EconSouth - Second Quarter 2008

Photo of Chris Cunningham
Chris Cunningham is a research economist and assistant policy adviser at the Federal Reserve Bank of Atlanta. He walks to work. Clearing the Roadways: The Case for Congestion Pricing

In 2005 the average American lost 38 hours—almost an entire work week—to traffic delays at an estimated annual cost to the economy of $78 billion. Two of the 10 most traffic-delayed cities in the United States are in the Southeast—Miami, at No. 5, and Atlanta, at No. 6. The underlying cause of congestion is readily apparent: More vehicles are using the roads than the roads were designed to accommodate.

The reason drivers continue to crowd onto the roads is that this limited supply of road capacity is available for free. In some areas, road capacity is a scarce resource; if it is given away for free, it will be overconsumed to the detriment of all. Economists refer to this phenomenon as the "tragedy of the commons." Each additional driver on the road slows down all the other drivers but does not "internalize" the cost of the lost time and expense suffered by the other drivers when deciding to make a trip.

Rationing is not the answer
In the early days of car travel, when roadway systems developed, the relatively high cost of cars and fuel served as an imperfect check on car travel generally. But as cars have become cheaper (in real terms) and development has followed new highways, congested roads have become endemic. Until recently, little could be done to address the problem because of the cost of monitoring and administering tolls over entire road networks.

A few cities attempted to alleviate congestion by rationing road access. For instance, Mexico City and São Paulo implemented a scheme that limits access to the central city on a given day based on the last digit of a car's license plate. In the United States some local governments, including greater Seattle, have attempted to keep demand from outstripping the supply of roads through a land use regulation called "concurrency." In this approach, new subdivisions are not permitted until road capacity—often financed by impact fees levied on the development—has been expanded to meet the additional need.

Both types of initiatives are a form of rationing, one explicitly rationing road access and the other indirectly rationing developable land. And both approaches have had only limited success. In Mexico City, many residents have simply bought a second car to double their access to the city. And in Seattle, concurrency regulations have had little effect beyond improving traffic on the arterial roads nearest to the restricted subdivisions. The impact of new housing on the broader road network has gone unchecked.

Even if these measures had worked in reducing congestion, rationing is a poor way to allocate a scarce resource. For example, a truck driver with a load of perishable produce sits in the same traffic jam as a teenager going out for a burger. The trucker, who is losing money by the minute, would be happy to pay the teenager to take the bus or even stay home, and the teenager, for the right price, would be happy to accept. Unfortunately, even if such an impossibly complex market could be formed, it couldn't prevent other drivers from entering and recongesting the road.

So now everyone sits in traffic and thus everyone loses. One could randomly kick a certain number of cars off the freeway (a la Mexico City) to speed up traffic. But ideally policymakers could target those individuals or groups who derive the least benefit from driving and pay them to stay home, carpool, take public transit, or change where they live, work, or shop.

Goodbye to the free ride
Two new strategies for checking the overconsumption of road infrastructure are gaining increased attention. These strategies—congestion pricing and a limited version of it called high-occupancy toll (HOT) lanes—are driven by spiraling congestion costs, a dwindling Highway Trust Fund, rising urban land prices (which limit road-widening projects), and the falling cost of information technology. Congestion pricing is, in its purest form, a variable rate tariff that rises as needed to deter enough drivers to keep traffic flowing on a particular road; in effect, the state becomes the market maker.

The most recent experiment with congestion pricing was implemented in central London in 2003. Drivers now pay a fee of £8 (about $16) a day to enter or drive within the city's core between 7 a.m. and 6:30 p.m. Since this fee was implemented, vehicle trips into central London have fallen by 20 percent and public transit use has increased by 13 percent. As a result, peak period congestion has dropped by 30 percent and travel times by 37 percent. Revenue from the fee has been directed to public transportation to serve the increased ridership there.

On March 31 of this year, the New York City Council voted to implement a congestion charge—an $8 toll for a private vehicle, for instance—for all trips into Manhattan south of 60th Street between 6 a.m. and 6 p.m. But the plan, which was expected not only to alleviate congestion but also to raise more than $400 million a year for transit, failed to pass in the state legislature mainly because of opposition from representatives in the outer boroughs.

Congestion pricing is HOT
A less comprehensive form of congestion pricing is the use of HOT lanes. These lanes charge a fee for single-occupant vehicles to drive in high-occupancy vehicle (HOV) lanes (while HOVs still use the lanes for free). The plan keeps existing lanes untaxed and thus congested but provides a fast travel option for those willing to pay for it. The toll charged to drive in the HOT lanes changes dynamically with the amount of congestion. As the lanes become more congested, the toll goes up. First initiated in Southern California in the mid-1990s, HOT lanes are being opened across the country, including in Florida, which will soon introduce them on I-95 between Miami and Fort Lauderdale.

Variable-rate tolling holds the promise of matching demand with supply with the smallest loss in social welfare. It should also inform policy. If drivers are willing to pay more to avoid congestion, then that willingness would suggest that more roads should be built. On the other hand, if only a modest toll deters enough drivers to speed up travel, then perhaps our existing road network is sufficient after all.