Transport infrastructures, spillover effects and regional growth: evidence of the Spanish case

Abstract

The impact of transport infrastructures on the economic growth of both regions and sectors, distinguishing among modes of transport, is analysed. An attempt is also made to capture the spillover effects associated with transport infrastructures. Two different methodologies are used: the first adopts an accounting approach based on a regression on indices of total factor productivity; the second uses econometric estimates of the production function. Very similar elasticities are obtained with both methodologies for the private sector of the economy, both for the aggregate capital stock of transport infrastructures and for the various types of infrastructure. However, the disaggregated results for sectors of production are not conclusive. The results confirm the existence of very substantial spillover effects associated with transport infrastructures.

Acknowledgements

The authors acknowledge the comments of two anonymous referees and the financial assistance of the Instituto Valenciano de Investigaciones Económicas (IVIE), Ministerio de Ciencia y Tecnología (projects SEC2001–2950 and SEC2002–03375) and the Oficina de Ciencia y Tecnología de la Generalitat Valenciana (project GRUPOS03/123).

Notes

Correspondence Address: Pedro Cantos, Departamento de Análisis Económico, Edificio departamental oriental, Universitat de Valéncia, Avda. de los Naranjos, s/n, E–46022 Valencia, Spain. Email: [email protected]

For a wide‐ranging bibliographical review, see Banister and Berechman (2000).

All these studies estimated the impact of public infrastructures, whereas the present study aims to estimate the effect of transport infrastructures, a magnitude clearly lower than the total of public infrastructures.

Although motorways are privately owned, they have been included with the stock of roads.

However, it is very important to take into account that De la Fuente also introduces human capital into the estimation.

The importance of the capital stock in roads obtained at the aggregate level for the private sector is a consequence of the positive results obtained for all sectors except construction, and of the major importance of these sectors in the total of the private sector (on average for the period analysed, 30% of private GAV is generated by industry and 41% by services).

The reason for taking 1983 as the reference year is that it marked the start of a process of heavy investment in transport infrastructures, particularly roads.

The argument is similar to that used in Holtz‐Eaking and Schwartz (1995).

Goerlich and Mas (2001), after comparing elasticities of public capital for regional and provincial levels of aggregation, also corroborate that greater dissagregation, such as the provincial level, implies a reduction of the elasticity associated with public capital, thus showing the importance of the spillover effects.

To calculate the weightings matrix on the basis of the importance of trade flows, the information supplied by the Instituto Nacional de Estadistica (INE) is used for interregional trade flows, for which a source‐destination matrix is available.

Alternatively, it is also possible to estimate separately the effect of the region's own capital stock (T) and the effect of the (weighted) capital stock of the rest of the regions (TS). However, the estimation loses precision because of the high number of regressors and the consequent problems of multicolineality.

As one of the referees points out, if better transport infrastructures between regions lead to changes in trade patterns, then using trade patterns as weights could introduce an endogeneity. For this reason, the estimator of instrumental variables is used.

Additional information

Notes on contributors

PEDRO CANTOS

Correspondence Address: Pedro Cantos, Departamento de Análisis Económico, Edificio departamental oriental, Universitat de Valéncia, Avda. de los Naranjos, s/n, E–46022 Valencia, Spain. Email: [email protected]