CESIS focal research questions pertain to sectoral innovation systems, international networks and innovative SMEs, and span five work areas.
Work Area 1: Dynamics of firms’ innovation efforts – strategies, resource bases and evolution
In both manufacturing and service sectors we can observe that a large share of all firms do not allocate resources to R&D and alike innovation efforts and that another group enters into innovation activities occasionally. Only a smaller set of firms spend resources on innovation persistently over extended time sequences. This work area relies on the theoretical building blocks of evolutionary economics and assesses the dynamics of firms’ innovation efforts. The research pertains to policy which aims (i) to support the transition of firms to alternative innovation strategies with long-term R&D commitments and to (ii) develop and restructure innovation networks of firms. The work area will deliver knowledge of how firms develop resource bases over time, and the role of R&D and network assets (links, nodes and their attributes) in these processes.
In the traditional view innovation has only transitory effects on firms’ profitability and growth by altering its competitive position in the short run. The introduction of innovation gives the firm a temporary monopoly power by increasing firms’ market-share, which allows for higher profit until other firms can imitate the innovation (Aghion and Howitt 1992, Klepper 1977). This approach is for instance common in the literature on first mover advantages and on patent races.
A second approach emphasizes that innovation intrinsically ‘characterizes’ a firm in that it creates a structural difference between innovating and non-innovating firms. Each firm is assumed to own different technological competencies that are firm-specific, difficult to imitate and emerge cumulatively from learning processes. A firm’s internal competencies together with its strategic behavior enable it to innovate, grow and survive in the market (see e.g. Malerba and Orsenigo 1995, Cohen and Levinthal 1989, Dosi et al. 1995). This approach corresponds to the resource-based view of the firm (Penrose 1958, Barney 1991).
The work area adheres to the second view and associates to problems of dynamic interdependencies and true path dependence in the evolution of firms and their capabilities and innovativeness. The research will carefully specify the time scales for different change processes, and uses econometric techniques complemented by interviews of selected firms to assess interdependencies. By interacting with a set of international research collaborators CESIS expects to advance existing research methods and econometric techniques to develop new insights into the role of R&D strategies and networks for productivity, evolution of capabilities and growth. The work area relates to the research frontier on the inter-connection of strategy, R&D, firm capabilities, networks and growth (see e.g. CBO 2005, Dosi and Nelson 2009, Cefis and Orsenigo 2001, Peters 2009 as well as Lööf and Johansson 2010). Methodologically the research area includes development of structural models, introduced by Lööf and Heshmati (2002, 2006).
The work area also associates to the recent literature on the role of sector and firm characteristics for innovation, which puts into focus their relative role in explaining innovation (see e.g. Peneder 2010, Srholec and Verspagen 2008). Firm data on innovation and performance consistently show much heterogeneity of behavior among individual companies. At the same time, sectoral data repeatedly demonstrate persistent and significant differences between sectors, e.g., with respect to average factor intensities, dominant corporate strategies, entry rates, or firm duration. For example, Malerba (2007) points at the apparent tension between these two stylized facts: while the first stresses variety, the latter emphasizes common contingencies among firms operating within the same markets. In most empirical analyses the tension between the micro- and the meso-levels of observation largely remains unresolved. Persistent differences between sectors draw attention towards specific technology fields, where observed regularities in industry data are interpreted as if they represent the behavior of the individual firms. Conversely, the variety of firm behavior causes many researchers to focus exclusively on micro-data, frequently discarding any aggregate levels of analysis .
Work area 1 will deal with the following issues:
- Temporal analysis of innovation efforts and economic performance, including application of wavelet techniques
- Time scales for resource-base development vis-à-vis firm performance, including application of wavelet techniques
- Transitions from non-persistent to persistent innovation efforts – probability and explanatory characteristics
- The role of network characteristics for firms’ resource-base and innovation performance, including networks for goods and services as well as knowledge networks
- The relative of role of firm and sector characteristics for in explaining innovation and R&D strategies of firms
The work area will employ firm-level register data with detailed information of the firms and their sectors and locations, export and import flows to and from different countries together with a sequence of Community of Innovation Surveys (CIS).
Work Area 2: International networks, export performance and innovation of firms and sector-grouped firms
The last five decades have witnessed an increasing role of international networks for firms, sectors and regions. The world economy is ‘globalizing’, bringing about a growing interdependence of economies, involving consumers, producers, suppliers and governments in different countries.
International networks are important in at least two ways. The first is that they are a vehicles for sales on foreign markets, and thus export revenues. In a small open economy like Sweden, openness – as measured by the total value of exports plus imports as a fraction of GDP – has increased from just above 20 % in the 1950s to over 90 % in the 2000s, and net exports contribute substantially to Sweden’s GDP (export revenues exceed imports). Also, international markets (exports) have grown in importance for many firms, SMEs as well as large firms (see e.g. OECD 1997). Andersson et al. (2008) show with Swedish firm-level data that 76 % of all firms in Sweden in manufacturing sectors are engaged in international trade, and that more than half (55 %) both exports and imports. Lööf (2010) reports that half of the Swedish firms exporting goods are service firms and that they account for a substantial and increasing share of the total value from exports of goods. Between 1997 and 2006 this fraction increased from 25 % to 34 %. Hence, international networks are important for generating revenues as well as securing input deliveries and product varieties from other countries.
The second way in which trade is important is that international networks are channels for knowledge and information flows. The endogenous growth literature has identified various channels of international knowledge spillovers. Based on the models of Grossman and Helpman (1991) and Rivera-Batiz and Romer (1991), recent studies have documented R&D cross-country knowledge spillovers through trade as an important engine of TFP growth in the industrialized countries. Eaton and Kortum (1996) suggest that even a large economy like the United States obtains over 40 percent of its growth from foreign innovation. Keller (2002) uses data that cover more than 65 percent of the world's manufacturing output and most of the world's R&D expenditures, and show that 20 % of the productivity of an industry in a country can be attributed to R&D expenditures in foreign industries, accessed through import flows. Though this literature more often than not focuses on imports of capital goods as a vehicle for technology flows, recent firm-level analyses also find that firms’ engagement in international trade have positive feedback effects on firms, so-called ‘learning-by-exporting’ (see e.g. Andersson and Lööf 2009, Baldwin and Gu 2003, Fernandes and Isgut 2007, Greenaway and Kneller 2008).
Work area 2 will assess the role of international networks in the form of exports and imports from both of the ways described above. First of all, the work area comprises studies where consequences of innovation efforts are measured in terms of export performance, where the latter may be recorded as network links to different destination markets, export value flows across destination links (e.g. Andersson and Johansson 2008, 2010), product quality measured as unit price premium across export links (e.g. Aiginger 1997, Andersson and Ejermo 2008, Johansson 2010), number of product varieties in the total export flow and specified for different destination links. Second, the work area includes studies that assess export and import flows as vehicles for technology diffusion, and will pay special attention to the role of link (e.g. temporal and product characteristics of export and import flows) and node attributes. Node attributes are essentially characteristics of destination and origin countries, where it should be observed that Acharya and Keller (2007) find that the combined effect of R&D investments in countries close to the world’s technology frontier is on average about three times as large as that of domestic R&D. In contrast to the bulk of the existing literature and as called for in the research literature (see e.g. Keller 2004), CESIS will primarily undertake firm-level analyses while considering the importance of industry and product-group aggregates. Third, the work area includes studies that assess the interdependence between innovation efforts and engagement in international trade. These analyses will draw on recent theorizing on the interaction between firms’ decision on innovation strategies and their decision to enter export markets (see e.g. Costantini and Melitz 2007 and Aw et al. 2008). In these models the interdependence goes in principle as follows: firms invest in R&D and physical capital, which can affect the path of future productivity for the firm. R&D investments, through their effect on productivity, increases the profits from exporting, and participation in the export market in turn raises the return to R&D investments.
Work area 2 will deliver knowledge of the role innovation efforts play for export revenues and GDP at the aggregate level, and what firm attributes and milieu characteristics are important for firms’ decision to participate in international trade. It will provide policy-relevant knowledge of international trade networks as vehicles for knowledge and technology diffusion. In general, these aspects are important for a small open economy like Sweden, where a large fraction of the firms are engaged in trade and export revenues are important for GDP. A specific policy issue concerns the interdependence between exports and innovation outlined above. The described interdependence imply for instance that export support can stimulate innovation efforts (large markets raise returns for R&D investments, which may be viewed as fixed costs), while R&D support may stimulate firms’ to export. This may suggest for instance that R&D support could be complemented by export support, and suggest a general interaction between innovation and export policies (e.g. VINNOVA and the Swedish Trade Council).
Work Area 3: Knowledge flow networks and innovation results for different sectors
Competitiveness and innovation is to a large extent dependent upon the ability to apply new knowledge and technology in products, services and their production processes. The creation and the diffusion of new ideas are processes which imply the integration and recombination of existing knowledge coming from different sources, locations and organizational positions. With rapid advancement of knowledge and technology, firms need to secure access to knowledge and information, and ‘knowledge flow networks’ play an important role in this context.
Within industrial economics and business studies, knowledge network concepts have been applied to the theory of the firm (Kogut 2000), to studies of organizations (Burt 2003), and to the analysis of strategic alliances for research, technology transfer, and standard setting (Andrews and Knoke, 2001). Yet, as Breschi and Lissoni (2004) remark, despite the frequent use of the knowledge network concept, quantitative research is still in its infancy.
The current work area will focus on assessing the role of knowledge flow networks for innovation and growth of firms, sectors and economy-wide. Knowledge flow networks refer to measurable flows associated to knowledge and information, and the work area focuses on knowledge flows as evidenced by patent citations, mobility of engineers and skilled labor as well as university graduates. For this, CESIS will employ firm-level EPO Worldwide Statistical Database (PATSTAT) including citations, matched employer-employee data and data describing interaction opportunities across space with spatial accessibility measures. In the accessibility analyses, CESIS will employ its developed techniques for estimating knowledge accessibility for different sectors and firms. International comparison using PATSTAT will rely on firm-level statistics from Compustat and AMADEUS databases as well as direct collaboration with other countries having access to registerdata on firms characteristics and trade statistics.
The use of patent citations has quite a well-established tradition in the economics of innovation. First, citations have been used along with patent re-classification and co-word analysis when searching for technology families, and comparing the knowledge base of different companies on the basis of their patent portfolios (Pilkington, Dyerson and Tissier 2002). Second, citations have been used to assess the quality of individual patents, which has been shown to increase with the number of citations received (eg. Ejermo 2009). This allows evaluating the economic value of companies’ patent portfolios. Third, citations have been increasingly interpreted as ‘paper trails’ left by knowledge flowing from the inventor or applicant of the cited document to the inventor/applicant of the citing one (Jaffe et al 1993). Moreover, patent citations analyses are also used to examine patterns of technical change in large technology systems. Fontana et al. (2008) state that identifying the structure of patent citations in different technology fields or sectors makes it possible to assess the main trajectories that have characterized its evolution.
In this work area, patent citations will primarily be used to identify knowledge flows from and to firms. Connectivity structures based on patent citations will be matched with the structure of international trade flows, flows of engineers and R&D workers and spatial accessibility measures describing the potential for interaction. This allows for an assessment of the correspondence between indirect knowledge flow networks and more direct measures. Mobility of knowledge labor between firms will be identified by Swedish matched employer-employee data.
Work area 3 will assess the following research issues:
- Patent citations as components in innovation networks of firms, sectors and countries
- Correspondence between knowledge flow networks, as evidenced by patent citations, and international trade networks, mobility of people and spatial accessibility measures
- The role of labor mobility for firms’ resource bases and innovation behavior
- Patents, citations and economic performance of firms, and aggregate of firms grouped by technology fields and sectors
- Accessibility measures that reveal the potential for knowledge flows across space to explain product development and other innovation efforts.
Work Area 4: The role of knowledge-intensive service sectors in innovation networks
Both large and small urban regions in the OECD group of countries are rapidly being transformed to economies that have an expanding share of services as well as an augmented knowledge intensity of the labor force. The change is clearly more accentuated in large agglomerations, and productivity and wage levels increase faster in places with a larger share of – in particular – knowledge intensive producer services, often referred to as KIBS (Storper and Venables 2004). This change process seems to reflect outsourcing of both standardized and knowledge-intensive services.
The growth of business services represents a qualitative new stage in the social structure of production in that firm-level scale economies with regard to knowledge and skill inputs are reduced, while being replaced by external deliveries of such inputs, thereby exploiting external scale economies (Gelaluff et al. 2004). The outsourcing, reflecting external economies of scale for the customers of KIBS suppliers, is one source of productivity enhancing change. An indirect effect obtains from spillovers that are generated by the innovation networks which are created through KIBS activities.
While the role of services for employment growth – especially knowledge intensive ones – is well established, there is a lack of systematic analyses on the role and function of services in innovation (Link and Siegel 2007). Most researchers hold that an accurate model of service innovation is still absent from the literature (Howells 2000). The general argument in the literature is that services adds value by integrating purchased physical technology, which embodies others’ R&D and innovation activity, into systems. Thereby, they play an important role in technological change and productivity growth by promoting standards and systems integration. Yet, knowledge of the structure and effects of linkages and transactions networks between knowledge intensive services and other firms in the economy is lacking.
Work Area 4 is outlined to investigate the role of knowledge-intensive business-service suppliers in innovation networks of all type of firms, including manufacturing, household services and business services. This includes a mapping of innovation networks of different business-service sectors in Sweden for a longer time period to corroborate structural changes. Focal questions include: what role has knowledge intensive services firms’ innovation processes, do they replace other networks for knowledge flows, and what sectors or markets do they serve? This comprise assessments of service sector firms as actors participating in other firms’ innovation efforts, the relationship between innovation efforts and performance amongst service sector firms, investigations of how service inputs affect customer firms’ export, import and technical renewal. As work area 3, the current work area will also employ spatial accessibility techniques developed by CESIS to provide insights into the role of proximity for business services networks. This will provide policy-relevant knowledge associated to urban planning and management of location and infrastructure.
Work Area 5: Entrepreneurship and innovation networks
The notion of entrepreneurship is loaded with complex meanings, ranging from finding a source of income when no jobs are available to the drive of individuals’ to create novelties, while the strive for temporary entrepreneurial rent remains the centre of entrepreneurial gravitation. However, many researchers emphasize institutional arrangements as an explanation of why the frequency of entrepreneurship varies between different, places, regions and countries.
In view of this, the work area asks: Which features make such structural differences remain invariant between decades, and how can such knowledge be employed in policy. Such analyses can now be carried out with the help of considerably long time series with clear opportunities to investigate dynamic interdependencies. Andersson and Koster (2010) find that spatially sticky and durable determinants of start-ups, which implies that there are sources of persistence to be examined in detail to increase the understanding of how certain functional regions can maintain a start-up rate which is close to double that of the average.
The start of new firms and the introduction of new products (goods and services) to the market is a process which on the micro level reveals a high frequency of entry and exit, whereas the macro pattern normally remains approximately invariant. This context concerns how entrepreneurs discover new business opportunities and how they develop and exploit networks for mobilizing joint innovation efforts. What is the pre-history of new entrepreneurs, and which networks do they carry with them when they leave an employment to start a company (Almeida and Kogut 1999). Relevant networks comprise both links to capital sources and knowledge technology and customer opportunities. Again, new data bases to which CESIS has access will help to illuminate these questions and provide guidelines to innovation network policies.
An recent example of CESIS research in this vein is Andersson, Baltzopoulos and Lööf (2010). Examining entrepreneurial ventures of ex-employees of firms with different R&D strategies three findings not well documented in the previous literature First, firms with persistent R&D investments with a general superiority in sales, exports, productivity, profitability and wages are less likely to generate entrepreneurs than firm with temporary or no R&D investments. Second, start-ups from knowledge intensive business service (KIBS) firms with persistent R&D investments have a significantly increased probability of survival. No corresponding association between the R&D strategies of incumbents and survival of entrepreneurial spawns is found for incumbents in manufacturing sectors. Third, spin-outs from KIBS-firms are more likely to survive if they start in the same firm, indicating the importance of inherited related knowledge. The findings suggest that R&D intensive firms spur fewer entrepreneurs, but their entrepreneurial spawns tend to be of higher quality.
Work area 5 associates to many pertinent policy issues. The most apparent concerns conditions conducive for entrepreneurship in the form of new firms. It also relate to work area 4 (knowledge intensive services) and its policy relevance. The frequency of product introduction and formation of new firms increases in knowledge-intensive service industries and especially knowledge intensive industries and the bulk of new firms is indeed knowledge intensive service firms. Moreover, entrepreneurial knowledge is spatially sticky, embodied in individuals and networks connecting relevant people and thereby tacit in nature. This suggests that spatial relocation and establishment of new interaction links are important policy measures in the development of sectoral networks.
 True path dependence is in contrast to spurious path dependence in panel data models, see e.g. Heckman (1981).
This issue has practical implications for innovation policy (Peneder 2010). On the one hand, without a proper understanding of the co-evolution of variety and contingency government authorities are easily mislead into an obsession for ‘high-tech’ industries. A biased perception of innovation potentials can thus lead to the misallocation of public funds, if, for example, innovative companies in traditional sectors find it more difficult to access public funding than firms with lower innovation potential in a ‘high-tech’ industry. On the other hand, industry characteristics matter and cannot be ignored. Their accurate understanding helps to design policy programs and tailor them more effectively to the needs of targeted firms.
Typical arguments are that exporting firms may accumulate knowledge and technology from their activities in foreign markets, such that exports have positive effects on firms’ knowledge and technology accumulation (Wagner 2007), that foreign buyers may provide product designs and technical assistance to domestic suppliers in order to improve the efficiency of their sourcing activities (Evenson and Westphal 1995) and that exporting may imply the reduction of X-inefficiencies and stimulate the renewal of development and production processes (Greenaway and Kneller 2007).
 A special category of firms is those forms which belong to multinational company groups, since these firms has their internal company group networks, while also having to decide where to locate their R&D facilities.