Spatial economic impacts of transport infrastructure investments

Estimating the spatial economic impacts of transport infrastructure is an unsolved issue that has plagued economic science for a long time. This paper will give an overview of the basic problem, the contributions of the empirical literature, the modelling approaches used until now, a recent application, and a perspective on how to proceed. The basic problem lies in establishing the 'anti monde', that is, the economic development that would have occurred without the investment in infrastructure. This basic problem is burdened with uncertainty about the direction of the impact of new transport infrastructure on the regions or nations affected. As infrastructure reduces the cost of both imports and exports of goods and services, the net effect is not clear. Macro economic research only gives indications about the demand and supply effects of bundles of historical investments in infrastructure. This only with considerable debate about the econometrics and the causality implied by the estimations. Moreover, macro research has only limited value when taking decisions on specific infrastructure projects. Surveys among firms and consumers have various measurement problems, but they also have the advantage of providing ex ante micro information. Among the major disadvantages are strategic answers, sample selection and the inability to capture indirect effects on non-using firms or consumers. Economic potential models for interregional infrastructure and land-use/ transportation interaction (LUTI) models for intra-urban infrastructure provide the best empirical answers to the question of the economic impacts of transport investments. Both approaches, however, show one major defect, namely an unsatisfactory foundation in economic theory. A promising alternative is provided by the new economic geography models that are evolving into more broadly based spatial computable general equilibrium models (SCGE models). Results of a recent Dutch application of a new SCGE model with 14 sectors and 548 municipalities on a proposed transrapid (magnetic levitation train) project from Schiphol Airport to Groningen City, confirm that the SCGE approach has a high potential. Moreover, its implementation appeared to be far easier than was expected. Some aspects and outcomes of this recent study will be discussed in more detail. Despite these difficulties, empirical literature has produced some general qualitative outcomes. First and foremost, both SCGE models and potential models show the same spatial pattern of impacts of infrastructure. On an isomorphic plane, investments in line infrastructure (road, rail, air and waterways) produce butterfly type of spatial patterns with diminishing impacts as the butterflies grow larger. Improvements in point infrastructure (terminals and harbours) produce concentric ring impact patterns. Second, there appears to be general agreement on the fact that new infrastructure produces minimal impacts in countries with abundant infrastructure services. This holds with one major exception. When new infrastructure resolves strong capacity limits in either point or line infrastructure, the local effects will be considerable but mostly at the expense of cities and regions close by.