Urban concentration and labour market linkages in the Norwegian ICT services sector Accepted for publication in European Planning Studies (2017)
Ingvild Jøranli1 Department of Sociology and Human Geography, University of Oslo E-mail:
[email protected] Phone number: +47 95796290 Address: P.O. Box 1096 Blindern N-0317 Oslo, Norway
Sverre J. Herstad Nordic Institute for Studies in Innovation, Research and Education E-mail:
[email protected] Phone number: +47 91649596 Address: P.O. Box 2815 Tøyen N-0608 Oslo, Norway
Inland Norway University of Applied Sciences E-mail:
[email protected] Address: P.O. Box 952 N-2604 Lillehammer, Norway
Abstract Building on research that emphasizes the dependence of services firms on the networks and experiences of individual employees, this paper investigates the urban concentration of ICT services employment in Norway from the perspective of labour market linkages. It finds that urban regions generally provide firms with access to sector-specific expertise. Beyond this, intrinsic region characteristics determine the position of individual firms in national labour markets for expertise: Firms in the dominant university town have strong contact points to academic labour markets, whereas firms in the industrial stronghold of the Western Capital region exploit a broader range of recruitment channels than firms in any of the other urban and non-urban locations. The results illustrate how capability building through recruitment is influenced by local conditions, and imply that the industry will continue to concentrate in the large-city regions where surrounding organizations provide firms with priveliged access to expertise. Implications for research, innovation policy and societal development more generally are drawn. Keywords: ICT services, urbanization, expertise, recruitment, Norway
1
Corresponding author:
[email protected]
1
Introduction This paper is motivated by three defining characteristics of the contemporary economic landscape. The first is growth in services industries, parallelled by increases in the services content of manufactruring and a downward trend in the direct contribution of this sector to employment (e.g. Tregenna, 2009). The second is urbanization, i.e. industry concentration and population growth favoring large-city regions. The third is rapid development and diffusion of modern information and communication technologies, with pervasive impacts on most aspects of modern life. The ICT services sector represent all these characteristics. It congtributes substantially to gross domestic product (Maryska, Doucek & Kunstova, 2012), and knowledge-intensive employment (e.g. Herstad & Sandven, 2017). Moreover, by being positioned at the intersection between rapidly evolving enabling technologies and the demand for advanced solutions based on them (Miles, 2005; Wood, 2006; Strambach, 2012), it exerts strong influences on innovation, competitiveness and growth in the broader economy. Hence, understanding the interplay between evolutionary and spatial dynamics in the ICT services sector is of fundamental importance to the understanding of the modern economy as whole; and thus to innovation policy making and implementation. Some have viewed ICTs as driving the development towards a ‘flatter’ economic landscape where firms and industries are less tied to specific locations (Friedman, 2005). Still, employment in the services segment of the ICT industry itself has exhibited exceptionally high levels of concentration in large, high-cost and often congested urban locations (Isaksen, 2004). This concentration is resilient, in spite of the growing importance of global innovation networks (Martin, Aslesen, Grillitsch & Herstad, forthcoming 2018) and the current trend of services export and outsourcing to low-cost countries such as India (Dossani & Kenney, 2009; Hermelin, Demir & Verhagen, 2011) that technological and regulatory changes have opened up for (Samiee, 1999; Javalgi, Griffith & White, 2004).
These trends suggest that industry concentration cannot be understood merely through the lenses of inter-firm collaboration in local value chains and innovation networks (Herstad & Ebersberger, 2015), and demand research attention devoted to the forces of localisation 2
arising out of on-the-job learning, knowledge diffusion through labour market mobility (Boschma & Frenken, 2011; Neffke, Henning & Boschma, 2011; Boschma, Eriksson & Lindgren, 2014) and frequent encounters between people in ‘crammed’ geographical spaces (Glaeser, Kallal, Scheinkman & Shleifer, 1992; Storper & Venables, 2004). The relevance of doing so is underscored by the dependence of knowledge intensive services provision on the networks and experiences of employees (Teece, 2003), and contributions that more generally have linked productivity (Boschma, Eriksson & Lindgren, 2009; Timmermans & Boschma, 2014) and innovation (Herstad, Sandven & Ebersberger, 2015; Herstad, 2017) to characteristics of firms’ human resource bases that are inherently shaped by the labour markets of their locations (Østergaard, Timmermans & Kristinsson, 2011; Aarstad, Kvitastein & Jakobsen, 2016; Fitjar & Timmermans, 2016).
Against this background, three inter-related research questions are considered. The first concerns the geographical distribution of employment and growth in the Norwegian ICT services industry during the two distinct periods that are bubble burst (2001-2004), and recovery (2005-2010). The second asks to what extent firms in urban locations have privileged access to different recruitment channels, and thus to valuable expertise. The third asks whether there are fundamental dividing lines between different types of ICT services organizations with respect to their exploitation of, and dependence on, the knowledge dynamics of urban labour markets.
In order to address these questions, a theoretical framework is developed that link research on innovation in knowledge-intensive services firms (KIS) to theories of agglomerations economies (Frenken, Oort & Verburg, 2007) and research on urban knowledge dynamics (McCann, 2007; Shearmur, 2015; Herstad, 2017). The questions are then explored empirically using linked employer-employee data from official Norwegian business and employment registers. The results are first discussed with emphasis on the interplay between evolutionary and spatial dynamics in the ICT services sector. Then, implications for services research, innovation policy and societal development more generally are drawn.
The knowledge dynamics of knowledge intensive services provision Much research on the micro-foundations of corporate competitiveness has emphasized the role of ‘organizational routines’ (Cyert & March, 1963) in transforming technology and 3
specialised individual knowledge into ‘collective’ capabilities that are expressed by firms (Teece, Pisano & Shuen, 1997). Because organisational routines are more difficult to replicate than the human resources and component technologies that they integrate (Wezel, Cattani & Pennings, 2006; Wang, He & Mahoney, 2009), they function as ‘isolating mechanisms’ preventing the firm from being imitated by competitors. In organizational research, concepts such as ‘embedded’ (e.g. Lam, 2000), ‘collective’ (e.g. Nonaka, 1994) and ‘firm-specific’ (e.g. Wang et al., 2009) knowledge have been used to capture the link between knowledge accumulation as a prerequisite for innovation, and the organizational contexts in which processes of knowledge exploration and exploitation occurs. At the aggregate level, theories of technological regimes (Breschi, Malerba & Orsenigo, 2000) use the term ‘Schumpeter Mark II’ to denote industries are particularly dependent on long-term accumulation of specialised skills that are integrated and transformed into capabilities by the organizational processes of firms.
This line of reasoning is strongly influenced by research on advanced manufacturing. The provision of KIS, by contrast, often involves the direct dissemination of knowledge that is embodied in the minds of professional experts (Teece, 2003; von Nordenflycht, 2011). In these cases, individual knowledge is not to the same extent processed by organizations. Instead, it evolves as professionals interact and co-create knowledge directly with clients and partners (Bettencourt, Ostrom, Brown & Roundtree, 2002; Teece, 2003; Dougherty, 2004; Doloreux & Shearmur, 2012), based on experience gained through prior projects and at prior places of employment (Toivonen, 2004; Skjølsvik, Løwendahl, Kvålshaugen & Fosstenløkken, 2007; Herstad & Ebersberger, 2014). Because the tacit knowledge obtained by the individual expert cannot be assumed articulated and shared before she moves on to the next project, or to a new employer (Alvesson, 2000; Dougherty, 2004; Lam & LambermontFord, 2010), this shifts attention away from the organizational contexts of knowledge exploration and exploitation, towards professional and geographical contexts that are external to firms (Lam, 2000).
In traditional professional services firms (denoted ' P-KIBS' by Miles, 2008), knowledge evolves gradually and the product provided is often defined by scholarly disciplines such as law, accounting, strategy or finance. Mintzberg (1979) refers to this as ‘pigeonholing’ of problems to be solved into categories that are institutionalized as professions and maintain by 4
the education system. The dependence on the individual expert is strong (Alvesson, 2000; Teece, 2003); yet, what defines and legitimizes the competence as such is education combined with ‘deep’ experience from within the domain of the profession in question. ICT services, by contrast, are characterized by commercial experimentation in highly volatile contexts of application, where individuals act as experts in contextualization (i.e. applying their experience-based knowledge to specific client problems), de-contextualization (i.e. learning through generalization of project-specific experiences) and subsequent re-application of what was learnt in order to solve new client challenges (Toivonen, 2004). While there are nuances to this that will be discussed later, this generally means that firms’ dependence on knowledge ‘encoded’ in the form of professions (Blackler, 1995) or embedded in their organizational systems is comparatively low, because these challenges are derivatives of the rapidly evolving technologies and markets that characterizes ‘entrepreneurial’ Schumpeter Mark I technological regimes (Breschi et al., 2000)
Agglomeration and urban knowledge dynamics The relevance of these perspectives on KIS is underscored by studies finding recruitment to be essential for competence upgrading (Keeble & Nachum, 2002; Tether, 2003), and demonstrating how the informal networks and experiences that key individuals have developed through their career paths exert strong influences on the strategic behaviour and performances of employer organizations (Todtling, Lehner & Trippl, 2006; Reihlen & Apel, 2007; Deprey, Lloyd-Reason & Ibeh, 2011; Herstad & Ebersberger, 2014). Accordingly, resources that are fundamental to KIS provision are also resources that are shaped by the career paths of individuals, and thus by the industrial structures of regions because this is level at which exchanges of employees between firms most intensively occurs (Eriksson & Lindgren, 2009; Fitjar & Timmermans, 2016). Generally, this issue can be approached from the perspective of agglomeration theory (e.g. Beaudry & Schiffauerova, 2009). ‘Localisation economies’ refers to the benefits arising from co-location of firms belonging to the same ‘skilled trade’ and points to the advantages arising from regional specialisation (Marshall, 1920; Beaudry & Schiffauerova, 2009) and the pools of specialised skills that are available to firms as a result of it. Accordingly, urban ICT services firms can be expected to benefit from access to experience-based knowledge developed in similar firms: 5
Proposition #1 - specialisation: ICT services firms in urban regions have privileged access to labour markets for sector-specific expertise The contrasting concept of ‘urbanization economies’ capture the benefits of locating in regions characterized by industrial diversity and density (Jacobs, 1969; Glaeser et al., 1992) that are a broader range of markets, collaboration partners and human resources available in the surroundings. Urban resources reflect not only the diversity of the local economy itself, but also the position of such regions as points of gravitation in mobility flows and industrial networks that operate at larger geographical scales, i.e. nationally as well as internationally (e.g. Simmie, 2003; Aslesen, Isaksen & Stambøl, 2008; Strambach, 2012; Herstad & Ebersberger 2015). As a result, theory and empirical research emphasizing the role of largecity regions in enabling entrepreneurial activity and fostering the growth of new industries (Glaeser et al., 1992; Duranton & Puga, 2001; McCann, 2007) are currently paralleled by studies demonstrating specifically how the human resources available therein shape innovation activity and foster linkages to networks at different spatial scales (Herstad & Ebersberger, 2014; Herstad, 2017). Based on such perspectives, a second research proposition is formulated: Proposition #2 - urbanization: ICT services firms in urban regions have privileged access to labour markets for expertise that extend into different (non-ICTs, non-services) domains of the external economy Shearmur (2012) and Herstad (2017) distinguish ‘urbanization economies’ associated with population size, market breadth and better communication infrastructures from ‘urban knowledge dynamics’ associated with interactive learning. This reflects the idea that different knowledge and insights are more likely to diffuse between different firms and industries when they are crammed into a confined geographical space. However, the size and diversity of the urban economy may also result in network fragmentation (Tödtling & Trippl, 2005), and the formation of multiple, unconnected labour market segments. Acknowledging this, a distinction is now made between agglomeration as unrelated variety that protect regions against sector-specific business cycle chocks, and agglomeration as ‘related’ variety that exposes local firms to knowledge and skills that are sufficiently novel to support learning and innovation yet sufficiently similar to allow understanding, transformation and exploitation (Frenken et al., 2007; Aarstad et al., 2016). In urban regions, variety of activity within in the 6
ICT services sector can be expected as a derivate of sector concentration and size. In turn, this variety could be reflected in the labour market positions of individual firms as the unique locational advantage that is access to human resources from different- yet-related branches of the larger ICT services complex: Proposition #3- Related variety: Firms in urban regions have privileged access to labour markets for expertise that reflect the diversity of ICT services activity in their surroundings
Urban differentiation Fundamental to this application of agglomeration theory is the assumption that local conditions are reflected in firms’ access to expertise, which in turn shape their commercial capabilities and growth prospects. Such local conditions are generally portrayed as derivatives of density and diversity, leading some to treat urban location as a binary characteristic of firms (e.g. Lee & Rodríguez-Pose, 2013). However, local conditions are also shaped by specific urban economy characteristics such as the presence of higher-order corporate and governmental functions, different types of research and higher education institutions, and evolutionary processes occurring at the sub-region level, in different business districts. Moreover, the ICT services industry itself may be unevenly distributed between different large-city regions and their business districts; in terms of overall employment levels and in terms of types of services provided based on what types of external expertise. Accordingly, recent research on inter-urban differentiation (Herstad, 2017) align with antecedent work on the ‘polycentric’ and internally differentiated nature of large cities (Suarez-Villa & Walrod, 1997; Brezzi & Veneri, 2014) to form the basis for a fourth research proposition: Proposition #4 – urban differentiation: General urban economy influences on the labour market positions of ICT services firms are paralleled by influences that reflect intrinsic characteristics of different urban regions
Organizational form and external labour market linkages For the sake of clarity in the baseline argumentation, the above conceptualized ICT services firms as small, entrepreneurial entities that operate in a context where pervasive yet unpredictable market and technology change create ample opportunities for innovation yet demand flexibility and responsiveness; i.e. ‘Schumpeter Mark I’ technological regimes (Breschi et al., 2000). This view has a micro-level parallel in organizational theory, where 7
Lam (2000) build on Mintzberg (1979) in using the term ‘operating adhocracies’ to capture organizations that gain their competitive edge from continuous reconfigurations of human resources in response to evolving circumstances. Because such organizations are highly innovative yet unstable, Lam (2000) underscores the importance of support from external ‘occupational labour markets’ (Eyraud, Marsden & Silvestre, 1990) that serve as containing social structures for collective knowledge accumulation. However, as ICTs become more and more essential to a broader range of industrial activities and the industry itself matures, this becomes less the complete industry picture. First, firms in manufacturing as well as other services domains of the economy are now establishing their own dedicated ICT business units, serving internal corporate functions such as data infrastructures support and software development or interacting with other business units in providing products or services to the market (e.g. customer self-service and support). Second, ICT services firms are not necessarily small entrepreneurial organizations that serve local clients and depend on local labour markets for expertise. They may themselves acquire or establish new business units in new areas and geographical locations, and grow to become large actors that provide a broad range of services on different markets. In the business registers, this is evident from the existence of enterprises (i.e. legal ‘firm’ entities) that consists of several establishments (i.e. physical ‘firm’ entities) in different sectors and locations. In these cases, administrative structures and intra-enterprise networks exist that potentially are paralleled by organizational labour markets linking different business units and locations. In combination with large firm size, this may reduce the direct dependence and receptiveness of individual ICT services business units to the occupational labour markets of their locations. Herstad et al. (2014) refer to this as a question of organizational form, and demonstrate how it is associated with the spatial distribution of KIS firms in general, and with their responsiveness to local contexts. A distinction must therefore be made between the small and independent service providers that are single business units in single locations exerting limited gravitational pull in the labour market for expertise; and nonindependent services providers that are constituent components of larger organizational
8
wholes that shelter them from local influences and exert stronger gravitational pulls 2 . From this follows a fifth and final research proposition: Proposition #5: Small and independent ICT services providers are particularly receptive to the labour market of their urban locations
The Norwegian landscape of ICT services Norway is a small, open and high- income economy specialized in deep-water oil and gas extraction technologies, defence technology, seafood, maritime equipment and metallurgical industries (Narula, 2002; Fagerberg, Mowery & Verspagen, 2009; Castellacci & Fevolden, 2014). The country has previously been broken down into 161 housing and labour market regions (Gundersen & Juvkam, 2013), which can be used to delineate the three non-capital urban agglomerations (BERGEN, STAVANGER, TRONDHEIM) from non-urban locations. This corresponds to the centrality classification of Statistics Norway, where 5 is the Capital housing and labour market region, 4 is other large-city regions and 1 – 3 is non-urban regions. Following prior Norwegian research (Isaksen, 2004; Herstad, 2017), intra-urban diversity can be acknowledged by breaking down Centrality 5 into the Central Capital City (CAPITAL C), the Western business district that extends into neighbouring counties (CAPITAL W) and the outer dwelling municipalities (CAPITAL O). The employment statistics and location quotients displayed in Table 1 are computed based on official register data covering all enterprises, establishments and individuals above age 16. The location quotients illustrate the distinctiveness of the Capital city in terms of size of the labour market and industry composition (Aslesen et al., 2008; Jakobsen & Lorentzen, 2015). One third of all Norwegian R&D personnel are working in the Capital labour market region that also account for over 40 per cent of industry expenditures on research, development and innovation (Foyn et al., 2011). The Western business district of the Capital is characterised by strong employment performance in the oil and gas sector and in high-tech manufacturing, ICTs and scientific and technical services, and the services industry in this district function as a node in national and international innovation networks (Herstad & Ebersberger, 2015). In the Central Capital, offshore oil and gas employment and manufacturing employment is
Please not that this refers to the organizational structure of the legal ‘enterprise’ entity, not its ownership structure. Thus, it does not capture whether the enterprise belongs to a larger (foreign or domestic) enterprise group defined as several legal enterprise entities with a common ownership structure. 2
9
under-represented and the specialisation in ICTs, scientific and technical services is less pronounced. This underscores the unique position of the Western Capital (Isaksen & Aslesen, 2001; Isaksen, 2004; Isaksen & Onsager, 2010) in the Norwegian industrial landscape. ----------------------------------------Table 1 approximately here ----------------------------------------Table 2 presents location quotients specifically for ICTs, while Table 3 describes employment growth and distribution on the different regions. The strong location quotients for CAPITAL C and CAPITAL W displayed in Table 2 reflect that 43 per cent and 10.7 per cent of total ICTs employment in 2010 occurred in the Central and Western business districts, respectively. By contrast, only 3.8 percent occurred in the more peripheral dwelling municipalities of the region (cf. also Table 3). In total, 57.5 per cent of Norwegian ICTs employment occurred in the Greater Oslo region as whole. Consistent with the idea that large industry concentrations are associated with ‘related’ industry variety, CAPITAL C and CAPITAL W have in common that they are specialised in a broader range ICT services than any of the other urban locations. Notably, independent firms (cf. Table 3) dominate employment in these two sub-regions. Web design services firms are over-represented in the central Capital (CAPITAL C) (table 2). As previous research has found web design firms to be connected to creative activity and the ‘symbolic’ knowledge of aesthetics and culture (Martin & Moodysson, 2011), this is consistent with the idea that the ‘creative class’ concentrate in certain urban areas (Florida, 2005a). By contrast, the more engineering-based data processing, software consultancy and programming services are over-represented in the Western business district (CAPITAL W), where the ICT industry was hit particularly hard by the bubble burst of the early 2000s. Current activity in the sector cannot be expected independent of the local demand conditions and labour market dynamics associated with strong activity in oil and gas, technologyintensive manufacturing and various advanced and internationally linked non-ICT business services industries (Herstad & Ebersberger, 2015). Outside the Capital, the most notable patterns are displayed by STAVANGER and TRONDHEIM. The role of the former as operational stronghold for the Norwegian offshore oil and gas industry is evident from the exceptionally strong location quotient. Moreover, it is 10
evident from over-representation of employment in the medium-tech manufacturing and technical services industries that supply the oil and gas sector with equipment, technology and support services. The growth rates described in Table 3 show that STAVANGER experienced impressive employment growth in ICTs from 2001 - 2010, even during the sectors’ downturn in 2001-2004. This growth was driven by non-independent establishments. The role of the region as host for IBMs Centre of Excellence in Chemicals and Petroleum (Mikkelsen & Langhelle, 2008) again point to the importance of impulses from the offshore oil and gas sector that was expanding rapidly at the time. --------------------------------------------Table 2 Approximately here --------------------------------------------The smallest large-city labour market region of TRONDHEIM exhibited particularly strong employment growth in scientific and technical services, and in public administration and teaching. Growth in independent services employment even during the years 2001-2004 (Table 3) must be viewed in light of a distinct specialisation in programming services that are presumably dependent on advanced technical skills (Table 2). Taken together, this suggest that the presence of Norway’s leading technical university and one of Europe’s largest applied industrial research institutes (Strand & Leydesdorff, 2013) provide unique labour market support for activity in independent programming firms. --------------------------------------------Table 3 Approximately here --------------------------------------------In 2010, 43 per cent of total employment in ICTs occurred in non-independent establishments, i.e. within enterprises consisting of two or more establishments. Table 4 below shows that these enterprises largely provided technical consultancy services (NACE 71) to other sectors of the economy. Only 16.9 per cent of such employment occurred within enterprises themselves classified as belonging to the ICT services industry3 . This underscores further the importance of distinguishing between independent and non-independent establishments.
3
Data disclosure rules prohibit us from breaking down this employment data on different regions.
11
--------------------------------------------Table 4 Approximately here ---------------------------------------------
Labour market linkages The next question is whether the propensities of firms to use different recruitment channels are influenced by the knowledge dynamics of the urban locations described above. In the business register for 2010, 4572 ICT services firms are identified amongst 289 775 public sector and private sector establishments in total. Each single establishment has been assigned a unique identifier, allowing the firm data to be linked to employment registers. Such datasets are referred to as ‘linked employer-employee’ registers (Boschma et al., 2009; Boschma et al., 2014; Timmermans & Boschma, 2014) and allow recruitment to be observed through year-toyear changes. The use of official register data imply that all Norwegian ICTs that meet the described criteria are included in the analysis. To even out yearly fluctuations, recruitment is observed over the three-year period 20082010. Then, the sample is restricted to the 3097 establishments that had positive (non-missing and larger than zero) employment during this period. This restriction is implemented to avoid influences from ‘sleeping’ firms with no real business activity, and influences from the large movements of labour from single sources that occur when new firms are established as (uncontrolled) spin-offs from existing ones or (controlled) spin-outs by them (e.g. Nås et al., 2003; Andersson and Klepper, 2013).
Dependent variables Three criteria are used to operationalize ‘the individual expert’. Following recent research (Timmermans & Boschma 2014; Herstad et al., 2015), the first criterion is formal education equivalent to a BA-degree or above. Because focus is on access to experience-based knowledge, the second criterion is current employment. The third criterion reflect recent contributions to agglomeration theory (Frenken et al., 2007), in which a distinction is made between recruitment from firms in the same industry (the labour-market expression of ‘localisation economies’, cf. Marshall, 1920); from different yet ‘related’ industries and from industries that are different and presumably ‘unrelated’ to the recruiting firm (‘urbanization economies’, cf. Jacobs (1969). In addition, the analysis acknowledges that recruitment from
12
the research system may provide firms with unique human resources support that cannot be understood merely through the lenses of specialisation versus diversity (Herstad et al., 2015). A large proportion of the sample consists of small and independent firms, for which expert recruitment is relatively rare. As a result, any continuous variable used to describe recruitment would be zero-inflated, and create estimation problems if standard regression techniques were to be applied. To circumvent this problem without restoring to more complex techniques, five binary dependent variables are used. Following prior operationalizations (Boschma et al., 2009; Herstad et al., 2015), these take on the value 1 if firms during the period 2008 – 2010 recruited professionals from within their own NACE 5-digit industry group (SPECIALISATION), from within the same NACE 2-digit group yet different 5-digit (RELATEDNESS), from different 2-digit groups (URBANISATION) or from research system institutions (universities & university colleges, research institutes; RESEARCH). Descriptive statistics are provided in the Appendix.
Estimation strategy The five dependent variables are estimated simultaneously using the Multivariate Probit estimator (mvprobit, cf. Chib & Greenberg, 1998; Cappellari & Jenkins, 2003). Control variables are used to isolate the effect of locations from effects of firm-level characteristics that are unevenly distributed between the different regions considered (cf. Table 1 and 2 above). The first mvprobit estimation include all firms (N= 3097). Reflecting research proposition #5, the second estimation includes only establishments with less than 10 employees at the start of the period 2008-2010, which remained independent throughout it. Dummy variables are used to capture the different urban locations described in Table 1, and estimates are compared to a reference consisting of non-urban regions (CENTRALITY 1-3). Following Herstad (2017), supplementary Wald´s tests are conducted to consider i) whether differences between urban regions (inter- and intra-urban differentiation) are significant in statistical terms, and ii) whether estimates for different urban regions are jointly significant (general urban economy effects) when they individually are not. --------------------------------------------Table 5 Approximately here ---------------------------------------------
13
Results Table 6 displays the results for the mvprobit estimations of recruitment from within firms’ 5digit industry group (SPECIALISATION), and from different 5-digit groups within their 2digit group (RELATEDNESS). From Model 1 Equation A, it is evident that ICTs in all but two urban labour market regions have a higher propensity to recruit within their own industry, compared to firms that are located in non-urban labour market regions. The two exceptions are CAPITAL O and BERGEN, in both of which ICTs employment is under-represented at the outset (Table 2). The supplementary Wald´s tests reveal that differences between the estimates for different urban locations are largely insignificant, and that the coefficient pairs CAPITAL W & CAPITAL C and TRONDHEIM & STAVANGER both are jointly significant when compared to the reference. Thus, the concentration of ICTs in urban locations is generally reflected in the propensity to recruit expertise from other firms in the same industry However, when only small and independent establishments are included in Model 2, the estimate for TRONDHEIM become insignificant, also when tested jointly with STAVANGER. Moreover, the significance of the latter is weakened substantially. The joint significance of CAPITAL C and CAPITAL W, by contrast, remains strong. This is a first indication that inter-region differences in recruitment propensities are more pronounced among the small and independent ICTs that by nature are particularly dependent on the external labour market. A second indication is the absence of positive individual and joint estimates in Model 1 (Equation B), compared to the significant individual estimates for STAVANGER and CAPITAL W and the joint significance of CAPTIAL C and CAPITAL W in Model 2 (Equation B). Small and independent firms in these two business districts of the Capital are significantly more likely to have recruited professionals from prior employment in ‘related’ industries, than are firms located outside the large urban agglomerations. --------------------------------------------Table 6 Approximately here --------------------------------------------Table 7 displays the results for the mvprobit estimations of recruitment outside firms’ own NACE 2-digit industry domains, and from the research system. Model 1 Equation C (URBANISATION) find firms located in the inner and central business districts of the Capital exhibiting significantly higher propensities to recruit outside their own NACE 2-digit industry 14
domain, than firms in the reference. However, while CAPITAL W and CAPITAL C jointly are significantly different from the reference and STAVANGER and TRONDHEIM are not; differences in the individual coefficient estimates are largely insignificant in the estimation that include all firms. This changes dramatically when only small and independent establishments are included in Model 2: Location in the Western industrial stronghold of the Capital is associated with recruitment propensities that are significantly higher than exhibited in any other urban locations, including the Central Capital. Substantively, this means that the access to sector-specialised expertise provided to small and independent firms in this and other urban locations (cf. Table 6) is paralleled by access to expertise developed outside the (NACE 2-digit) domains of services firms that is unique to the Western business district of the Capital.
Last, the propensity to recruit from research system institutions is considered. Model 1 Equation D reveals a striking pattern where firms in the university towns of BERGEN and TRONDHEIM exhibit significantly higher RESEARCH propensities than firms in the Capital, even though firms in CAPITAL C exhibit propensities that are significantly higher than in the reference. When only small and independent firms are included, individual estimates for CAPITAL C, BERGEN and TRONDHEIM remain strong but the difference between CAPITAL C and BERGEN is no longer significant. Keeping in mind the overrepresentation of non-independent ICTs employment in Bergen, this suggests that access to spillovers from the research system has motivated enterprises to establish ICT units in this city. In Trondheim and the Central Capital, by contrast, independent services firms do benefit from privileged access to expertise from the local research system.
--------------------------------------------Table 7 Approximately here ---------------------------------------------
Discussion and implications Concentrations of ICT services employment in urban regions are associated with the formation of specialised labour markets around firms in different brances of the industry (‘localisation economies’). Underneath this general urban economy characteristic that is most pronounced in the Central and Western Capital, the picture is one of inter- and intra-urban 15
differentiation in the learning impetuses that firms receive from the labour market. In the Western business district of the Capital, small and independent ICT services firms thrive and are provided with access to expertise that reflect the diversity and position of the region in the Norwegian industrial landscape (‘urbanization economies’); and its role as national stronghold for a broad range of ICT services activities (‘related variety’). The existence of learning impetuses not only from specialisation but also from related and unrelated variety is unique to this business district. Independent ICT services firms also thrive in Trondheim. Here, labour markets for scientific expertise are exploited by a technically sophisticated segment of the industry within which intra-industry exchanges of employees are not particularly strong. Clear contrasts to these diversity-based (the Western Capital) and research-driven (Trondheim) knowledge dynamics are found in the second largest city of Bergen, and in the oil and gas industry stronghold of Stavanger. In Bergen, employment in the industry is generally under-represented, particularly weak in the web design segment and dominated by non-independent firms engaged in data processing and maintenance services. The support provided to independents from specialised labour markets is weak even compared to rural locations; and their positions in the labour markets for scientific expertise that benefit the local industry as a whole are not particularly strong. In Stavanger, exceptionally strong industry growth rates during the whole 2001-2010 period where driven by non-independent establishments, and resulted in the formation of specialised (‘localisation’) and diverse (‘related variety) labour markets for industry-specific expertise. In this region, industry growth and subsequent agglomeration effects cannot be understood independently of demand from the technologically sophisticated oil and gas sector that was booming during the period. The results have implications for research on urban knowledge dynamics more generally. By demonstrating how small and independent firms are particularly receptive to the labour markets of certain urban locations, the analysis sheds new light on mechanisms at play in allowing these regions to serve as ‘nurseries’ for entrepreneurial ventures and venues for continuous experimentation with new combinations of knowledge originating in different domains of the larger economy (Duranton & Puga, 2001; McCann, 2007; Shearmur, 2015). Thus, it aligns with other recent studies (Herstad & Ebersberger, 2015; Herstad, 2017) in demanding attention to urban differentiation and the unique knowledge dynamics of Capital regions. Such is required to capture the specific locational factors that are ignored in studies 16
treating all urban locations as the same (Lee & Rodríguez-Pose, 2013); considering exclusively the central-periphery dimension of locations (Herstad et al., 2013), or overlooking the intra-urban differentiation that characterizes the larger cities (Jakobsen & Lorentzen, 2015). Potential future challenges to policy are evident in extension of the analysis and empirical results. On the one hand, industry maturity and decelerating rates of technological change may entail that KIS generally and the ICT services industry specifically become more dependent on specialized internal capability building through ‘collective’ organizational processes. On the other hand, if these industries retain the key characteristic that is dependence on learning through external knowledge pools, they will continue to thrive in those urban regions where such markets are most vibrant. In this case, further increases in the relative contribution of services to employment combined with strengthened dependences of manufacturing industries on ICTs and other KIS may shift the locus of industrial development further away from the organizational labour markets and business networks of firms, towards the occupational labour markets and informal networks of locations. Instead of regional convergence as promised by the space-transcending nature of ICTs themselves (Friedman, 2005), further structural change towards a services-based economy may result in a more uneven geographical distribution of industrial activity (Florida, 2005b) that favour large-city regions in general and certain vibrant business districts of capital regions in particular. As the labour markets of such locations make ideas flow more easily and provide the basis for matching of individual expertise with specific firm needs, accelerated urbanisation may aid structural change and strengthen creativity, innovation capacity and productivity in the economy as a whole. However, it may also cause urban regions to grow beyond levels of tolerance that are defined by physical space, transportation infrastructures and housing markets, at the expense of their own citizens as well as other, less well-positioned urban and rural regions. Moreover, re-allocation of human resources from the relatively stable ‘organizational’ labour markets of the advanced manufacturing economy to the more flexible ‘occupational’ labour markets of urban services-based economies transfers market and technology risk from firms to individuals, and influences income distribution and working conditions (e.g. Wessel, 2013) in ways that are not necessarily desirable. This points to the need for policy awareness, and research that in the spirit of Lundvall (1996) considers specifically the social implication of the urbanized, services-based learning economy. 17
Limitations This paper analysed the labour market positions of ICT services firms without empirically considering how they influence capability building. In order to overcome this first limitation, additional data allowing innovation output and productivity estimations are required. Second, while the analysis acknowledged differentiation within the urban hierarchy, it failed to acknowledge differentiation outside it by treating all non-urban locations in the reference as the same. Third, the results inherently reflect specificities of the Norwegian economy. Still, neither one of these limitations reduces the relevance of the overall conclusion, which is that independent ICT services firms are particularly responsive to intrinsic characteristics of their urban locations, and concentrate in those regions where occupational labour markets provide strongest support.
Acknowledgements Research for this article benefitted from funding from the Research Council of Norway under the DEMOSREG program, project number 209769. Tore Sandven assisted with data preparation, and Bjørnar Sæther with comments and frequent discussions. This support is gratefully acknowledged, as are the thoughtful comments received on earlier versions of this paper from two anonymous reviewers and the Editor. Yet, all interpretations and any errors are the sole responsibility of the authors.
18
References Aarstad, J., Kvitastein, O.A., & Jakobsen S.-E. (2016). Related and unrelated variety as regional drivers of enterprise productivity and innovation: A multilevel study. Research Policy, 45, 844-856. Alvesson, M. (2000). Social identity and the problem of loyalty in knowledge-intensive companies. Journal of Management Studies, 37, 1101-1123. Andersson, M., & Klepper, S. (2013). Characteristics and performance of new firms and spinoffs in Sweden. Industrial and Corporate Change, 22, 245-280. Aslesen, H.W., Isaksen, A., & Stambøl, L.S. (2008). Knowledge intensive business services as innovation agent through client interaction and labour mobility. International Journal of Services Technology and Management, 9, 138-153. Beaudry, C., & Schiffauerova, A. (2009). Who's right, Marshall or Jacobs? The localization versus urbanization debate. Research Policy, 38, 318-337. Bettencourt, L.A., Ostrom, A.L., Brown, S.W., & Roundtree, R.I. (2002). Client coproduction in knowledge intensive business services. California Management Review, 44, 100-128. Blackler, F. (1995). Knowledge, Knowledge Work and Organizations: An Overview and Interpretation. Organization Studies, 16, 1021-1046. 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, 169–190. Boschma, R., Eriksson, R., & Lindgren, U. (2014). "Labour Market Externalities and Regional Growth in Sweden: The Importance of Labour Mobility between Skill-Related Industries." Regional Studies, 48, 1669-1690. Boschma, R., & Frenken, K. (2011). The emerging empirics of evolutionary economic geography. Journal of Economic Geography, 11, 295-307. Breschi, S., Malerba, F., & Orsenigo, L. (2000). Technological regimes and schumpeterian patterns of innovation. The Economic Journal, 110, 388-410. Brezzi, M., & Veneri, P. (2014). Assessing Polycentric Urban Systems in the OECD: Country, Regional and Metropolitan Perspectives. European Planning Studies, 23, 1128-1145. Cappellari, L., & Jenkins, S.P. (2003). Multivariate probit regressions using simulated maximum likelihood. The Stata Journal, 3, 278-294. Castellacci, F., & Fevolden, A. (2014). Capable Companies or Changing Markets? Explaining the Export Performance of Firms in the Defence Industry. Defence and Peace Economics, 25, 549-575. Chib, S., & Greenberg, E. (1998). Analysis of multivariate probit models. Biometrika, 85, 347-361. Cyert, R.M., & March, J.G. (1963). A behavioral theory of the firm. Blackwell Publishers Inc.: Malden, MA. Deprey, B.D., Lloyd-Reason, L., & Ibeh, K.I.N. (2011). The internationalization of small and medium-sized management consultancies: an exploratory study of key facilitating factors. The Service Industries Journal, 32, 1609-1621. Doloreux, D., & Shearmur, R. (2012). Collaboration, information and the geography of innovation in knowledge intensive business services. Journal of Economic Geography, 12, 79-105. Dougherty, D., (2004). Organizing Practices in Services: Capturing Practice-Based Knowledge for Innovation. Strategic Organization, 2, 35-64. 19
Duranton, G., & Puga, D. (2001). Nursery Cities: Urban Diversity, Process Innovation, and the Life Cycle of Products. American Economic Review, 91, 1454-1477. Eriksson, R., & Lindgren, U. (2009). Localized mobility clusters: impacts of labour market externalities on firm performance. Journal of Economic Geography, 9, 33-53. Eyraud, F., Marsden D., & Silvestre, J.-J. (1990). Occupational and internal labour markets in Britain and France. International Labour Review, 129, 501-518. Fagerberg, J., Mowery, D.C., & Verspagen, B. (2009). The evolution of Norway's national innovation system. Science and Public Policy, 36, 431-444. Fitjar, R.D., & Timmermans, B. (2016). Regional skill relatedness: towards a new measure of regional related diversification. European Planning Studies, 25, 516-538. Florida, R. (2005a). Cities and the Creative Class. New York: Routledge. Florida, R. (2005b). The World is Spiky: Globalization Has Changed the Economic Playing Field, but Hasn’t Leveled It. The Atlantic Monthly, October, 48-51. Foyn, F., Gundersen, F., Gunnes, H., Langhoff, K., Nås, S.O., Onsager, K., Røsdal, T., Sandven, T., Skyrud, T., & Wilhelmsen, L. (2011). Regionale sammenligninger av FoU og innovasjon [Regional comparisons of R&D and innovation]. In K. Wendt (Eds.), Det norske forskningssystemet 2011 [Science and technology indicators for Norway 2011] (139-181). Oslo, The Research Council of Norway. Frenken, K., Oort, F.V., & Verburg, T. (2007). Related Variety, Unrelated Variety and Regional Economic Growth. Regional Studies, 41, 685-697. Friedman, T.L. (2005). The World Is Flat: A Brief History of the Twenty-first Century. New York: Farrar, Straus and Giroux. Glaeser, E., Kallal, H., Scheinkman J., & Shleifer, A. (1992). Growth in Cities. Journal of Political Economy, 100, 1126-1152. Gundersen, F., and Jukvam, D. (2013). Inndelinger i senterstruktur, sentralitet og BA-regioner [Categorizations of centrality and labor market regions]. By- og regionforskningsinstituttet. Heckman, J.J. (1979). Sample selection bias as a specifcation error. Econometrica, 47, 153161. Herstad, S., & Ebersberger, B. (2014). Urban agglomerations, knowledge intensive services and innovation activity: Establishing the core connections. Entrepreneurship & Regional Development, 26, 211-233. Herstad, S., & Ebersberger, B. (2015). On the link between urban location and the involvement of knowledge intensive business services in collaboration networks. Regional Studies, 49, 1160-1175. Herstad, S., & Sandven, T. (2017). Towards regional innovaton systems in Norway? NIFU REPORT (draft #2, January 2017). Herstad, S., Sandven, T., & Ebersberger, B. (2015). Recruitment, knowledge integration and modes of innovation. Research Policy, 44, 138-153. Herstad, S. (2017). Innovation strategy choices in the urban economy. Urban Studies, Online February 1st . Isaksen, A. (2004). Knowledge-based clusters and urban location: The clustering of software consultancy in Oslo. Urban Studies, 41, 1157-1174. Isaksen, A., & Aslesen, H.W. (2001). Oslo: In what way an innovative city? European Planning Studies, 9, 871-887. Isaksen, A., & Onsager, K. (2010). Regions, networks and innovative performance: The case of knowledge-intensive industries in Norway. European Urban and Regional Studies, 17, 227-243. Jacobs, J. (1969). The Economy of Cities. New York: Vintage. 20
Jakobsen, S.-E., & Lorentzen, T. (2015). Between bonding and bridging: Regional differences in innovative collaboration in Norway. Norwegian Journal of Geography, 69, 80-89. Javalgi, R., Griffith, D., & White, S. (2004). An empirical examination of factors influencing the internationalization of service firms. The Journal of Services Marketing, 17, 185201. Keeble, D., & Nachum, L. (2002). Why do business service firms cluster? Small consultancies, clustering and decentralization in London and southern England. Transactions of the Institute of British Geographers, 27, 67-90. Lam, A. (2000). Tacit knowledge, organizational learning and innovation: A societal perspective. Organization Studies, 21, 487-513. Lam, A., & Lambermont-Ford J.P. (2010). Knowledge sharing in organisational contexts: a motivation‐based perspective. Journal of Knowledge Management, 14, 51-66. Lee, N., & Rodríguez-Pose, A. (2013). Original Innovation, Learnt Innovation and Cities: Evidence from UK SMEs. Urban Studies, 50, 1742-1759. Lundvall, B.-A. (1996). The social dimension of the learning economy. DRUID Working Papers, 96, 1-24. Marshall, A. (1920). Industrial Organization, Continued. The Concentration of Specialized Industries in particular Localities. In A. Marshall (Eds.), Principles of Economics. London: Macmillan. Martin, R., Aslesen, H.W., Grillitsch, M. & Herstad, S. (Forthcoming 2018). Regional innovation systems and global knowledge flows. In A. Isaksen, R. Martin & M. Trippl (Eds.), New Avenues for Regional Innovation Systems – Theoretical Advancements, Empirical Cases and Lessons for Policy, Springer. Martin, R., & Moodysson J. (2011). Innovation in Symbolic Industries: The Geography and Organization of Knowledge Sourcing. European Planning Studies, 19, 1183-1203. Maryska, M., Doucek, P., & Kunstova, R. (2012). The Importance of ICT Sector and ICT University Education for the Economic Development. 3rd International Conference on New Horizons in Education - Inte 2012, 55, 1060-1068. McCann, P. (2007). Sketching Out a Model of Innovation, Face-to-face Interaction and Economic Geography. Spatial Economic Analysis, 2, 117-134. Mikkelsen, A., & Langhelle, O. (2008). Arctic oil and gas: Sustainability at risk? New York, Routledge. Miles, I. (2005). Knowledge intensive business services: prospects and policies. Foresight, 7, 39-63. Miles, I. (2008). Patterns of innovation in services industries. IBM Systems Journal, 47, 115128. Mintzberg, H. (1979). The structure of organizations. NJ: Prentice Hall. Narula, R. (2002). Innovation systems and ‘inertia’ in R&D location: Norwegian firms and the role of systemic lock-in. Research Policy, 31, 795-816. Neffke, F., Henning, M., & Boschma, R. (2011). How do regions diversity over time? Industry relatedness and the development of new growth paths in regions. Economic Geography, 87, 237-265. Nonaka, I. (1994). A dynamic theory of organizational knowledge creation. Organization Science, 5, 14-37. Nås, S.O., Sandven, T., Eriksson, T., Andersson, J., Tegsjö, B., Lehtoranta, O., & Virtaharju, M. (2003). High-Tech Spin-Offs in the Nordic Countries: Main report. STEP Report.
21
Reihlen, M., & Apel B.A. (2007). Internationalization of professional service firms as learning - a constructivist approach. International Journal of Service Industry Management, 18, 140-151. Samiee, S. (1999). The internationalization of services: trends, obstacles and issues. The Journal of Services Marketing, 13, 319-336. Shearmur, R. (2015). Far from the Madding Crowd: Slow Innovators, Information Value, and the Geography of Innovation. Growth and Change, 46, 424-442. Simmie, J. (2003). Innovation and urban regions as national and international nodes for the transfer and sharing of knowledge. Regional Studies, 37, 607-620. Skjølsvik, T., Løwendahl, B.R., Kvålshaugen, R., & Fosstenløkken, S.M. (2007). Choosing to learn and learning to choose: Strategies of client co-prodution and knowledge development. California Management Review, 49, 110-128. Storper, M., & Venables, A J. (2004). Buzz: face-to-face contact and the urban economy. Journal of Economic Geography, 4, 351-370. Strambach, S. (2012). Knowledge Dynamics and Knowledge Commodification of KIBS in Time and Space. In E. Di Maria, R. Grandinetti and B. Di Bernardo (Eds.), Exploring Knowledge-Intensive Business Services. Hampshire: Palgrave Macmillan. Strand, Ø., & Leydesdorff L. (2013). Where is synergy indicated in the Norwegian innovation system? Triple-Helix relations among technology, organization, and geography. Technological Forecasting and Social Change, 80, 471-484. Suarez-Villa, L., & Walrod, W. (1997). Operational strategy, R&D and intra-metropolitan clustering in a polycentric structure: The advanced electronics industries of the Los Angeles basin. Urban Studies, 34, 1343-1380. Teece, D. (2003). Expert talent and the design of professional service firms. Industrial and Corporate Change, 12, 895-916. Teece, D., Pisano, G., & Shuen, A. (1997). Dynamic capabilities and strategic management. Strategic Management Journal, 18, 509-533. Tether, B.S. (2003). The sources and aims of innovation in services: Variety between and within sectors. Economics of Innovation and New Technology, 12, 481-505. Timmermans, B., and Boschma, R. (2014). The effect of intra- and inter-regional labour mobility on plant performance in Denmark: the significance of related labour inflows. Journal of Economic Geography, 14, 289-311. Todtling, F., Lehner, P., & Trippl, M. (2006). Innovation in knowledge intensive industries: The nature and geography of knowledge links. European Planning Studies, 14, 10351058. Toivonen, M. (2004). Expertise as business. Long-term development and future prospects of knowledge-intensive business services (KIBS). Doctoral dissertation series 2004, 2, Helsinki University of Technology. Tregenna, F. (2009). Characterising deindustrialisation: An analysis of changes in manufacturing employment and output internationally. Cambridge Journal of Economics, 33, 433-466. Tödtling, F., & Trippl, M. (2005). One size fits all? Towards a differentiated regional innovation policy approach. Research Policy, 34, 1203-1219. von Nordenflycht, A. (2011). What is a professional service firm? Towards a theory and a taxonomy of knowledge-intensive firms. Academy of Management Review, 35, 155-174. Wang, H.C., He, J.Y., & Mahoney, J.T. (2009). Firm-specific knowledge resources and competitive advantage: The roles of economic and relationship-based employee governance mechanisms. Strategic Management Journal, 30, 1265-1285. 22
Wessel, T. (2013). Economic Change and Rising Income Inequality in the Oslo Region: The Importance of Knowledge-Intensive Business Services. Regional Studies, 47, 10821094. Wezel, F.C., Cattani, G., & Pennings, J.M. (2006). Competitive Implications of Interfirm Mobility. Organization Science, 17, 691-709. Wood, P. (2006). Urban development and knowledge-intensive business services: Too many unanswered questions? Growth and Change, 37, 335-361. Østergaard, C.R., Timmermans, B., & Kristinsson, K. (2011). Does a different view create something new? The effect of employee diversity on innovation. Research Policy, 40, 500-509.
23
Appendix --------------------------------------------Table 8 Approximately here ---------------------------------------------
24
Table 1: Location quotients and employment in the Norwegian urban system. 2010. Capital labour market region Central Western Outer
Other urban labour market regions Bergen Stavanger Trondheim
Centrality 1 - 3
Employment
Agriculture, forestry and fisheries Offshore oil & gas, mining
0,04 0,06
0,18 1,45
0,56 0,05
0,40 1,34
0,99 5,42
0,58 0,54
1,57 0,82
57 830 55 163
High-tech manufacturing Medium high-tech manufacturing Medium low-tech manufacturing Low-tech manufacturing
0,85 0,25 0,24 0,49
1,65 0,53 0,71 0,25
0,27 0,51 0,58 0,91
0,78 1,11 1,14 0,86
0,64 1,03 1,19 0,94
1,41 0,51 0,58 0,97
1,13 1,37 1,32 1,27
11 40 99 81
Infrastructure Construction Wholesale and retail trade Transportation Hotels & restaurants
0,80 0,74 0,95 1,01 1,07
0,36 0,77 1,18 0,63 0,79
0,68 0,96 1,56 1,83 1,00
1,05 0,99 0,93 1,15 1,00
0,61 0,94 0,89 0,91 1,07
0,85 0,99 0,95 0,86 1,17
1,21 1,12 0,97 0,93 0,96
28 023 199 500 362 494 146 769 81 848
Telecom, software and publishing Financial & real estate services Scientific and technical services Business services, other
2,36 1,89 1,50 1,42
3,17 1,70 2,41 1,00
0,47 0,47 0,78 0,97
0,90 1,26 1,08 1,08
0,78 0,75 1,19 1,15
1,01 1,11 1,51 1,11
0,48 0,69 0,67 0,82
86 927 72 740 123 894 134 991
Public administration & teaching Health services Culture, sports & membership org.
1,33 0,72 1,57
0,59 0,88 0,78
0,90 1,02 0,85
0,92 1,03 1,01
0,74 0,82 0,87
1,17 0,99 0,98
0,96 1,12 0,86
345 531 499 430 87 924
17,9 %
3,7 %
5,9 %
8,1 %
6,8 %
5,3 %
52,3 %
2 515 990 (100 %)
Share of employment
Note: Computations based on business register data from 2010. Industry, employment and location is identified at the individual establishment level. Location quotients are computed as region share of Norwegian employment in sector over region share of all employment in Norway.
855 493 293 285
Table 2: Location quotients, ICT services in Norway, 2010. Capital labour market region
Other urban labour market regions
Central
Western
Outer
Bergen
Stavanger
Trondheim
All ICT services
2,40
2,88
0,65
0,89
1,10
1,10
Programming (NACE 62.010)
1,85
2,57
0,50
0,81
1,02
2,07
Consultancy (NACE 62.020, 62.090)
2,11
4,06
0,96
0,81
1,28
0,81
Maintenance (NACE 62.030)
3,65
0,25
0,19
1,21
0,90
0,90
Processing (NACE 63.110)
2,32
3,27
0,13
1,13
1,24
0,64
Web design (NACE 63.120)
4,12
1,36
0,39
0,56
0,15
0,25
Note: Own computations based on business register data for 2010.
Table 3: ICT Services employment and growth rates by organizational form, Norway 20012010. Employment Growth
Region ICT employment by organizational form
Proportion of domestic ICT services employment
2010
2010
2001-2004
2005-2010
2001-2010
Norway all Non-independent est. Independent est. All establishments
-11,01 % -11,36 % -11,21 %
30,17 % 30,04 % 30,10 %
18,32 % 23,48 % 21,19 %
Non-urban locations Non-independent est. Independent est. All establishments
-2,02 % 1,56 % 0,22 %
33,83 % 23,08 % 26,75 %
28,12 % 34,86 % 32,35 %
36,09 % 63,91 % 100,00 %
7,94 % 14,06 % 22,00 %
Trondheim Non-independent est. Independent est. All establishments
-15,13 % 4,37 % -7,25 %
32,07 % 41,92 % 36,38 %
21,27 % 49,40 % 32,64 %
54,45 % 45,55 % 100,00 %
3,17 % 2,65 % 5,81 %
Stavanger Non-independent est. Independent est. All establishments
34,12 % 0,62 % 13,39 %
71,61 % 44,54 % 59,44 %
178,32 % 17,91 % 79,05 %
59,25 % 40,75 % 100,00 %
4,42 % 3,04 % 7,46 %
Bergen Non-independent est. Independent est. All establishments
-4,55 % -4,21 % -4,41 %
43,54 % 18,26 % 30,89 %
26,03 % 41,47 % 32,57 %
54,78 % 45,22 % 100,00 %
3,92 % 3,23 % 7,15 %
Capital O Non-independent est. Independent est. All establishments
1,70 % -17,98 % -7,42 %
-31,81 % 30,77 % -3,36 %
-37,30 % 16,01 % -12,59 %
38,48 % 61,52 % 100,00 %
1,48 % 2,36 % 3,84 %
Capital W Non-independent est. Independent est. All establishments
-53,11 % -20,40 % -34,29 %
61,89 % 14,16 % 27,58 %
-25,01 % -0,19 % -10,74 %
35,69 % 64,31 % 100,00 %
3,83 % 6,91 % 10,74 %
43,36 % 56,64 % 100,00 %
Capital C Non-independent est. -7,91 % 22,89 % 18,71 % 43,29 % Independent est. -18,11 % 38,80 % 22,86 % 56,71 % All establishments -13,61 % 31,43 % 21,03 % 100,00 % Note: Own computations based on annual waves of business register data, 2001 – 2010.
18,61 % 24,39 % 43,00 %
Table 4: Employment in non-independent ICT services establishments by enterprise industry group, 2010. Enterprise industry group NACE NACE NACE NACE Other Sum
71: Technical consultancy services 63: Data processing & web services 68: Real estate 06 - 09: Mining, extraction of crude oil and natural gas
Share of non-independent ICT services employment 77,30 % 16,86 % 3,31 % 2,04 % 0,49 % 100,00 %
Table 5: Control variables Variable
Definition
Motivation
AGE
Continuous variable. Log of firm age in 2008.
Age may influence the labour market position of the firm, and the need to engage in expert recruitment.
GROWTH
Continuous variable. Log of firm growth during the period for which recruitment is observed, i.e. 2008 to 2010: Ln (size 2010 / size 2008) = Ln (size 2010) – Ln (size 2008) ( Cf. Coad & Rao, 2008)
Fast-growing firms may be more inclined to recruit on a broad basis, compared to firms that experience slow or negative growth and therefore recruit more selectively if at all.
INDEPENDENT
Binary variable. Value 1 if the firm did not at any point during the 2008-2010 period belong to a multi-establishment enterprise (Cf. Herstad & Ebersberger, 2014)
Capture the existence (value 0) or absence (value 1) of a larger organizational structure around the observed firm that may influence its receptiveness to the external labour market
SECTOR
4 sector dummies capture the 5 sector groups described in Table 2
The type of ICT service provided is likely to be an important determinant of the extent to which highly skilled and experienced employees are needed, and thus of the labour market position of the individual firm
REPLACEM ENT
The proportion of employees replaced annually, averaged over the three year period 2008-2010 (Cf. Herstad & Ebersberger, 2014).
Increases the need for firms to recruit, and thus the propensity to engage in expert recruitment.
SIZE
Continuous variable. Log of firm size in 2008.
Influences the gravitational pull of firms in the labour market, and may correlate with regions.
Table 6: Within-industry recruitment (‘localisation economies’)
1 CAPITAL C 2 CAPITAL W 3 CAPITAL N 4 BERGEN 5 STAVANGER 6 TRONDHEIM 7 CENTRALITY 1-3
Equation A: SPECIALISATION Modell 1: Model 2: All firms Small & independent firms Coeff. SE Coeff. SE 0,363 0,093*** 0,369 0,126*** 0,354 0,131*** 0,327 0,176* -0,181 0,214 -0,320 0,291 -0,019 0,153 -0,622 0,343* 0,418 0,145*** 0,390 0,205* 0,320 0,153** 0,151 0,246 Reference Reference
Equation B: RELATEDNESS Modell 1: Model 2: All firms Small & independent firms Coeff. SE Coeff. SE 0,097 0,093 0,141 0,132 0,152 0,132 0,376 0,174** -0,453 0,235* -0,323 0,289 0,024 0,144 -0,097 0,237 -0,088 0,155 0,352 0,213* 0,132 0,155 -0,067 0,274 Reference Reference
SIZE AGE GROWTH REPLACEM ENT INDEPENDENT
0,769 -0,324 -0,182 1,741 -0,423
0,771 -0,315 -0,234 1,622 0,062
Walds Chi2 test of inter-region differences CAPITAL C vs. CAPITAL W CAPITAL C vs. CAPITAL O CAPITAL C vs. BERGEN CAPITAL C vs. STAVANGER CAPITAL C vs. TRONDHEIM CAPITAL W vs. CAPITAL O CAPITAL W vs. BERGEN CAPITAL W vs. STAVANGER CAPITAL W vs. TRONDHEIM
0,036*** 0,066*** 0,072** 0,273*** 0,099***
0,00 6,53** 6,13** 0,14 0,08 5,33** 4,29** 0,14 0,04
0,831 -0,481 -0,072 1,384
0,074*** 0,109*** 0,117 0,352***
0,06 5,77** 8,50** 0,01 0,82 4,32** 6,90** 0,07 0,42
0,037*** 0,068*** 0,074*** 0,284*** 0,103
0,12 5,49** 0,25 1,39 0,05 5,72** 0,55 1,72 0,01
Walds Chi2 tests of joint coefficient significance Test 1 (Regions) (1 & 2): 17.02** (1 & 2): 9,05** (1 & 2): 1,80 Test 2 (Regions) (5 & 6): 10.85** (5 & 6): 3,69 (4 & 6): 0,73 Model statistics Wald chi2(60) = 2281.98*** Wald chi2(56) = 921.42*** N 3097 2316 Coefficient estimates and standard errors from multivariate probit regression Models 1 and 2, equations A and B, estimated with 50 draws. ***, ** and * indicate significance at 1 per cent, 5 per cent and 10 per cent levels respectively. Four sector dummies are in cluded.
0,786 -0,564 -0,125 1,658
0,078*** 0,114*** 0,119 0,360***
1,85 2,59 1,02 0,99 0,58 5,09** 3,25** 0,01 2,24
1 & 2: 4,73* -
Table 7: Outside-industry recruitment (‘urbanisation economies’)
1 CAPITAL C 2 CAPITAL W 3 CAPITAL N 4 BERGEN 5 STAVANGER 6 TRONDHEIM 7 CENTRALITY 1-3
Equation C: URBANIZATION Modell 1: Model 2: All firms Small & independent firms Coeff. SE Coeff. SE 0,171 0,080** 0,113 0,094 0,342 0,111*** 0,433 0,127*** -0,202 0,149 -0,184 0,167 0,103 0,125 0,034 0,153 0,133 0,136 0,065 0,170 0,194 0,137 0,007 0,182 Reference Reference
SIZE AGE GROWTH REPLACEM ENT INDEPENDENT
1,004 -0,048 -0,159 2,622 0,086
Walds Chi2 test of inter-region differences CAPITAL C vs. CAPITAL W CAPITAL C vs. CAPITAL O CAPITAL C vs. BERGEN CAPITAL C vs. STAVANGER CAPITAL C vs. TRONDHEIM CAPITAL W vs. CAPITAL O CAPITAL W vs. BERGEN CAPITAL W vs. STAVANGER CAPITAL W vs. TRONDHEIM
0,035*** 0,057 0,063** 0,228*** 0,100
2,27 6,06** 0,28 0,07 0,03 10,29** 2,55 1,72 0,85
1,139 -0,106 -0,150 2,805
0,056*** 0,072 0,085* 0,264***
6,11** 3,08* 0,26 0,08 0,33 10.69*** 5,09** 3,68* 4,42**
Equation D: RESEARCH Modell 1: Model 2: All firms Small & independent firms Coeff. SE Coeff. SE 0,322 0,135** 0,489 0,230** 0,285 0,193 0,483 0,311 -0,738 0,583 a) 0,696 0,175*** 0,804 0,307*** -0,203 0,249 a) 1,054 0,173*** 1,068 0,306*** Reference Reference 0,635 0,033 0,145 1,647 0,098
0,047*** 0,091 0,106 0,420*** 0,132
0,04 3,35 5,01** 4,73** 18,95** 2,95* 3,63* 3,06* 12,76***
0,564 0,161 0,604 2,320
0,122*** 0,160 0,185*** 0,550***
0,00 1,30 4,39** 0,00 2,95*
Walds Chi2 tests of joint coefficient significance Test 1 (Regions) (1 & 2): 10,97** (1 & 2): 11,62*** (1 & 2): 5,98* (1 & 2): 4.82* Test 2 (Regions) (4, 5 & 6): 2,87 (4, 5 & 6): 0,18 Note: Coefficient estimates and standard errors from multivariate probit regression Models 1 and 2, equations C and D, estimated with 50 draws. ***, ** and * indicate significance at 1 per cent, 5 per cent and 10 per cent levels respectively. Four sector dummies are included. a): Predicts failure perfectly.
Table A1: Descriptive statistics & bivariate correlations. N = 3097. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
SPECIALISAT ION RELAT ED URBANIAT ION RESEARCH CAPIT AL C CAPIT AL W CAPIT AL N BERGEN ST AVANGER T RONDHEIM SIZE AGE GROWT H REPLACEMENT INDEPENDENT PROGRAMMING CONSULT ING MAINT ENANCE PROCESSING WEB
11 12 13 14 15 16 17 18 19 20
SIZE AGE GROWT H REPLACEMENT INDEPENDENT PROGRAMMING CONSULT ING MAINT ENANCE PROCESSING WEB
Mean 0,144 0,128 0,318 0,052 0,285 0,093 0,074 0,078 0,058 0,057 1,133 2,124 0,014 0,107 0,879 0,338 0,537 0,037 0,048 0,040
SD 0,351 0,334 0,466 0,222 0,452 0,290 0,261 0,268 0,235 0,231 1,304 0,631 0,444 0,143 0,326 0,473 0,499 0,190 0,214 0,196
Min 0 0 0 0 0 0 0 0 0 0 0 0 -5,421 0 0 0 0 0 0 0
Max 1 1 1 1 1 1 1 1 1 1 7,200 4,710 3,114 1 1 1 1 1 1 1
1 1 0,398 0,385 0,327 0,084 0,025 -0,084 -0,013 0,067 0,039 0,534 0,036 0,082 0,121 -0,280 -0,049 0,059 0,079 -0,062 -0,041 11 1 0,254 0,219 0,044 -0,358 -0,029 -0,097 0,191 0,084 0,041
2
3
4
5
6
7
8
9
10
1 0,379 0,312 0,065 0,018 -0,082 0,004 0,016 0,036 0,513 0,048 0,070 0,105 -0,197 0,044 -0,045 0,113 -0,054 -0,044
1 0,271 0,082 0,028 -0,118 0,003 0,046 0,031 0,673 0,098 0,129 0,198 -0,220 0,001 -0,103 0,106 0,086 0,062
1 0,058 -0,005 -0,060 0,051 -0,015 0,119 0,407 0,100 0,087 0,046 -0,159 0,020 -0,057 0,107 0,009 -0,018
1 -0,202 -0,178 -0,184 -0,158 -0,155 0,088 -0,012 0,010 0,031 0,092 -0,024 -0,019 0,011 0,005 0,090
1 -0,090 -0,093 -0,080 -0,078 -0,002 -0,007 -0,003 -0,002 0,047 -0,016 0,042 -0,040 -0,020 -0,009
1 -0,082 -0,070 -0,069 -0,135 -0,024 -0,007 0,001 0,082 0,005 0,031 -0,043 -0,040 -0,007
1 -0,073 -0,071 0,004 -0,002 -0,014 -0,004 -0,021 0,013 -0,017 0,019 0,019 -0,029
1 -0,061 0,063 0,027 0,012 0,018 -0,064 -0,030 0,022 0,031 0,002 -0,016
1 0,040 -0,003 0,016 -0,030 -0,051 0,029 -0,011 -0,012 -0,003 -0,029
12
13
14
15
16
17
18
19
20
1 -0,083 -0,390 -0,003 0,014 -0,030 0,047 0,034 -0,041
1 0,029 0,060 0,001 0,015 -0,048 -0,013 0,021
1 -0,025 -0,019 -0,031 0,029 0,008 0,087
1 0,091 0,021 -0,209 -0,056 -0,010
1 -0,769 -0,141 -0,161 -0,146
1 -0,212 -0,242 -0,220
1 -0,044 -0,040
1 -0,046
1
