Do Public Transport Improvements Increase Agglomeration Economies? A Review of Literature and an Agenda for Research

Abstract

Public transport improvements may increase economic productivity if they enable the growth and densification of cities, downtowns, or industrial clusters and thereby increase external agglomeration economies. It has been argued that the potential agglomeration benefits are large; if so, understanding them better would be useful in making funding decisions about public transport improvements. We reviewed theoretical and empirical literature on agglomeration as well as a small number of articles on transportation's role in agglomeration. The theoretical literature is useful in understanding possible avenues by which transportation improvements might affect agglomeration, although there is little discussion of public transport specifically. Relevant empirical studies tend to focus on metropolitan regions and use a generalized measure of transportation cost. But public transport impacts on agglomeration are likely to be different from road investment impacts. We identified several ways of conducting research building on this literature that would help evaluate the agglomeration impacts of public transport proposals: tracing the links between transport, agglomeration, and productivity; better motivating research using theories of agglomeration mechanisms; taking scale and redistribution into account; exploring the functional form of agglomeration economies; accounting for endogeneity in model structure; and considering development context.

Acknowledgements

This paper is based on an ongoing study funded by the Transit Cooperative Research Program of the US National Academies entitled “Methodology for Determining the Economic Development Impacts of Transit Projects” (TCRP Contract H-39). Niels Voorhoeve provided research assistance. This paper draws upon material prepared by the authors for an interim report for that project (Chatman et al. 2009) for which Lars Rognlien, Kaan Ozbay, Peter Bilton, and Niels Voorhoeve were co-authors and to which Daniel Graham, Deva Deka, and Joseph Berechman also contributed. The authors thank three anonymous reviewers for their useful feedback that improved this paper.