Long-term sustainability of clusters: A dynamic theory of declusterisation

ABSTRACT

In this paper, we build a System Dynamics simulation model to study the long-term counterintuitive consequences of internationalisation strategies of machinery producers in industrial clusters and proposes an explanation for declusterization. To test this hypothesis, we collected empirical data on declusterization and clusters’ exports towards newly industrialized countries, we formalised our hypotheses on the causes of the observed phenomenon in a model, and we explored the behaviour of the model through computer simulation. Comparing the adherence of model-simulated and real data we tested our hypotheses on endogenous causes of declusterization. Our work contributes to widening our perspective on the study of industrial clusters’ dynamics at least under two perspectives. First, we propose an evolutionary perspective that highlights the need to understand the development of clusters’ lifecycle. Second, we warn on long-term, possibly undesired, consequences of clusters’ internationalisation.

KEYWORDS:

• Computer modelling
• system dynamics
• cluster sustainability
• relational capital
• international competition

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

1. Declusterization is also defined in economic literature as cluster decline (Menzel & Fornahl, 2009).

2. Knowledge can be divided into two categories (Kogut & Zander, 1992): information and know-how. Information is the knowledge that can be transmitted without loss of integrity once the syntactical rules for deciphering it are known. Know-how is the accumulated practical skill. Knowledge as information implies knowing what something means. Know-how implies knowing how to do something.

3. The model is available for inspection from authors upon request.

4. Italian Bureau of Statistics.

5. According to the above-mentioned law, Italian regions have been asked to map regional specialisation areas (distretti).

6. It is the association represented all the Italian districts.

7. Montebelluna (BL), Riviera del Brenta (VR), Vigevano (PV), Parabiago (MI), S. Croce sull’Arno (PI), Valdinievole-Leporecchia (PT), Macerata-Fermo (MC), Fusignano-Bagnacavallo, S. Mauro Pascoli, Val Vibrata, Barletta, Castrano, Napoli.

8. Ventana Systems Inc. The version of the software is 5.9c.

9. In the diagram, the “+” (“-”) sign implies a positive (negative) correlation between two variables as follows: $x\underset{+}{\to }yyields\frac{\mathrm{\partial }y}{\mathrm{\partial }x}>0$ and. $x\underset{-}{\to }yyields\frac{\mathrm{\partial }y}{\mathrm{\partial }x}<0$