Comments

Richard Voith: This ambitious paper attempts to shed light on the determinants of new rail investment, the likely impacts of these investments on transit use, and their possible welfare impacts. By looking across cities, the authors can search for systematic factors that determine the relative success or failure of this type of infrastructure spending. This is significant, as there has been little serious cross-city analysis of rail transit investment. In addition, the application of a differences-in-differences approach to measuring the ridership impact of new transit supply provides insight at the microlevel of this type of investment on overall transit use.

The dominant view of economists has been that rail transit investments generally have been ineffective and expensive, and the benefits do not justify the costs. Baum-Snow and Kahn present evidence that can be interpreted as supporting the dominant view. In particular, the authors argue that: 1) older cities with existing rail transit experienced large declines in transit market share in the 1970–2000 period, although there are notable exceptions; 2) investments in new rail transit have failed to stem the decline in transit market share in most cities, again with some exceptions; 3) the primary impacts of new rail transit investments are to cause transit mode shifts from bus to rail within the transit market; and 4) the primary reason that many cities have undertaken new rail investment has been because federal subsidies have significantly lowered rail investment costs to state and local governments.

While the cross-city approach taken by Baum-Snow and Kahn is an interesting and useful perspective, actual public transit markets are intensely local, with specific characteristics of the transit service, markets served, and investment in alternatives ultimately determining the transit market share. This local perspective is lacking (understandably) in the Baum-Snow and Kahn approach, and I argue it leads to some excessively negative conclusions regarding rail transit investment.

I examine four areas that could materially affect the interpretation of the Baum-Snow and Kahn analysis: differences in alternative investment in alternative transportation (roads) across cities and neighborhoods; rail transit operating characteristics that differ significantly in quantity, quality, and scope; households, as well as household preferences and characteristics, can change over time and are affected, in part, by transportation and other public investments in neighborhoods; and the focus on market share obscures the performance of rail transit in markets where it is a viable option.

Alternative Transportation Investment

Baum-Snow and Kahn note that more than $25 billion has be expended on rail transit investment in the sixteen cities building new rail transit systems from 1970 through 2000. To put this in perspective, $128.5 billion was spent on highways in 2000 alone.¹ Extensive investment in highways affects transit ridership in three ways. First, it often competes directly with rail transit services for delivery of people to the central business district (CBD). In Philadelphia, for example, parts of interstate highway I-95 that run through central Philadelphia were completed during the 1970s at a cost of $500 million. An additional $200 million was spent on I-676, which also runs through Center City Philadelphia. These highways competed directly with existing rail transit services in the city.² It is not surprising, therefore, that Baum-Snow and Kahn find sharply declining market shares for transit in Philadelphia.

Second, extensive highway investment lowers vehicular transportation costs, especially to locations outside of existing CBDs, which in turn encourages decentralization of employment. As noted in the Baum-Snow and Kahn analysis, employment centralization is a key factor affecting transit market share. Again in Philadelphia, significant new highway construction, notably I-476 and other improvements west of the city, shifted the relative attractiveness of employment and retail locations away from the older, more centralized city. Finally, the geographic expansion associated with new highway investment generally places additional demands on public transit resources that typically lowers the average quality and quantity of transit services in high density, high ridership areas, while increasing the geographic scope of services. Since most transit providers are regional in their organization and governance, as activity shifts outward, resources are redeployed to serve...