A spatial analysis of growth and convergence in Italian provinces: the role of road infrastructure

ABSTRACT

A spatial analysis of growth and convergence in Italian provinces: the role of road infrastructure. Regional Studies. The role of road infrastructure in reducing economic disparities is debated in both Europe and the United States. During the 1960s, huge investments in highways significantly reduced the time run between north and south Italy. A spatial analysis of convergence is performed. Results show that the effect of road infrastructure investment was significant but accompanied by a strong polarization between north and south (also emerging from non-parametric analysis) that it failed to prevent, possibly because investment in the south was not large enough to close the accessibility gap with the centre–north, a gap that appears to persist during the subsequent years.

摘要

意大利各省中的成长与聚合之空间分析：道路建设的角色。Regional Studies. 在欧洲与美国，道路基础建设之于降低经济落差的角色同样受到争议。1960年代高速道路的巨幅投资，显着减少了意大利南北之间的行车时间。本研究进行聚合的空间分析。研究结果显示，道路基础建设投资具有显着影响，但却同时造成了南北部之间无法避免的大幅两极化（同时也从非参数分析中浮现），而这可能是因为南部的投资并不足以弥合其与中北部可及性之间的差距，该差距似乎在接下来的数年中持续存在。

RÉSUMÉ

Une analyse spatiale de la croissance et de la convergence des provinces italiennes: le rôle de l’infrastructure routière. Regional Studies. En Europe et également aux États-Unis le rôle de l’infrastructure routière afin de réduire les disparités économiques fait l’objet d’un débat. Pendant les années 1960, d’importants investissements en infrastructure routière a sensiblement réduit le temps de parcours entre le nord et le sud de l’Italie. On fait une analyse spatiale de la convergence dont les résultats laissent voir que l’impact de l’investissement en infrastructure routière s’est avéré important. Cependant il est accompagné d’une polarisation nord–sud tellement forte (ce qui ressortit aussi d’une analyse paramétrique) qu’il n’a pas réussi à l’empêcher. Il est possible que l’investissement affecté au sud ne suffisait pas pour combler l’écart en matière d’accessibilité avec le centre–nord, un écart qui semble persister pendant les années suivantes.

ZUSAMMENFASSUNG

Räumliche Analyse von Wachstum und Konvergenz in den italienischen Provinzen: die Rolle der Straßeninfrastruktur. Regional Studies. Die Rolle der Straßeninfrastruktur für die Verringerung der wirtschaftlichen Disparitäten wird sowohl in Europa als auch in den USA diskutiert. In den sechziger Jahren verkürzte sich durch umfangreiche Investitionen in Autobahnen der Zeitaufwand für Fahrten zwischen Nord- nach Süditalien erheblich. Wir führen eine räumliche Analyse der Konvergenz durch. Aus den Ergebnissen geht hervor, dass die Auswirkungen der Investitionen in Straßeninfrastruktur signifikant waren, aber mit einer starken Nord-Süd-Polarisierung einhergingen (die auch aus der nicht-parametrischen Analyse hervorgeht), die sie nicht zu verhindern vermochten, möglicherweise weil die Investitionen im Süden nicht umfangreich genug ausfielen, um die Erreichbarkeitslücke gegenüber dem Zentrum und Norden zu schließen – eine Lücke, die in den folgenden Jahren offenbar fortbesteht.

RESUMEN

Análisis espacial de crecimiento y convergencia en las provincias italianas: el papel de la infraestructura vial. Regional Studies. Tanto en Europa como en los Estados Unidos se debate el papel de la infraestructura vial en la reducción de las desigualdades económicas. Durante los sesenta, las enormes inversiones en las autopistas disminuyeron drásticamente el tiempo de desplazamiento entre el norte y el sur de Italia. Aquí realizamos un análisis espacial de la convergencia. Los resultados muestran que el efecto de la inversión en la infraestructura vial fue significativo pero vino acompañado de una fuerte polarización entre el norte y el sur (que también aparece en el análisis no paramétrico) que no pudo impedir, posiblemente debido a que la inversión en el sur no fue lo suficientemente grande como para cerrar la brecha de accesibilidad en comparación con el centro y norte, una brecha que parece persistir en los años siguientes.

KEYWORDS:

• highways
• convergence
• parametric analysis
• non-parametric analysis
• spatial econometrics

关键词:

• 高速道路
• 聚合
• 参数分析
• 非参数分析
• 空间计量经济

MOTS-CLÉS:

• routes nationales
• convergence
• analyse paramétrique
• analyse non-paramétrique
• économétrie spatiale

SCHLÜSSELWÖRTER:

• Autobahnen
• Konvergenz
• parametrische Analyse
• nicht-parametrische Analyse
• räumliche Ökonometrie

PALABRAS CLAVES:

• autopistas
• convergencia
• análisis paramétrico
• análisis no paramétrico
• econometría espacial

JEL:

• C14
• C21
• R11
• R12

DISCLOSURE STATEMENT

No potential conflict of interest was reported by the authors.

SUPPLEMENTAL DATA

Supplemental data for this article can be accessed at https://doi.org.10.1080/00343404.2017.1334117.

ORCiD

Stefania Cosci http://orcid.org/0000-0002-8120-8887

Loredana Mirra http://orcid.org/0000-0001-8386-2439

Notes

1. It is very difficult to compare the highway effect in the United States, an environment dominated by large firms, with the effect in Europe, an environment often dominated by small firms.

2. The analysis excludes all the segments of the highway that are free of payment.

3. See Geurs and Van Wee (2004) for a review of the accessibility measures.

4. See Appendix A in the supplemental data online for details.

5. According to Schürmann and Talaat (2000), captures the potential market for firms producing services, while evaluates the potential markets for firms producing goods.

6. NUTS = Nomenclature des Unités Territoriales Statistiques.

7. Annual provincial data are not available for the five decades.

8. The regions of Friuli-Venezia Giulia, Molise and Sardegna are not broken down into provinces as done by Cosci and Mattesini (1995) (see Appendix D in the supplemental data online).

9. See Durlauf and Quah (1999), Silverman (1986) and Appendix B in the supplemental data online for details.

10. The stochastic kernels are calculated by means of a Matlab routine developed by Magrini.

11. The Italian government decided to develop the Italian highway network in 1953. When Amintore Fanfani (the chief of the government during the early 1960s) forced the highway to pass through his town, it was for reasons that must be considered exogenous. See Table E3 in Appendix E in the supplemental data online.

12. The spatial weight matrix is row standardized such that all elements in each row sum to 1 (Anselin & Bera, 1998; Anselin, Bera, Florax, & Yoon, 1996; Niebuhr, 2001; Badinger, Müller, & Tondl, 2004). See Appendix C in the supplemental data online for details.

13. Whenever Moran’s I test on OLS errors reveals the presence of residual autocorrelation (Cliff & Ord, 1981), this suggests that the standard growth regression should be extended explicitly to include spatial elements, and it is typically estimated via maximum likelihood (ML)-based techniques.

14. See also Arbia, Battisti, and Di Vaio (2010), Seya, Morito, and Yoshiki (2012), Chapman and Meliciani (2016), and Meliciani (2016) for an application of the SDM specification to the spatial analysis of growth and regional convergence.

15. Lagrange multiplier (LM) tests and their robust versions are used to test the OLS versus the SAR and SEM LM tests and their robust versions are used to test the OLS versus the SAR and SEM; likelihood ratio (LR) tests are used for testing the SAR and SEM versus the SDM, while the test of the SLX versus the SDM is a t-test on the coefficient of the spatial lag of the dependent variable in the SDM. If the (robust) LM tests point to another model than the LR tests, then the SDM is adopted. This is because this model generalizes both the spatial lag and the spatial error model.

16. In addition, the robustness of results is checked using spatial weight matrices based on the k-nearest neighbours with k = 10, 15, 20, 25 neighbours (Ertur & Koch, 2006). A k-nearest neighbours weight matrix is computed from the distance between the regions’ centroids and implies that each region is connected to the same number k of neighbours. Complete results on robustness analysis are available from the authors upon request.

17. All computations in this paper were carried out using SpaceStat 1.91 (Anselin, 1999) and the spatial econometrics toolbox in Matlab (LeSage, 1999).

18. For an application of spatial regimes analysis, see Allers and Elhorst (2005) and Chapman, Cosci, and Mirra (2012).

19. Provinces of the centre–north are those belonging to the regions of Val d’Aosta, Piemonte, Lombardia, Veneto, Trentino-Alto Adige, Liguria, Emilia-Romagna, Toscana, Umbria, Marche and Lazio. The remaining provinces are included in the group named ‘south’.

20. See Table E1 in Appendix E in the supplemental data online for a detailed description.

21. The correlation between I^truck70 and the variable Roads in 1971 is 0.07.

22. The highway linking Naples to Pompei was opened in 1923; that linking Pompei to Salerno in 1961.

23. We included the variables usually considered as potential determinants of local growth (Ferri & Mattesini, 1997; Aiello & Scoppa, 2000; d’Acunto, DeStefanis, & Musella, 2004) available at the provincial level for the whole period. We report in Table E5 in Appendix E in the supplemental data online only the ones that we found significant at least in one regression. We introduce in our regressions only variables referred to in the initial year of each period. For this reason population growth was not included, also considering that in the post-war period population migrated toward growing provinces. In any case, we have tried to control also for this variable and the main results are unaffected (the results are available from the authors upon request).

24. All the checks for the presence of possible outliers were done (we performed data trimming on the variables included in our econometric analysis and no outliers were detected).

25. A decrease in spatial dependence in Italy during the 1960s is reported also by Arbia and Basile (2005).