A Schumpeterian approach to innovation clustering in a low-tech technology in a peripheral region: the case of garments in Mezzogiorno.

SUMMARY

This paper examines the characteristics of innovation in a low-tech industry in a peripheral region. Its theoretical framework draws mainly upon the Schumpeterian hypothesis concerning the links between innovation and firm size, merged with the recent contributions on economics of knowledge production and on the path-dependence approach. The methodology involved both quantitative and qualitative analysis, coupling the use of correspondence analysis and appreciative theorizing. Consistent with industrial district literature, informal knowledge accumulation proved to be crucial. Here innovation mainly emerges from the exploitation of knowledge coming from linkages with final markets, rather than from local interactions with subcontractors. Size matters and only larger firms appear successful in taking advantage of positive feedback from network externalities.

KEY WORDS

industrial districts; innovation; Schumpeterian legacy; path dependence; final markets; firm size; informal knowledge accumulation

1. INTRODUCTION

The goal of this paper is the analysis of innovation dynamics in an industrial district localized in the South of Italy, and specialized in the production of a low-tech product; i.e., garments and textiles. Industrial districts are mainly characterized by the fragmentation of the production process and the widespread presence of small firms. On the other hand, the Schumpeterian legacy stresses the role of large firms as carriers of innovation within the economic system. Hence we wonder to what extent size can matter also in contexts characterized by the presence of network externalities and easier knowledge circulation due to higher levels of trust among agents. Moreover the case of lock-in, stemming from the embeddedness in local cliques, may emerge as an additional and critical issue.

The paper is organized as follows. In Section 2 the analytical context is provided, both reviewing the literature dealing with innovation and industrial districts, and elaborating the working hypotheses. In Section 3 the empirical analysis is presented, merging the quantitative and the qualitative approach. Section 4 presents the main conclusions stemming from the discussion.

2. ANALYTICAL CONTEXT

The analysis of innovation in low-tech industries and peripheral regions is poor represented in literature. On the one hand one can find plenty of studies concerning the dynamics of knowledge creation and innovation in high-tech sectors, like biotechnologies or semiconductors. On the other hand, when mature sectors come into the analysis, the attention is mainly focused on economically advanced areas. Moreover it is very rare to find studies focused on innovation dynamics within labor intensive industries, like the garment one, while one can easily find analyses of capital intensive sectors like mechanics or automotives.

The garment industry clearly is fashion-driven sectors, in which there is not great scope for technological innovation. We again refer here to the work of Schumpeter (1934), who defined basically innovations as the introduction of either new products, or new processes, or new organizational forms, or new raw materials or new final markets. Surely product innovation in our case represents the main source of innovation among local firms, and this has sound implications on the dynamics feeding creativity.

Background

Firms innovate in order to react to the pressures of the economic environment (Schumpeter, 1942). In particular different interpretations of these pressures have long shaped the economic debate, which can be grouped into two main strands: the factor costs inducement (Hicks, 1932; Fellner, 1961; Kennedy, 1964) and the demand pull theories (Young, 1928; Kaldor, 1957). In the theory of localized technological change firms innovate because of changes in factor costs and in demand levels, in order to save switching costs due to irreversibilities in learning activities (Antonelli, 1995). Technological change is localized in factor markets, in product markets, in sectors, learning processes and geographical contexts. Localization then emerges as a consequence of the appreciation of path dependence in economic choices (Antonelli, 1999).

Innovation is the eventual result of the knowledge creation process, which turns out to stem from the combination of four kinds of inputs: internal codified knowledge; internal tacit knowledge; external codified knowledge; and, external tacit knowledge (Antonelli, 1999, 2001). The role of external knowledge has been strongly emphasized in the literature dealing with the non-linear and interactive character of innovation processes (Gibbons et al., 1994, Lundvall, 1992, Von Hipple, 1988).

It is well known that in Schumpeter's work two main approaches can be found. On the one hand, in the Theory of the Economic Development (1934) he stresses the role of entrepreneurs as basic engine of innovative activity within an economic system characterized by the presence of small firms and low barriers to entry. On the other hand, in Capitalism, Socialism and Democracy (1942), the role of oligopolistic rivalry is appreciated, and innovations are the eventual outcome of industrial strategies, pursued by large firms through R&D expenditures.

In this context, the empirical evidence that small firms, in specific market conditions, are basic innovative agents, attracts the attention to the inputs that they can employ in knowledge production. In particular, knowledge production models have strongly emphasized the role of R&D activities, which are usually carried out within research institutions or large firms. So, the issue of where small firms get knowledge inputs, then emerges. Audretsch (1995) argued that small firms resort to third parties, such as just research institutions and large firms, in order to get knowledge. The latter, in turn, may spill over from the context of production and can be applied by other firms (Acs et al., 1992).

Knowledge spillovers have proved to be geographically clustered (Audretsch and Feldman, 1996; Baptista and Swann, 1998), and firms are likely to base their location choices on the opportunities of taking advantages of the positive feedbacks associated to knowledge externalities. The spatial concentration applies above all when informal rather than formal cooperation ties are at work (Audretsch and Stephan, 1996).

Different approaches to the clustering of technological knowledge production can be observed in literature, in which the idiosyncratic character of the context of production seems to be a common ingredient. The level of communication costs is influenced by transaction costs, so that the likelihood of technological knowledge percolation among different agents depends on the probability that neighbor agents share the same set of norms and rules and the same knowledge framework, such as the language and the vocabulary used by technicians; and that they are characterized by a strong commitment in cooperating (Saxenian, 1990; Porter, 1990; Storper, 1995a and 1995b; Storper and Scott, 1995; Cooke et al., 1997).

Knowledge creation is cumulative not only in a synchronic perspective, but also in a diachronic one. The internalization of knowledge from the surrounding environment is the result of agents' intentional efforts. Firms are likely to absorb external knowledge, in function of the prior knowledge they possess, which has been cumulated through internal learning processes (Cohen and Levinthal, 1990). In view of this, localization emerges mainly as a consequence of dynamic irreversibility; that is of the repeated actions of economic agents in a well defined set of techniques, relationships, markets, organizational structure (Antonelli and Quere, 2001).

The appreciation of the historical flavor characterizing both learning processes, and communication and clustering dynamics, makes the concept of path dependence relevant. History matters in a very peculiar way, in which the phenomenology at the time t is dependent on the choices made at the time t-1. The existence of a multiplicity of alternatives make the new state only one of the possible outcomes, and then it is not possible to fully anticipate the last outcome starting from the initial state (David, 2001; Antonelli, 2003a).

Industrial districts appear to provide an environment conducive to creativity because of two sets of arguments. Firstly, as far as proximity matters, within industrial districts there is a strong interplay between the geographical and the technological dimension. Secondly, the institutional asset, both in terms of formal organizations and of norms and trust, has proved to be quite supportive to innovation in such areas (Brusco, 1982; Becattini, 1991; Patrucco, 2005). Moreover recent empirical findings shows that in Italian industrial districts intentional innovative activity can be found beside the working of externalities and knowledge spillovers (Cainelli and De Liso, 2003).