Who Are the Knowledge Brokers in Regional Systems of Innovation? A Multi-Actor Network Analysis

Abstract

Kauffeld-Monz M. and Fritsch M. Who are the knowledge brokers in regional systems of innovation? A multi-actor network analysis, Regional Studies. The discussion on regional innovation systems emphasizes the duality of local and global links. This empirical analysis of eighteen German regional innovation networks shows that public research organizations, especially universities, are profoundly involved in knowledge-exchange processes and possess more central (broker) positions within their regional innovation networks than private firms. This results, in part, from public research's ‘gatekeeper function’, which can be particularly important in lagging regions that typically suffer from a lack of large firms that often fill this role in advanced regions. The transferred knowledge is absorbed, especially, by private firms without inter-regional research and development cooperation activity.

Kauffeld-Monz M. et Fritsch M. Les agents de la connaissance dans les systèmes d'innovation régionaux, qui sont-ils? Une analyse de réseau d'agents multiples, Regional Studies. Le débat sur les systèmes d'innovation régionaux met l'accent sur la dualité des liens à la fois locaux et mondiaux. Cette analyse empirique de dix-huit systèmes d'innovation régionaux en Allemagne laisse voir que les instituts de recherche publics, notamment les universités, se sont engagés à fond aux processus de l'échange de la connaissance et ont plus de positions centrales (en tant qu'agents) au sein de leur réseaux d'innovation régionaux que ne l'ont les entreprises privées. En partie, cela remonte à la fonction de la recherche publique comme ‘gardienne’, ce qui peut s'avérer particulièrement important dans les régions en perte de vitesse qui souffrent comme à leur habitude d'un manque d'entreprises de taille qui jouent souvent ce rôle dans les régions en pleine croissance. La connaissance transférée est absorbée, notamment par les entreprises privées qui ne font pas de la recherche et du développement interrégionaux.

Systèmes d'innovation régionaux Réseaux d'innovation Analyse de réseau Agent de la connaissance Gardienne

Kauffeld-Monz M. und Fritsch M. Wer sind die Wissensbroker in regionalen Innovationssystemen? Eine Netzwerkanalyse mit verschiedenen Akteuren, Regional Studies. Der Ansatz der regionalen Innovationssysteme betont die Bedeutung der Dualität globaler und lokaler Austauschbeziehungen für Innovationsprozesse. Unsere empirische Analyse von 18 regionalen Innovationsnetzwerken in Deutschland zeigt, dass öffentliche Forschungseinrichtungen – insbesondere Universitäten – intensiv in die Wissensaustauschprozesse dieser Netzwerke involviert sind und mehr zentrale (Wissensvermittler-)Positionen einnehmen als die in den untersuchten Netzwerken vertretenen Unternehmen. Dies resultiert zum Teil daraus, dass die öffentliche Forschung in Regionen mit Entwicklungsrückstand eine ‘Gatekeeper-Funktion’ wahrnimmt, welche in besser entwickelten Regionen typischerweise größeren Unternehmen zukommt. Das in das Netzwerk eingespeiste Wissen wird insbesondere von denjenigen Unternehmen absorbiert, die über keine eigenen regionsexternen FuE-Partnerschaften verfügen.

Regionale Innovationssysteme Innovationsnetzwerke Netzwerkanalyse Wissenstransfer Gatekeeper

Kauffeld-Monz M. y Fritsch M. ¿Quiénes son los intermediarios del conocimiento en los sistemas de innovación regionales? Un análisis de las redes de varios actores, Regional Studies. El debate sobre los sistemas de innovación regionales pone de relieve la dualidad de los vínculos locales y globales. En este análisis empírico de dieciocho redes de innovación regionales en Alemania mostramos que las organizaciones públicas de investigación, especialmente las universidades, están muy involucradas en los procesos de intercambio de conocimientos y poseen posiciones más centrales (de intermediarios científicos) en sus redes de innovación regionales que las empresas privadas. Esto se debe en parte a que la investigación pública tiene una función de ‘guardián’, lo que puede ser especialmente importante en regiones cuyo desarrollo ha quedado rezagado y que suelen carecer de las grandes empresas que con frecuencia cubren esta función en regiones avanzadas. El conocimiento transferido es absorbido especialmente por empresas privadas sin investigación interregional ni actividad de cooperación al desarrollo.

Sistemas de innovación regional Redes de innovación Análisis de redes Intermediario del conocimiento Guardián

Keywords:

• Regional innovation systems
• Innovation networks
• Network analysis
• Knowledge broker
• Gatekeeper

JEL classifications:

• D83
• D85
• L14
• O18

Notes

At the firm level, Tushman and Katz (1980) found that gatekeepers positively affect the performance of R&D projects within R&D units.

All these regions are of about the same geographical size.

For example, biotechnology, medical technology, automotive, innovative textiles, phytopharma, health industry, and musical instruments.

Five of the networks that were involved in the InnoRegio initiative have been excluded from the study either because of very small numbers of participating actors or because of their particular innovation objectives (for example, ‘social’ innovations such as barrier-free tourism).

The number of organizations: 142 manufacturing firms, eighty service firms, thirty-five universities, twenty-seven non-university PROs, twenty-eight private research organizations, and twenty-six other organizations (for example, educational institutions and regional agencies of business development). The majority of the PROs belong to the Fraunhofer Association. Max-Planck Institutes are hardly involved in the networks.

More than 500 R&D projects were conducted and granted in the programme. They differ considerably with regard to their research topics, duration, financial volume, and partners involved. However, the subsidies were basically restricted to the early stage of innovation.

The networks were restricted to organizations that have been funded by the policy programme.

It is assumed that an organization has transferred information and knowledge to a certain network member if it was named by this network member as an important partner. Absorption takes place if an organization named a certain network member as an important partner. Thus, mutual information and knowledge exchange only occur if two organizations name each other as important partners.

For measurement details, see the fourth section.

Probably for this reason, Ahuja (2000) found that indirect connections among firms positively affect innovation, although the effect is moderated by direct ties.

Private firms' share of knowledge they transferred amounts to 48% (PROs = 43%). Thus, the numerical dominance of private firms does not crowd out PROs' meaningful transfer value.

Knowledge transfer as well as knowledge absorption of non-university PROs turns out to be significantly lower than those of the universities (at the 5% level; Mann–Whitney U-test).

The standardized measure corresponds to the degree of an organization divided by the maximal possible degree that is calculated on the basis of the total number of organizations, multiplied by 100. Thus, the standardized measure takes the network size effects into consideration.

Statistically significant at the 1% level (Mann–Whitney U-test).

The correlation coefficient is 0.125 (statistically significant at the 5% level). With respect to the universities, a positive, but insignificant, correlation coefficient of 0.144 was found. The correlation coefficient for the non-university PROs had a non-significant negative value.

The PRO located at the middle of the top in Fig. 3 may serve as an example. To calculate the broker measure, the organizations' direct relations (ego-network) are taken into account, which amount to five. Thus, for this actor, a maximum of twenty broker positions (n*(n – 1)) is attainable. According to Fig. 3, this PRO is linked to five pairs of organizations that are not connected directly. Additionally, the organization connects four other pairs of organizations that are not linked reciprocally, but only in one direction. Such links in which knowledge is only transferred in one direction are only counted as 0.5. As a result, the calculation of the number of broker positions accounts for the exchange directions. Altogether, the examined PRO attains seven broker positions (5 + (4*(0.5)).

All mentioned differences between brokers and non-brokers are statistically significant at the 5% level of significance.

In three out of the eighteen networks, one university has an enormous number of broker positions (367, 94.0 and 92.5 broker positions, respectively).

Those seven universities that do not assume a broker position in the networks under study also show an extremely low level of knowledge exchange with network partners. In cases where a university does not have at least one broker position in a network, the innovation activity of the network does not predominantly rely on academic knowledge. It is known from the inquiry that these universities do also exchange knowledge with other actors, but these actors do not participate in the respective network.

The correlation coefficient (Pearson) for product innovation is 0.474 (statistically significant at the 1% level); for process innovation it is 0.337 (statistically significant at the 5% level).

Significant differences at the 10% level between the two groups ‘with/without inter-regional cooperation activity in product innovation’.

The correlation coefficient (Pearson) is 0.243 (statistically significant at the 5% level).