies were found compared to firm productivity.

162

Knowledge-intensive business services, creative labour inflow and firm productivity

previous employment. This is a distinct feature of the present analysis. By focusing on occupational relatedness instead of industry relatedness we capture what people do with their skills since, even at a very fine industrial classification, there are expected variations in the distribution of occupations within firms. Due to collinearity, intra-regional and inter-regional inflows are estimated separately. Due to limited space and only very small differences in effects, they are not included in the estimations when we test the effect of inflow of labour with higher education on firm productivity.

Turning to detailed results, general labour inflow has a negative effect on firm productivity in KIBS firms as well as non-KIBS firms. These results are robust throughout the alternative models both when differentiating incoming between types of inflow but also when controlling for regional- and firm characteristics. By calculating the point elasticities one can see that the negative effect is larger for non-KIBS firms (-0.023) than for KIBS firms (-0.0018).

8

The main argument for this is that the general labour inflow increases the labour stock but incoming labour does not add to the existing knowledge base and thus has a negative effect on firm productivity. A reservation should be mentioned, due to that large labour inflows can indicate some type of firm reorganization and to see any positive effects one needs to differentiate between types of labour inflows. From this, the following conclusion can be drawn:

General labour inflow has a negative productivity effect for all types of firms but the negative

effect is smaller for KIBS firms than for non-KIBS firms.

The above results are in line with the results by Boschma et al. (2009) concerning firm productivity effects of labour inflows into Swedish firms.

Whereas the general labour inflow empirically has been examined before, the novel results of the present paper mostly concern creative labour inflows. The results in Table 4 indicate that the creative labour inflow has a positive effect on firm productivity in the KIBS sector. A majority of occupations in the KIBS sector are associated with consultancy and related work tasks. These have been shown to rely on creativity, and a renewal of the existing knowledge base of

KIBS firms is critical to preserve their competitiveness. This can only be managed by a steady inflow of creative labour to these firms. In a more explicit way this means that knowledge and skills embodied in individuals with creative occupations have a larger impact on firm productivity in the KIBS sector than on firm productivity in the non-KIBS sector. While the creative labour inflow

8

Point elasticities and marginal effects for log-linear models are performed as respectively.

163

and

Jönköping International Business School

has a positive effect on firm productivity in KIBS firms, it is also positive for other firms. However, the positive effect is larger for KIBS firms with a point elasticity of 0.0135 compared with non-KIBS firms where the elasticity is almost 30 per cent lower (0.00374). Thus, we draw the following conclusion;

Creative labour inflow has a positive effect on firm productivity in all types of firms but the

effects is larger for KIBS firms than for non-KIBS firms.

The above conclusion is derived from two main circumstances. The first is the heavy dependence on creative labour in the knowledge-intensive business sector. For KIBS firms, this creative labour inflow is also strongly correlated with the inflow of labour inflow with higher education. Following the rich literature explaining the role of knowledge for firm performance, these results are not new. However, for non-KIBS firms creative labour inflow does not have such large overlap with the inflow of labour with higher education, indicating that the creative labour inflow adds something to the existing knowledge base, which cannot be explained by education only. When rerunning the model again, substituting the creative labour inflow with inflows of labour with higher education, the results are relatively robust. The effect is again stronger for KIBS firms than for non-KIBS firms. Thus, some of the effects of the general labour inflow are captured by the inflow of educated labour. The size of the z-values is consistently larger for those models controlling for creative labour inflows, which strengthens this argument further (9.58 and 8.93 for KIBS firms, 6.48 and 4.77 for non-KIBS firms).

We now turn to the six variables designed to catch the effects of labour inflows with varying occupational similarities and with different spatial origin.

The two types of firms accompany each other in terms of the direction of effects and significance. However, the effects are consistently larger for KIBS firms than for non-KIBS firms. Matching prior expectations, incoming labour with very similar work tasks has a significant negative effect on firm productivity irrespective of whether this similar labour comes from the same region or from another region. Incoming labour with related occupational work tasks has a significant positive impact on productivity while that with unrelated occupational work tasks has a significant negative effect. This can be compared with studies showing that local absorptive capacities are needed to transform external knowledge into something that generates value added or economic growth (Boschma, Eriksson, & Lindgren, 2009; Boschma & Iammarino, 2009).

By only adding similar knowledge, the extra knowledge does not add to the existing knowledge stock in such a way that it can be adopted and transformed into something contributing to productivity at the firm level. In contrast to this,

164

Knowledge-intensive business services, creative labour inflow and firm productivity

related knowledge is likely to be absorbed and be transformed into something that improves firm productivity. The z-value value related to this variables is consistently also consistently higher than those for similar and unrelated work tasks. Knowledge that is completely unrelated would theoretically add to the existing knowledge base but it is too different to be absorbed and adopted into something useful for the firm. As already mentioned, the direction of the effects is the same irrespective of whether labour comes intra-regionally or inter-regionally but the z-values are higher for intra-regional similar inflow than

inter-regional similar inflow and are also higher for intra-regional related inflow than for inter-regional related inflow. The opposite appears for inter-regional unrelated (z- value of -5.11), which is higher than intra-regional unrelated (z-value of -4.30).

Hence, the results clearly show neither that there is a lock-in effect of incoming labour from the same region nor that incoming labour from other regions is more difficult to absorb. This can be summarised with the following conclusion:

9

The effects of general labour inflows into KIBS firms and non-KIBS firms are dependent on spatial scale but also on the degree of similarity of occupational work tasks at the old and new

job.

4.2 Variables controlling for firm characteristics and regional economic milieu

Plant size has by far the largest effect on firm productivity, which is no surprising result. This effect is largest for non-KIBS firms. Descriptive statistics in Tables 2 and 3 reveal that KIBS firms are smaller on average and are presumably characterized by lower efficient scale economies. Knowledgeintensive business services are to a lesser extent focused on standardised production and tend to have a customised supply of services from firms with a close relation to their clients. Hence, size matters less for KIBS firms than for non-KIBS firms. Firm age has a negative effect on firm productivity in KIBS firms but a positive effect in non-KIBS firms. The at the correlation matrices in

Tables A6 and A7 show that the bivariate correlation between age and size is not that large. By focusing on firm characteristics only, the following conclusion can be drawn:

9

Corresponding models have been estimated, substituting the creative labour inflow in favour of high education labour inflow and the results are robust and follow those of the model presented here.

165

Jönköping International Business School

Larger firms have higher productivity than smaller firms do. While older non-KIBS firms have higher productivity than younger non-KIBS firms, older KIBS firms have lower

productivity than younger KIBS firms.

The final group of variables comprises those associated with the regional economic milieu. The assumption is that firms get economic advantages of being located in regions offering a high potential for knowledge externalities spurred by agglomeration. Incoming labour brings new knowledge, which is easier to absorb and adapt in regional milieus with potential knowledge externalities, and this should be particularly true for firms with a creative labour inflow.

As one would expect, the group of variables related to firm characteristics shows results that are highly robust within the group of KIBS firms and the group of non-KIBS firms. However, the results differ between the two groups of firms. Related variety indicates whether there are real learning opportunities within firms. We can envisage that the employees use the same technical language applied to different but related work tasks, which can stimulate learning processes within firms. Unrelated varieties may not stop knowledge diffusion but, contrary to related variety, it may hamper the pace of it and can cause problems of understanding and interpretation of knowledge within firms.

As expected, related variety is significantly positive for non-KIBS firms as well as for KIBS firms and unrelated variety affects firm productivity in a significant negative way. Point-elasticities of these variables show that the positive effect is larger for firms outside the KIBS sector than for firms in the

KIBS sector. Contrary to this, the negative effect of unrelated variety on firm productivity is larger for KIBS firms than for non-KIBS firms. Therefore, the results can illustrate potential problems that may be caused by heterogeneity within firms. Firms with a too differentiated stock of knowledge do not create an optimal milieu for knowledge sharing and knowledge absorption. The variable measuring firm size is significantly positive and has a larger effect on the productivity of KIBS firms than on the productivity of non-KIBS firms.

The final group of variables capture the influence of the local economic milieu and characteristics of the functional region. Regional size positively affects productivity in KIBS firms. Supply intensity has constantly the largest z- value, ranging from 14.26 to 14.50 for KIBS firms and from 37.43 to 37.77 for non-KIBS firms. For non-KIBS firms this is a higher value than the other variables describing the region. The coefficients of regional supply are also consistently higher for non-KIBS firms than for KIBS firms, though the differences are very small.

166

Knowledge-intensive business services, creative labour inflow and firm productivity

For the other three regional variables, the size of coefficients is larger for

KIBS firms compared with those for non-KIBS firms. The positive effect of variables describing regional size meets prior expectations, i.e., being located in a large region is ceteris paribus more important for KIBS firms than for non-

KIBS firms. The model has also been estimated controlling for the three largest functional regions in Sweden, Stockholm, Göteborg and Malmö but with the same results for regional size. Firms in the KIBS sector largely produce distance-sensitive products, and larger regions offer better matching between customers and suppliers of such services. Information technology and social media can provide some communication opportunities but cannot replace the knowledge diffusion through face-to-face contacts (Andersson & Beckmann,

2009). This is specifically true for knowledge associated with complex problem solving and embodied skills, which is highly relevant for KIBS firms. Being located in a creative region is more important for KIBS firms than for non-

KIBS firms. From this we can conclude the following:

KIBS firms as well as non-KIBS firms are positively affected by regional size but the effect is

larger for KIBS firms than for non-KIBS firms.

5. Conclusions and suggestions for future research

This paper has examined how labour inflow in general and creative labour inflow in particular affects productivity, measured as value added per employee in KIBS firms and non-KIBS firms, respectively. It argues that creative labour inflow, defined as changes in labour stock between the years 2001 and 2008, matters for firm productivity and that KIBS firms are specifically dependent upon such incoming knowledge. Generally speaking the findings suggest that labour inflow per se does not add to the existing knowledge base in firms and thus does not contribute to firm productivity. However, creative labour inflow has a positive effect for both types of firms, but the effect is significantly stronger for firms in the KIBS sector than for firms outside the KIBS sector.

While previous research has focused on labour inflow, this paper, by subdividing it into degrees of relatedness and spatial origin of incoming labour in general, has its focus on a sector that has shown a remarkable development during post-industrialism and rapid technological development and a fast growth of the knowledge-intensive business service sector. People employed in

KIBS firms are particularly good at solving complex problems based on new

167

Jönköping International Business School

knowledge, fast adaption and non-standardised solutions. One way to define these individuals is to classify them as creative individuals in line with Florida

(2002c). This group of individuals have been shown to dominate in firms in the knowledge-intensive business service sector.

To understand in detail the role of the creative labour inflow, the data were was split between non-KIBS and KIBS firms. The analysis was conducted as a pooled OLS for the period 2001-2008 and in addition to analysing the effects of the general labour inflow and the creative labour inflow, we also controll for intra-regional and inter-regional labour flows. Such a sub-division partly builds on previous work, but the present analysis focuses on occupational relatedness rather than industrial relatedness and this is a distinct and important new feature of the analysis in this paper.

There were only small differences between intra-regional and inter-regional labour inflows and also small differences between KIBS and non-KIBS firms in this respect. However, the results strongly suggest that there are significant differences in absorption capacity when incoming labour applies similar skills,

related skills or unrelated skills respectively. Similar work tasks in the previous and the new employment add nothing to the existing knowledge base but the contrary is true in the case of labour inflows with related skills. Unrelated skills do not seem to be absorbed into something that contributes to firm productivity. As control variables, firm size and firm age were included and the results follow prior expectations.

KIBS firms are smaller than non-KIBS firms and their productivity is negatively affected by firm age. All variables capturing the regional milieu are highly robust for KIBS and non-KIBs firms but the size of the effects vary between the two types of firms. Productivity in firms categorised as KIBS firms are more dependent upon regional size per se, and the importance of subsidiary services is larger than for non-KIBS firms.

To control for the possible similar effects of creative labour inflow and inflow of labour with higher education, a model variation was estimated. When substituting creative labour inflow inflow of labour with higher education, the results are robust for non-KIBS firms. However, there is still much room for advancements in this area of research and one suggestion is to analyse the factors determining the productivity of KIBS firms further. They differ from non-KIBS firms in terms relation with and dependence upon the regional economic milieu. It would therefore be of high interest to analyse more in detail how location of KIBS firms and how productivity is affected by differences in locations.

168

Knowledge-intensive business services, creative labour inflow and firm productivity

References

Abowd, J. M., Haltiwanger, J., McCue, K., McKinney, K., & Sandusky, K. (2005). The relation among human capital, productivity and market value- building up from micro evidence. In C. Corrado, J. Haltiwanger & D. Sichel (Eds.), Measuring Capital

in the New Economy (pp. 153-198). Chicago and London: University of Chicago Press.

Almeida, P., & Kogut, B. (1999). Localization of Engineers in Regional Networks.

Management Science, 45(7), 905-917.

Amabile, T. M. (1983). The Social Psychology of Creativity. New York, NY: Springer-Verlag.

Andersson, M., & Hellerstedt, K. (2009). Location Attributes and Start-ups in

Knowledge-Intensive Business Services. Industry & Innovation, 16(1), 103 - 121.

Andersson, M., & Karlsson, C. (2006). Regional Innovation Systems in Small and

Medium-Sized Regions: A Critical Review and Assessment. In B. Johansson, C.

Karlsson & R. Stough (Eds.), The Emerging Digital Economy: Entrepreneurship, Clusters,

and Policy (pp. 55-82). Berlin: Springer-Verlag.

Andersson, M., & Thulin, P. (2008). Globalisering, arbetskraftens rörlighet och produktivitet (in Swedish) (Vol. Underlagsrapport ). Stockholm: Globaliseringsrådet

Andersson, Å. E. (1985). Kreativitet- storstadens framtid: En bok om Stokholm (in Swedish).

Stockholm: Prisma.

Andersson, Å. E., & Beckmann, M. J. (2009). Economics of Knowledge: Theory, Models and

Measurements. Cheltenham, UK: Edward Elgar.

Angel, D. (1989). The Labor Market for Engineers in the US Seminconductor Industry.

Economic Geography, 65(2), 99-112.

Bettencourt, L. A., Ostron, A. L., & Brown, R. I. (2002). Client co-production in knowledge-intensive business services. California Management Review, 44(4), 100-128.

Boschma, R., Eriksson, R., & Lindgren, U. (2009). How does labour mobility affect the performance of plants? The importance of relatedness and geographical proximity.

Journal of Economic Geography, 9(2), 169-190.

Boschma, R., & Iammarino, S. (2009). Related Variety, Trade Linkages, and Regional

Growth in Italy. Economic Geography, 85(3), 289-311.

Braunerhjelm, P. (2007). Behövs entreprenörerna? Om deras betydelse för kunskapsutveckling, kommersialisering och tillväxt (in Swedish). Ekonomisk Debatt,

35 (6), 19-33.

Bresnahan, T., Gambardella, A., & Saxenian, A. (2001). ‘Old Economy’ Inputs for

‘New Economy’ Outcomes: Cluster Formation in the New Silicon Valleys. Industrial

and Corporate Change, 10(4), 835-860.

Cooper, D. P. (2001). Innovation and reciprocal externalities: Information transmission via job mobility. Journal of Economic Behavior & Organization, 45(4), 403-425.

Czarnitzki, D., & Spielkamp, A. (2003). Business services in Germany: bridges for innovation. The Service Industries Journal, 23(2), 1 - 30.

Dahl, M. S., & Sorenson, O. (2007). Home sweet home? Social capital and location decisions.

Paper presented at the Academy of Management Annual Meeting, Philadelphia, PA. den Hertog, P. (2000). Knowledge-intensive business services as co-producers of innovation. International Journal of Innovation Management, 4(4), 491-528.

Duranton, G., & Puga, D. (2004). Micro-foundations of Urban Agglomeration

Economies. In J. V. Henderson & J.-F. Thisse (Eds.), Handbook of Regional and Urban

Economics. Vol. 4: Cities and Geography (pp. 2063-2117). Amsterdam: Elsevier.

169

Jönköping International Business School

EMCC, (2005). Sector Futures: the knowledge-intensive business services sector. In E.

M. C. o. Change (Ed.). Dublin: European Foundation for the Improvement of

Living and Working Conditions.

Feldman, M. P. (2000). Location and Innovation: The New Economic Geography of

Innovation, Spillovers and Agglomeration. In G. L. Clark, M. P. Feldman & M. S.

Gertler (Eds.), The Oxford Handbook of Economic Geography (pp. 373-394). Oxford:

Oxford University Press.

Florida, R. (2002a). Bohemia and economic geography. Journal of Economic Geography.

Florida, R. (2002b). The Economic Geography of Talent. Annals of the Association of

American Geographers, 92(4), 743-755.

Florida, R. (2002c). The Rise of the Creative Class: and how it's transforming work, leisure,

community & everyday fife. New York, NY: Basic Books.

Florida, R., Mellander, C., & Stolarick, K. (2008). Inside the black box of regional development-human capital, the creative class and tolerance. Journal of Economic

Geography, 8(5), 615-649.

Fox, J. T., & Smeets, V. (2011). Does input quality drive measured differences in firm productivity? International Economic Review, 52(4), 961-989.

Fujita, M., & Thisse, J.-F. (2002). Economics of Agglomeration: Cities, Industrial Location, and

Regional Growth. Cambridge, UK: Cambridge University Press.

Gallouj, F. (2002). Knowledge-intensive business services: processing knowledge and producing innovation. In J. Gadrey & F. Gallouj (Eds.), Productivit, Innovation and

Knowledge in Services (pp. 256-284). Cheltenham, Northampton: Edward Elgar.

Glaeser, E. L. (2005). Review of Richard Florida's The Rise of the Creative Class. Regional

Science and Urban Economics, 35(5), 593-596.

Glaeser, E. L. (2008). Cities, Agglomeration and Spatial Equilibrium. Oxford, UK: Oxford

University Press.

Glaeser, E. L., Kallal, H. D., Scheinkman, J. A., & Shleifer, A. (1992). Growth in Cities.

Journal of Political Economy, 100(6), 1126-1152.

Glaeser, E. L., & Saiz, A. (2004). The Rise of the Skilled City [with Comments by G.T

Burtless & W.S Strange]. Brookings-Wharton Papers on Urban Affairs, 5, 47-105.

Hoyman, M., & Faricy, C. (2009). It Takes a Village: A Test of the Creative Class, Social

Capital, and Human Capital Theories. Urban Affairs Review, 44(3), 311-333.

Hsieh, C.-T., & Klenow, P. J. (2009). Misallocation and Manufacturing TFP in China and India. The Quarterly Journal of Economics, 124(4), 1403-1448.

Iammarino, S., & McCann, P. (2006). The structure and evolution of industrial clusters:

Transactions, technology and knowledge spillovers. Research Policy, 35(7), 1018-1036.

Jacobs, J. (1969). The Economy of Cities. New York, NY: Random House.

Johansson, B., Klaesson, J., & Olsson, M. (2002). Time distances and labor market integration. Papers in Regional Science, 81(3), 305-327.

Johansson, B., & Quigley, J. M. (2004). Agglomeration and Networks in Spatial

Economies Papers in Regional Science, 83(1), 165-176.

Karlsson, C. (2011). Clusters, networks and creativity. In D. E. Andersson, Å. E.

Andersson & C. Mellander (Eds.), Handbook of Creative Cities. Cheltenham, UK and

Northampton, MA: Edward Elgar.

Krugman, P. (1991). Geography and Trade. Cambridge, MA: MIT Press.

Kuznets, S. (1960). Population Change and Aggregate Output. With Comments by R.E

Quandt and M.Freidman. . In J. C. Ansley (Ed.), Demographic and Economic Change in

Developed Countries (pp. 340-367). Princeton, NJ: Princeton University Press.

170

Knowledge-intensive business services, creative labour inflow and firm productivity

Leiponen, A. (2005). Organization of Knowledge and Innovation: The Case of Finnish

Business Services. Industry and Innovation, 12(2), 185-203.

Lundvall, B. A. (1998). The Learning Economy: Challenges to Economic Theory and

Policy. In K. Nielsen & B. Johnson (Eds.), Institutions and economic change.

Cheltenham, UK: Edward Elgar.

Markusen, A. (2006). Urban development and the politics of a creative class: evidence from a study of artists. Environment and Planning A, 38(10), 1921-1940.

Marlet, G., & van Woerkens, C. (2004). Skills and Creativity in a Cross-section of

Dutch Cities Discussion Paper Series 04-29. Utrecht: Tjalling C.Koopmans Reserach

Institute.

Marshall, A. (1890/1920). Principles of Economics. London, UK: Macmillan.

Marshall, J. N., Damesick, P., & Wood, P. (1987). Understanding the location and role of producer services in the United Kingdom. Environment and Planning A, 19(5), 575-

595.

Maskell, P., & Malmberg, A. (1999). Localised learning and industrial competitiveness.

Cambridge Journal of Economics, 23(2), 167-185.

Mellander, C. (2009). Creative and Knowledge Industries: An Occupational

Distribution Approach. Economic Development Quarterly, 23(4), 294-305.

Mellander, C., & Florida, R. (2007). The Creative Class or Human Capital?- Explaining

Regional Development in Sweden Working Paper Series in Economics and Institutions of

Innovation. Stockholm: CESIS, Royal Institute of Technology.

Miles, I. (1999). Services in National Innovation System: From Traditional Services to

Knowledge-Intensive Business Services. In G. Schienstock & O. Kuusi (Eds.),

Transformation Towards a Learning Economy: the Challenge for the Finnish Innovation System

(pp. 57-98). Helsinki.

Miles, I. (2001). Knowledge-intensive business services revisites Nijmegen Lectures on

Innovation Management. Antwerpen: Maklu.

Miles, I. (2003). Services and the Knowledge-Based Economy. In J. Tidd & F. M. Hull

(Eds.), Service Innovation, Organizational Responses to Technological Opportunities & Market

Imperatives (pp. 81-112). London, UK: Imperial College Press.

Miles, I., Kastrinos, N., Flanagan, K., Bilderbeek, R., den Hertog, P., Huntink, W., &

Bouman, M. (1995). Knowledge -Intensive Business Services-Users:Carriers and Sources of

Innovation. Manchester: PREST.

Muller, E., & Zenker, A. (2001). Business services as actors of knowledge transformation: the role of KIBS in regional and national innovation systems.

Research Policy, 30(9), 1501-1516.

Nooteboom, B., Van Haverbeke, W., Duysters, G., Gilsing, V., & van den Oord, A.

(2007). Optimal cognitive distance and absorptive capacity. Research Policy, 36(7),

1016-1034.

Noteboom, B. (2000). Learning and Innovation in Organizations and Economies. Oxford, UK:

Oxford University Press.

Nutek. (1998). Småföretag och regioner i Sverige 1998: med ett tillväxtperspektiv för hela landet Report B 1998:10. Stockholm: Närings- och teknikutvecklingsverket:

Nutek.

O'Mahony, M., Robinson, C., & Vecchi, M. (2008). The impact of ICT on the demand for skilled labour: A cross-country comparison. Labour Economics, 15(6), 1435-1450.

Porter, M. E. (1990). The Competitive Advantage of Nations. New York, NY: Free Press and

London, UK:Macmillan.

171

Jönköping International Business School

Power, D., & Lundmark, M. (2004). Working Through Knowledge Pools: Labour

Market Dynamics, the Transference of Knowledge and Ideas, and Industrial

Clusters. Urban Studies, 41(5/6), 1025-1044.

Runco, M. A. (2004). Everyone Has Creative Potential. In R. J. Sternberg, E. L.

Grigorenko & J. L. Singer (Eds.), Creativity: From Potential to Realization (pp. 21-30).

Washington, DC: American Psychological Association.

Saxenian, A. (1994). Regional Advantage: Culture and Competition in Silicon Valley and Route

128. Cambridge, MA: Harvard University Press.

Shepherd, W. G. (1986). On the Core Concepts of Industrial Economics. In H. W. de

Jong & W. G. Shepherd (Eds.), Mainstreams in Industrial Organization. Dordrecht:

Martinus Nijhoff Publisher.

Smith, G., Carlsson, I., & Danielsson, A. (1984). Experimental Examinations of Creativity.

Lund: Lund University.

Syverson, C. (2004). Product Substitutability and Productivity Dispersion. Review of

Economics and Statistics, 86(2), 534-550.

Syverson, C. (2011). What Determines Productivity? Journal of Economic Literature, 49(2),

326-365.

172

Knowledge-intensive business services, creative labour inflow and firm productivity

Appendix

Table A1 outlines the main statistical sectors that encompass KIBS with industrial classifications codes NACE 72-74.

Table A1 Major KIBS sectors in the empirical analysis

72

72.1 Hardware

72.2

Computer and related activities

Software consultancy and supply

72.3 Data activities

72.5

72.6

73

73.1

73.2

74

Maintenance and repair of office, accounting and computing machinery

Other computer-related activities

Research and development

Research and experimental development on natural science and engineering

Research and development

Other business activities

74.11 Legal

74.12

74.13

74.14

74.15

Accounting, book-keeping and auditing activities; tax consultancy

Market research and public opinion polling

Business and management consultancy activities

Management activities of holding companies

74.2

74.3

74.5

Architectural and engineering activities and related technical consultancy

Technical testing and analysis

74.4 Advertising

Labour recruitment and provision of personnel

74.6 Investigation

74.7 Industrial

74.8 Miscellaneous business activities n.e.c

173

Jönköping International Business School

Table A2 outlines the main statistical sectors that encompasses the creative sector. Occupations in health and education are not included.

Table A2 Codes and names of occupational groups included in the empirical analysis

241

242

244

245

247

248

311

312

313

321

324

342

343

344

347

SSYK Occupational group

111

112

121

122

123

131

211

212

213

214

221

231

412

521

Legislators and senior government officials

Senior officials of special-interest organisations

Directors and chief executives

Production and operations managers

Other specialist managers

Managers of small enterprises

Physicists, chemists and related professionals

Mathematicians and statisticians

Computing professionals

Architects, town and traffic planners

Life science professionals

College, university and higher education teaching professionals

Business professionals

Legal professionals

Economists

Writers and creative or performing artists

Public service administrative professionals

Administrative professionals of special-interest organisations

Physical and engineering science technicians

Computer associate professionals

Optical and electronic equipment operators

Agronomy and forestry technicians

Life science technicians

Business service agents and trade brokers

Administrative associate professionals

Customs and border inspectors

Artistic, entertainment and sports associate professionals

Numerical clerks

Fashion

174

Knowledge-intensive business services, creative labour inflow and firm productivity

Table A3 Fixed effects, dependent variable firm productivity,

113 745 firms, (z-values in italics and parentheses)

Fixed effect

(robust standard errors)

,

, all firms, 7 years and

FE with time effects

(Constant) 2.92***

(99.99)

Inflow characteristics

∆ All labour inflow t,j

-1.86e

-4

**

(-2.39)

∆ Creative inflow t,j

3.57e

-4

**

(2.13)

Firm characteristics

Age

t,j (>20 years=1; 0 otherwise)

Firm size

t,j

(ln)

Regional milieu

Regional size

t,r

(ln)

Rho

5.74***

(40.29)

6.65e

-5

(1.00)

4.06e

-4

(1.91)*

-0.49***

(-153.37)

0.32***

(114.98)

-0.49***

(-154.48)

0.04***

(3.20)

0.87 0.85

Prob>F 0.000 0.000

*0.1, **0.05, ***0.001

175

Jönköping International Business School

Regional size

r

(ln)

Supply intensity

r

(ln)

Knowledge

intensity

r

(ln)

Creative

region

r

(ln)

Table A4 Correlation matrix KIBS: N= 125 258, n (number of KIBS firms)= 17 894, T

(years)=7

P j

All inflow

j

Creative inflow

j

Long education inflow

j

Age

j

Firm size

(ln)

j

Regional size

r

(ln)

Supply intensity

r

(ln)

Knowledge intensity

r

(ln)

Creative region

r

(ln)

P j

1

All inflow

j

Creative inflow

j

Long education inflow

j

Age

j

Firm size

j

(ln)

0.052 1

0.071 0.895

0.078 0.833

-0.000 0.016

0.126 0.397

1

0.915

0.019

0.320

1

0.017

0.331

1

0.085

1

Regional size

r

(ln)

Supply intensity

r

(ln)

Knowledge intensity

r

(ln)

Creative region

r

(ln)

0.086 0.045

-0.033 -0.009

0.080 0.040

0.074 0.022

0.049

-0.012

0.043

0.029

0.059

-0.021

0.055

0.042

-

0.034

0.034

-

0.040

-

0.037

0.006

0.003

0.002

0.005

1

-0.651

0.932

0.925

1

-0.699 1

-0.651 0.973 1

Table A5 Correlation matrix non-KIBS: 670 943, n (number of non-KIBS firms) = 95 849, T

(years)=7

P j

All inflow

j

Creative inflow

j

Long education inflow

j

Age

j

Firm size

j

(ln)

Regional size

r

(ln)

Supply intensity

r

(ln)

Knowledge intensity

r

(ln)

Creative

region

r

(ln)

P j

1

All inflow

j

Creative inflow

Long

j

education inflow

j

Age

j

Firm size

j

(ln)

0.055 1

0.062 0.764

0.031 0.712

0.033 0.027

0.212 0.315

1

0.597

0.020

0.207

1

0.005

0.140

1

0.094

1

0.032 0.024

0.115 0.014

0.025 0.016

0.017 0.003

0.034

0.006

0.029

0.020

0.028

-0.015

0.027

0.023

-

0.084

0.109

-

0.101

-

0.100

-0.014

0.123

-0.025

-0.017

1

-0.416

0.903

0.900

1

-0.479 1

-0.409 0.964 1

176

JIBS Dissertation Series

No. 001 Melander, Anders: Industrial wisdom and strategic change – The

Swedish pulp and paper industry 1945-1990, 1997 (Business

Administration)

No. 002 Marmefelt, Thomas: Bank-industry networks and economic evolution

– An institutional-evolutionary approach, 1998 (Economics)

003 Wiklund, Johan: Small firm growth and performance –

Entrepreneurship and beyond, 1998 (Business Administration)

No. 004 Braunerhjelm, Pontus: Knowledge capital, firm performance and network production, 1999 (Economics),

No. 005 Frankelius, Per: Företagande över tid – Kontextuellt perspektiv och visuellt beskrivningsspråk, 1999 (Business Administration)

No. 006 Klaesson, Johan: A study of localisation economies and the transport sector, 2001 (Economics)

No. 007 Hatemi-J, Abdulnasser: Time-series Econometrics Applied to

Macroeconomic Issues, 2001 (Economics)

No. 008 Alhager, Eleonor: Mervärdesskatt vid omstruktureringar, Iustus förlag AB, 2001 (Commercial Law)

No. 009 Hugoson, Peter: Interregional Business Travel and the Economics of

Business Interaction, 2001 (Economics)

No. 010 Pettersson, Lars: Location, Housing and Premises in a Dynamic

Perspective, 2002 (Economics)

No. 011 Paulsson, Thomas: Decision Making, Risk, and Utility – Assessments and Applications of Alternative Decision Models, 2002 (Economics)

No. 012 Brundin, Ethel: Emotions in Motion – The Strategic Leader in a

Radical Change Process, 2002 (Business Administration)

No. 013 Wiklund, Hans: Arenas for Democratic Deliberation – Decisionmaking in an Infrastructure Project in Sweden, 2002 (Political

Science)

No. 014 Florin Samuelsson, Emilia: Accountability and Family Business

Contexts - An Interpretive Approach to Accounting and Control

Practices, 2002 (Business Administration)

No. 015 Ahl, Helene J.: The Making of the Female Entrepreneur – A

Discourse Analysis of Research Texts on Women’s

Entrepreneurship, 2002 (Business Administration)

177

Jönköping International Business School

No. 016 Olsson, Michael: Studies of Commuting and Labour Market

Integration, 2002 (Economics)

No. 017 Wigren, Caroline: The Spirit of Gnosjö – The Grand Narrative and

Beyond, 2003 (Business Administration)

No. 018 Hall, Annika: Strategising in the context of genuine relations: An interpretative study of strategic renewal through family interactions,

2003 (Business Administration)

No. 019 Nilsson, Ulf: Product costing in interorganizational relationships – A supplier’s perspective, 2003 (Business Administration)

020 Samuelsson, Mikael: Creating new ventures: A longitudinal investigation of the nascent venturing process, 2004 (Business

Administration)

No. 021 Bruns, Volker: Who receives bank loans? A study of lending officers’ assessments of loans to growing small and medium-sized enterprises,

2004 (Business Administration)

No. 022 Gustafsson, Veronica: Entrepreneurial Decision-Making: Individuals, tasks and cognitions, 2004 (Business Administration)

No. 023 Agndal, Henrik: Internationalisation as a Process of Strategy and

Change – A Study of 16 Swedish Industrial SMEs, 2004 (Business

Administration)

No. 024 Ejermo, Olof: Perspectives on Regional and Industrial Dynamics of

Innovation, 2004 (Economics)

No. 025 Barenfeld, Jesper: Taxation of Cross-Border Partnerships: Double-

Tax Relief in Hybrid and Reverse Hybrid Situations, 2005 (Law)

No. 026 Hilling, Maria: Free Movement and Tax Treaties in the Internal

Market, 2005 (Law)

No. 027 Brunninge, Olof: Organisational self-understanding and the strategy process, 2005 (Business Administration)

No. 028 Blombäck, Anna: Supplier brand image – a catalyst for choice:

Expanding the B2B brand discourse by studying the role corporate brand image plays in the selection of subcontractors, 2005 (Business

Administration)

029 Nordqvist, Mattias: Understanding the role of ownership in strategizing: a study of family firms, 2005 (Business Administration)

No. 030 Karlsson, Tomas: Business Plans in New Ventures: An Institutional

Perspective, 2005 (Business Administration)

178

JIBS Dissertation Series

031 Johnson, Andreas: Host Country Effects of Foreign Direct

Investment: The Case of Developing and Transition Economies,

2005 (Economics)

No. 032 Nyström, Kristina: Entry and Exit in Swedish Industrial Sectors, 2006

(Economics)

No. 033 Salvato, Carlo: Micro-Foundations of Organizational Adaptation. A

Field Study in the Evolution of Product Development Capabilities in a Design Firm, 2006 (Business Administration)

No. 034 Gråsjö, Urban: Spatial Spillovers of Knowledge Production – An

Accessibility Approach, 2006 (Economics)

No. 035 Dahlqvist, Jonas: Assessing New Economic Activity – Process and

Performance in New Ventures, 2007 (Business Administration)

No. 036 Andersson, Martin: Disentangling Trade Flows – firms, geography and technology, 2007 (Economics)

No. 037 Nilsson, Désirée: Essays on Trade Flows, Demand Structure and

Income Distribution, 2007 (Economics)

No. 038 McKelvie, Alexander: Innovation in New Firms: Examining the role of knowledge and growth willingness, 2007 (Business Administration)

No. 039 Garvi, Miriam: Venture Capital for the Future - Implications of

Founding Visions in the Venture Capital Setting, 2007 (Business

Administration)

No. 040 Rosander, Ulrika: Generalklausul mot skatteflykt, 2007 (Law)

041 Hultman, Jens: Rethinking adoption – Information and communications technology interaction processes within the Swedish automobile industry, 2007 (Business Administration)

No. 042 Hilling, Axel: Income Taxation of Derivatives and other Financial

Instruments – Economic Substance versus Legal Form: A study focusing on Swedish non-financial companies, 2007 (Law)

Sjölander, Pär: Simulation-Based Approaches in Financial

Econometrics, 2007 (Economics)

No. 044 Hang, Min: Media Business Venturing: A Study on the Choice of

Organizational Mode, 2007 (Business Administration)

No. 045 Lövstål, Eva: Management Control Systems in Entrepreneurial

Organisations – A Balancing Challenge, 2008 (Business

Administration)

No. 046 Fridriksson, Helgi-Valur: Learning processes in an interorganizational context – A study of krAft project, 2008 (Business

Administration)

179

Jönköping International Business School

No. 047 Naldi, Lucia: Growth through Internationalization: a Knowledge

Perspective on SMEs, 2008 (Business Administration)

No. 048 Wiberg, Daniel: Institutional Ownership - the Anonymous Capital:

Corporate Governance and Investment Performance, 2008

(Economics)

No. 049 Eklund Johan E.: Corporate Governance, Private Property and

Investment, 2008 (Economics)

No. 050 Glemdal, Michael: Gubben på kullen - Om den smärtsamma skillnaden mellan politiska intentioner och praktiska resultat, 2008

(Political Science)

No. 051 Bohman, Helena: Trade, Knowledge and Income Distribution, 2008

(Economics)

No. 052 Rendahl Pernilla: Cross-Border Consumption Taxation of Digital

Supplies - A Comparative Study of Double Taxation and

Unintentional Non-Taxation of B2C E-Commerce, 2008 (Law)

No. 053 Mellander, Charlotta: The Wealth of Urban Regions - On the

Location of Creative Individuals and Firms, 2008 (Economics)

No. 054 Johansson, Monica: Organizing Policy - A Policy Analysis starting from SMEs in Tuscany and the County of Jönköping, 2008 (Political

Science)

No. 055 van Weezel, Aldo: Entrepreneurial Strategy-Making Mode and

Performance: A Study of the Newspaper Industry, 2009 (Business

Administration)

No. 056 Hellerstedt, Karin: The Composition of New Venture Teams: Its

Dynamics and Consequences, 2009 (Business Administration)

No. 057 Hunter, Erik: Celebrity Entrepreneurship and Celebrity

Endorsement: Similarities, differences and the effect of deeper engagement, 2009 (Business Administration)

No. 058 Gerson, Anna: Compensation of Losses in Foreign Subsidiaries within the EU: A Comparative Study of the Unilateral Loss-

Compensation Mechanisms in Austria and Denmark, 2009 (Law)

No. 059 Dahlström, Tobias: Causes of corruption, 2009 (Economics)

No. 060 Languilaire, Jean-Charles: Experiencing work/non-work - Theorising individuals' process of integrating and segmenting work, family, social and private, 2009 (Business Administration)

No. 061 Nguyen Tan, Phat: Transfer Pricing: The Vietnamese System in the

Light of the OECD Guidelines and the Systems in certain Developed and Developing Countries, 2009 (Law)

180

JIBS Dissertation Series

No. 062 Karlsson, Ann Britt: Institutionalisering av ansvar i kommunal revision – Lärande organisering, 2009 (Political Science)

No. 063 Johansson, Sara: Knowledge, Product Differentiation and Trade,

2010 (Economics)

No. 064 Ots, Mart: Understanding value formation - A study of marketing communications practices at the food retailer ICA, 2010 (Business

Administration)

No. 065 Raviola, Elena: Paper meets Web: How the institution of news production works on paper and online, 2010 (Business

Administration)

No. 066 Palmberg, Johanna: Family Ownership and Investment Performance,

2010 (Economics)

No. 067 Borgström, Benedikte: Supply chain strategising: Integration in practice, 2010 (Business Administration)

No. 068 Wikner, Sarah: Value co-creation as practice: On a supplier’s capabilities in the value generation process, 2010 (Business

Administration)

No. 069 Karlsson, Peter: Issues of Incompleteness, Outliers and Asymptotics in High-Dimensional Data, 2011, 2011 (Statistics)

No. 070 Helin, Jenny: Living moments in family meetings: A process study in the family business context, 2011 (Business Administration)

No. 071 Markowska, Magdalena: Entrepreneurial Competence Development:

Triggers, Processes & Consequences, 2011 (Business Administration)

No. 072 Cui, Lianguang: Innovation and network development of logistics firms, 2011 (Business Administration)

No. 073 Norbäck, Maria: Making Public Service Television: A study of institutional work in collaborative TV production, 2011 (Business

Administration)

No. 074 Dzansi, James: Essays on Financing and Returns on Investment,

2011 (Economics)

No. 075 Månsson, Kristofer: Issues of multicollinearity and conditional heteroscedasticy in time series econometrics, 2012 (Statistics)

No. 076 Balkow, Jenny: In the Middle: On Sourcing from China and the Role of the Intermediary, 2012 (Business Administration)

No. 077 Karlsson, Hyunjoo Kim: Dynamics of macroeconomic and financial variables in different time horizons, 2012 (Economics)

No. 078 Bjerke, Lina: Knowledge flows across space and firms, 2012

(Economics)

181

Was this manual useful for you? yes no
Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Related manuals

Download PDF

advertisement