Entrepreneurship dynamics and economic cycles: an analysis for local systems and industrial districts

ABSTRACT

We propose a research framework that expressly takes into account the moderating and differentiated influence that can be exerted on the dynamics of entrepreneurship capital by both the type of local system and the economic situation. Moreover, we investigate whether industrial districts are a more favourable context for the evolution of the entrepreneurship capital. We apply this framework to data of municipalities, grouped in local labour systems, of three Spanish regions and in two sub-periods of, respectively, crisis (2007–2011) and incipient recovery (2011–2014). Results are consistent with a set of sensible relations that can be extracted from the background literature.

KEYWORDS:

• Entrepreneurship capital
• local labour systems
• industrial districts
• economic cycles

Disclosure statement

No potential conflict of interest was reported by the authors.

ORCID

Maria J. Ruiz-Fuensanta http://orcid.org/0000-0003-2021-0040

Marco Bellandi http://orcid.org/0000-0002-9044-5630

Notes

1. See Sternberg (2011) for a complete and clear description of the determinants of entrepreneurship from a territorial perspective.

2. Due precisely to this coexistence of random and fixed effects in the same equation, multilevel models or hierarchical models are also known as mixed effects models.

3. Corrado and Fingleton (2012) claim the utility of combining multilevel analysis with spatial econometrics methods. In particular, the potential limitations of the SAR model can be overcome with the structure of a mixed model, that is, a model in which there is a combination of fixed and random effects (Dong & Harris, 2015).

4. The partial adjustment approach has been applied to a wide range of topics ranging, for example, from microeconomics and macroeconomics (e.g. Coen-Pirani, 2004) to capital capital structures (e.g., Mai, Meng, & Ye, 2017), etc.

5. This spatial lag is constructed as a weighted average using a first-order binary contiguity matrix of (n × n) items, in which item w_ij takes a value of 1 if the local labour systems i and j are geographically adjacent and 0 otherwise. For more information, see Anselin (1988).

6. If the total variance of the EC is denoted:(2) ${\sigma }^2={\sigma }_{\mathrm{r}\mathrm{e}\mathrm{g}}^2+{\sigma }_{\mathrm{L}\mathrm{L}\mathrm{S}}^2+{\sigma }_{\mathrm{m}\mathrm{u}\mathrm{n}}^2$(2) where, ${\sigma }_{\mathrm{r}\mathrm{e}\mathrm{g}}^2$ is the component of variance between regions, ${\sigma }_{\mathrm{L}\mathrm{L}\mathrm{S}}^2$ is the variance of LLS between regions and ${\sigma }_{\mathrm{m}\mathrm{u}\mathrm{n}}^2$ is the variance associated with municipalities within regions and LLS, then the ICC for municipalities within regions is calculated as(3) $\mathrm{I}\mathrm{C}{\mathrm{C}}_{\mathrm{r}\mathrm{e}\mathrm{g}}=\frac{{\sigma }_{\mathrm{r}\mathrm{e}\mathrm{g}}^2}{({\sigma }_{\mathrm{r}\mathrm{e}\mathrm{g}}^2+{\sigma }_{\mathrm{L}\mathrm{L}\mathrm{S}}^2+{\sigma }_{\mathrm{m}\mathrm{u}\mathrm{n}}^2)}$(3) and the ICC for municipalities within LLS and regions is computed as(4) $\mathrm{I}\mathrm{C}{\mathrm{C}}_{\mathrm{L}\mathrm{L}\mathrm{S}/\mathrm{r}\mathrm{e}\mathrm{g}}=\frac{({\sigma }_{\mathrm{r}\mathrm{e}\mathrm{g}}^2+{\sigma }_{\mathrm{L}\mathrm{L}\mathrm{S}}^2)}{({\sigma }_{\mathrm{r}\mathrm{e}\mathrm{g}}^2+{\sigma }_{\mathrm{L}\mathrm{L}\mathrm{S}}^2+{\sigma }_{\mathrm{m}\mathrm{u}\mathrm{n}}^2)}.$(4)

7. In all four cases, the Moran I test statistic (Cliff & Ord, 1981) had previously confirmed that EC is positively spatially auto-correlated. In other words, LLS with re1ative1y high EC (resp. low) are located near other LLS with a re1atively high EC (resp. low), more often than if their locations were merely random.