Advancing innovation in manufacturing firms: knowledge base combinations in a local productive system

ABSTRACT

Industry 4.0 requires that manufacturing firms not only innovate but also generate more radical and different varieties of innovation, often incorporating new types of knowledge. To advance such types of innovation, several studies in Innovation and Economic Geography foreground that firms need to combine knowledge in novel ways. The contribution of this paper is to investigate in-depth how manufacturing firms with traditional roots combine new generative knowledge in and beyond a local productive system (LPS), what enables them to access and integrate such knowledge from external sources, and how this relates to the firms’ innovation performance, with a focus on radical and varied forms of innovation. Findings show that firms standing out in terms of innovation performances combine complementary types of knowledge through internal and external sources, particularly at national and international levels. Moreover, firms that have complementary knowledge internally are able to access new knowledge beyond the borders of the LPS.

KEYWORDS:

• Industry 4.0
• knowledge bases
• local productive system
• innovation
• manufacturing firms

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1 A local productive system (LPS) is an identified place that is associated to specific productive specialisations well rooted in the local community (Bellandi et al. 2021, p. 1).

2 As often used in the literature, we proxy radical innovations with innovations that according to the most common OECD classifications is at least new-to-the-market in the specific sector (cf Hervas-Oliver et al., 2022).

3 The Veneto region and in particular the province of Vicenza has numerous LPSs concentrating on traditional low and medium tech sectors of “made in Italy” specializations, such as fashion, furniture, agro-food, sport equipment and mechanics; all having quite dynamic international markets. Recent statistical data of the chamber of commerce shows that the percentage of mechanics and electro-mechanical employees double the Italian average.

4 The mechatronics specializations present at local level are still partly based on technical skills resembling almost ‘artisan work’. The production of tailored machineries is particularly suitable for responding to the incremental adaptation of products and services and peculiar clients’ needs which find correspondence in the ‘Made in Italy’ typology of production as well as in similar niches productions in GVCs.

5 E.g. Plechero and Chaminade (2013); Martin et al. (2018).

6 Firms were asked to judge on a scale ranging from 0 (not at all important) to 5 (very important) the presence within the company of different types of knowledge base (a. scientific, b. technical/engineering and 3. creative knowledge) supporting innovation activities. We considered the knowledge base of strategic importance only when firms provided values 4 or 5.

7 One of the main aims of the local Ph.D. programme is to strengthen the scientific training of future human resources for the research activity related to the local industry.

8 Concretely, in the survey question related to sourcing we asked firms if in the years 2014–2016 the company had purchased externally (and at which geographical level) from other companies/organizations factors and activities related to the different knowledge bases reported in figure 2. In a subsequential question, we asked if the company had (within the same time frame and in relation to the three activities of basic/scientific research; technical and development and aesthetic and creative design) some active collaborations (also at informal level) with external partners. The original survey is available from the authors upon request.

9 We gave value 1 to each degree of innovation for any of the 5 different introduced types of innovation activities during 2014-2016, and summed up the values.

10 Since the innovation index measures with a count variable the varied innovation, we considered more appropriate to use a partition in 3 sets and not 2 as the former highlights the tails of the distribution.

11 We used different regression models according to the typology of the dependent variable applied. The logit model has been used for the regressions with dependent dummy variables: ‘product innovation’ and ‘top innovation’. The poisson model has been used for the regression with the counting variable ‘innovation index’. All the models have been checked for the robust standard errors that adjust for the heteroscedasticity.

12 The description of the variables and related correlations are reported in Appendix B.