15 September 2019

The Congestion Relief Benefit of Public Transit: Evidence from Rome

In a recent CESifo Working Paper, Martin W. Adler, Federica Liberini, Antonio Russo, and Jos N. van Ommeren analyze traffic in Rome, Italy based on new data. They find out how travel time can be reduced for both private and public transport participants. A summary and the full version of this interesting working paper are available on our website.

Image, Featured Paper, CESifo, Public Transit, Rome

Key issue

Most cities devote substantial resources to subsidizing public transport. Economic theory tells us that because the external costs of road congestion are generally not reflected in the price of car travel, subsidies for public transport are justified as a second-best instrument to reduce congestion. However, the relationship between public transport systems and congestion is not fully understood. For instance, how much congestion can be reduced by providing public transport is still an open question. Furthermore, expanding the provision of public transport may generate a virtuous circle because bus users may benefit from less congestion themselves. However, this effect has never been measured. In addition, little is known about the effects of dedicating road space to public transport vehicles to insulate them from traffic.

Approach and methodology

We estimate the effect of public transport supply on travel times of cars (and motorcycles) as well as bus users in Rome, Italy. Our methodology exploits hourly information on public transport service disruptions during strikes. Furthermore, we combine car and bus traffic data to estimate the effect of reallocating road space to bus lanes.

Key findings and conclusions

We find that the effect of public transport on road congestion is sizeable. Furthermore, bus travel time gains account for an important share of this benefit. A ten percent reduction in public transit supply increases the travel time of car and bus users by about 1.6 percent in the morning peak. In monetary terms, this reduction produces extra congestion costs that are equal to at least 49 percent of the operator cost reduction from the downscaling. These findings imply that the current level of subsidies in Rome (which, at about 75 percent of operating costs, is already large) is less than the optimal level. We also find that dedicated bus lanes reduce bus travel time by at least 29 percent. Although car travel time increases, we can single out several mixed-traffic roads where introducing a bus lane would most likely increase welfare, suggesting that bus lanes are undersupplied in Rome. These results have policy implications for large cities in emerging and less developed economies, where buses are the mainstay of the public transport system.