«Der Open-Access-Publikationsserver der ZBW – Leibniz-Informationszentrum Wirtschaft The Open Access Publication Server of the ZBW – Leibniz ...»
The aim of this paper is to explore to what extent spatial reach of collaboration linkages determines innovation orientation and innovative behavior. Therefore we use descriptive and explorative approaches. Having in mind the findings of the last two sections we want to analyse if these concomitant intra- and interregional links are a precondition for innovation. In contrast, given a successful regional innovation system, local links could be a substitute for more far reaching collaboration activities.
That means: Innovation performance could be positively related to (a) to a high intensity of local collaboration; (b) the intensity of international collaboration, or (c) neither regional nor (inter)national collaborations.
We build our analysis on a data set collected from a sample of firms in the district of the city of Jena and the adjacent counties (Landkreise) with a maximum distance of 25 kilometers. The information was collected during the time period August 2009 to March
2010. The sample includes several different industries and service sectors from manufacturing to trade and IT-services. In comparison to several studies quoted in sections 2 and 3 we cover a broad range of industries and services.
The basic population comprises 811 firms with at least 5 employees. This population data set relies on information provided by the firm registers of two renowned commercial private data banks (Creditreform and Hoppenstedt). All these firms were contacted by phone in order to identify partners for “face-to-face” interviews with a sound knowledge as to the firm innovation behavior and economic conditions. Finally we conducted personal interviews lasting 40-60 minutes with 280 enterprises, representing a response rate of 35%. Due to a lack of any innovative behavior of firms on the one hand and other data collecting problems (e.g. incomplete answers, interview cut-offs) on the other hand this finally resulted in a sample size of 216 interviews with SME, which pursued innovation projects within the last three years.
Projekt KompNet2011 Erfolgsfaktoren regionaler Innovationsnetze SME are defined according with EU policy as firms with up to 250 employees. As to large firms the reach of collaboration without exception covers the national and very often the international dimension (Fritsch 2000, Freel et al. 2009). Hence differences with regard to the spatial reach of collaboration are likely to be particularly revealing with regard to small and medium enterprises.
The concentration on innovators is possible, because our aim is not to distinguish innovators and non-innovators but we are interested in explaining the interdependence of innovation success on the one hand and the geographical reach of collaboration on the other hand.
In order to cope with self-selection problems the contacts by phone always included a question as to the reason of the refusal to participate. The answers corroborate the idea that there is no systematic and non-random influence as to the non-participators.
Given the efforts devoted to the data collection process, the survey is only to a limited extent plagued by the self-selection bias problem found in so many empirical studies.
The questionnaire used closed-ended but nonetheless detailed questions designed to catch the supposed inherent complexity and various meanings of first, the term innovation, second, the collaboration channels (i.e. the knowledge transfer processes) and third the spatial reach of collaboration.
As to innovation the questions ground on the Oslo-manual definitions of innovative behavior. We distinguish product innovation, process innovation, organizational innovation and marketing innovation. With regard to the first two forms of innovation the questionnaire includes innovations new to the market or new to the firm as well as improvements of existing products or processes.
Concerning cooperation as an instrument to promote innovation, the transfer of knowledge becomes particularly relevant. Innovation always has its roots in new knowledge. Thus we pay special attention to the diverse aspects of knowledge transfer processes. Therefore, 16 different collaboration and knowledge transfer options were identified and requested. The types of different channels vary from formal cooperation (personal contract based work, test jobs, etc.) to informal cooperation (workshops, attending of fairs, personal non-contract based work, etc.). By means of this detailed range of transfer channels we should be able to identify differences in the variety of possible collaboration behaviors.
In order to investigate the importance of these different types of knowledge transfer possibilities more precisely, the questions measure the intensity of use of these channels on a 6-point scale (Likert-scale-type).
In addition we ask for the innovation relevance as well as the sectoral and spatial dispersion of collaboration activities. With regard to the latter we distinguish four geographical dimensions: local, regional, national and international linkages.
Besides we include all control variables discussed in section 3:
The variable SIZE is defined as the total number of full-time equivalent employees in 2009 to control a linear effect on the innovation performance. Furthermore we add the square term of SIZE to allow for a curvilinear relationship (SIZE²).
The independent variable AGE is simply the age of the firm, i.e. years since founding.
Based on the classification of economic activities (Federal Statistical Office Germany, Edition 2008) we aggregated the following twelve industry sectors (see
Table 1 Industry dummies and classification
To capture sectoral patterns in the innovation performance we control for significant effects of BR_2, BR_3 and BR_4. The value of these industry dummies takes 1, when the firm belongs to the corresponding sector; otherwise it takes 0.
The independent variable HOUSE measures if a firm claims in-house development to be the most important type of innovation development (value = 1). This dummy variable allows us to look for the influence of absorptive capacity on the innovation performance.
We asked the respondents to what extent different cost factors, including equity financing and debt financing of innovation projects as well as too high innovation costs, have inhibited their innovation activities. We sum up the individual evaluations of the three cost barriers to calculate the variable FINA.
Moreover, we include the share of graduates on the total number of employees by the variable GRAD and the share of research and development employees by the variable RND. Both have a continuous index with a range between 0 and 1.
We control the variety of transfer relations in two different ways: we are able to analyze the number of transfer channels (N_CHAN) and the number of transfer partners (N_PART) used. Therefore we compile a list of 16 transfer channels and seven partners. In both cases the collaboration intensity is evaluated on a 6-point Likert-scale from “0 – not important” till “5 – very important”. We count the number of channels and partners evaluated with 4 or 5.
In addition we analyze the impact of relationships with scientific partners measured by the variable SCIEN, which sums up the intensity of collaboration with universities, universities of applied sciences and research institutes.
In order to identify the influence of the strategic performance on the innovation activities, we determine the relevance of cost leadership (COST) as well as quality leadership (QUAL) on a 6-point Likert-scale from “0 – not important” till “5 – very important”.
To control for the impact of the intensity of competition, often mentioned as a relevant determinant on the innovation activity, we asked for the importance of Porter‟s five competitive forces on a 6-point Likert-scale from “0 – not important” till
“5 – very important”. Simply adding these numbers for suppliers, substitutes, customers, potential and current competitors (Porter 2004) we build an index of intensity of competition named COMP.
In order to identify the relevance of legal regulations (LEGA) as barriers to innovation, the questionnaire included the relevance on a 6-point Likert-Scale from “0 – not important” till “5 – very important”.
5 Descriptive findings and econometric resul ts In this section a first step summarizes the relevant variables and presents several descriptive analyses. In this respect we first examine the regional reach of collaboration in general. Second, we focus on the collaboration relevant for innovation and its geographical pattern. Third, the links of different forms of innovation and their regional reach are under scrutiny. In a final step we discuss the results of a basic regression model.
Table 2 Descriptive statistics of dependent and independent variables
Table 2 provides a brief overview on the variables included in the analyses, their value ranges, acronyms and descriptive location parameters.
The survey comprises information with regard to the importance of several innovation types (evaluated on a 6-point Likert-scale). We include product and process innovations both new to the market as well as marketing and organizational innovations. Hence we are able to provide a more detailed analysis of innovations in comparison to previous studies. As to the spatial distance of collaboration we measure the percentage of transfer partners at the same place of location (LOCAL), in the remaining federal state (REGIO), in the rest of the country (NATIO) and abroad (INTER).
On a general level, that means considering all the 16 knowledge transfer channels of our questionnaire, we identify four different patterns of spatial reach (see Pfeil et al.
2011 and table 3):
I. Collaboration with dominant role of local relationships, e.g. student trainees II. Collaboration with dominant role of supraregional relationships, e.g. advanced training of firm members III. Collaboration with uniform distribution over the local, regional and national distance, e.g. economic consulting IV. Collaboration with distance paradox, e.g. research contracts The expression distance paradox refers to the fact, that there is first no smooth decline with increasing spatial distance and second no uniform spatial distribution of collaboration activities (Rosenfeld & Roth 2004). To the contrary, a clear dip at the regional distance level emerges.
Chart 3 illustrates the patterns of collaboration with regard to the four geographical areas and underlines the differences between the transfer channels examined.
In the following analyses we do not explore the geographical dispersion of all of the transfer channels, but concentrate on those channels which are highly relevant for the innovation projects of a specific firm.
55 firms evaluate none of the transfer channels as very innovation relevant and 50 enterprises practice only one transfer channel to speed up the development of innovations. At the maximum nine important transfer channels are identified. The specific collaboration channels, which are rated as highly relevant for innovation projects, also vary from firm to firm. The number of firms, which declare the relevant transfer channel as highly innovation relevant, ranges from advanced training (N=85) and workshops (N=68) to doctoral thesis (N=4) and lectureships (N=3).
Projekt KompNet2011 Erfolgsfaktoren regionaler Innovationsnetze In order to examine the relevance of the four cooperation regions we calculate the mean spatial concentration of those transfer channels, which are evaluated as highly relevant for the firm‟s innovation projects at the level of each firm. Table 4 presents the results with regard to the four geographical regions. Overall about one-third of the innovation-relevant collaborative activities are aimed at the local level and the same holds as to the national level. Only 6% of all cooperations are international oriented, while the remaining 22% are allocated to regional collaborations. Thus, when we examine the cooperative relationships relevant for innovation on an aggregated level, we find the spatial pattern of collaboration called “distance paradox”.