Primer on congestion pricing

David Fettig | Editor

Published April 1, 1996  | April 1996 issue

Our congested highways, airports and other public facilities pose a dilemma for government officials. While population projections point toward increasing demands on the nation's infrastructure and public facilities, many people oppose their expansion. Economic theory, though, provides a method to address this dilemma.


Most economic transactions are viewed as involving two parties: the provider of the goods and services and the consumer of them. Many transactions, however, affect people other than those directly involved, sometimes negatively. For example, an upstream factory discharges effluent, which reduces fishing downstream. On the other hand, a beekeeper benefits neighboring farmers by incidentally supplying pollination services. The gist of these third-party or spillover effects, called externalities, is that their costs or benefits are not reflected in their prices. Thus, markets will supply too many of the goods with negative externalities and too few of those with positive externalities. To mitigate goods with negative externalities, they should be taxed, and to enhance those with positive externalities, they should be subsidized.

Congestion is a negative externality. As additional people use a facility or service, the quality of service received by all users declines. Consider a fire station, a school, an airport or a sewage disposal plant. One extra consumer after a certain point is reached will dilute the quality of service provided to other consumers.

Congestion amenable to pricing

When drivers use a highway, they do not pay the costs that they may be imposing on other drivers. Ideally, traffic should flow smoothly at the speed limit. But as traffic increases, eventually the addition of one more vehicle will slow the flow and increase the travel time of other vehicles. At this point congestion (the negative externality) begins.

Cost curves illustrate this development (see Chart 1). The vertical axis measures the costs per vehicle mile, and the horizontal axis measures trips per hour. The curve, labeled private costs PC, includes the costs drivers impose on themselves such as gasoline and oil usage, and wear and tear on automobiles. The other curve, labeled social costs SC, includes PC plus costs that each driver imposes on all others.

On the horizontal axis to the left of No the curves coincide, and cars travel at the speed limit. But to the right of No, congestion sets in, and the two curves separate and slope upward. PC shows how a driver's costs increase as speeds drop below the limit and travel times increase. SC rises at a faster rate than PC because the additional drivers increase the gasoline, oil, time and other costs to all other drivers. Supply curves are another title for these cost curves, for they show how costs change as trips per hour increase.


Highway usage is where the supply curve bisects the demand curve, labeled D (see Chart 2). From the demand curve's viewpoint, the costs per vehicle mile, C, indicate the prices motorists would pay for a given number of trips on a highway. D is downward sloping, indicating the higher the price the fewer trips taken. Prices also reflect the benefits a driver receives from a trip.

If usage of the highway is free to drivers, then N number of trips will be taken, the point where D bisects PC. Up to that point, drivers' willingness to pay for trips exceeds their private costs, and beyond it they do not. At N, however, travel is underpriced because drivers pay C for trips instead of C**, the price which includes the congestion costs imposed on other drivers. Thus, too many trips are taken.

If drivers are required to pay for the congestion they impose on other drivers, N* becomes the optimal number of trips. This is where D intersects SC. Travel times decrease as N-N* fewer trips are taken. To reduce trips by this amount, a toll equal to the difference between SC and PC at N*, or C*-C1, could be collected, so drivers would pay the full cost of their trips. (In effect, the personal cost curve is shifted to the left.)

Society is now better off with congestion pricing. The increase in the value of well-being from decreased congestion exceeds the value of the loss of N-N* trips (see Chart 3). The area under the demand curve between N* and N, labeled B, approximates the value of the loss of trips. The area under the social cost curve between N* and N (areas A and B) is the value of the increase in well-being from decreased congestion, and it exceeds area B by the area labeled A.

Note, this solution does not eliminate all congestion. Congestion still occurs, but only up to the point that the incremental value of a trip is equal to the incremental social cost, or until drivers feel that the incremental costs imposed by congestion start exceeding the incremental value of using the highway. The toll would impel some drivers to seek other travel options—public transportation or car pools, for example.

Congestion pricing feasible

While the preceding example sketches out how social cost pricing can reduce highway congestion and potentially make all drivers better off, will it work? For about the last three decades economists have advocated congestion pricing to control highway usage, but toll collection was an impediment. Stopping traffic to collect tolls is self-defeating. Recent electronic toll collection devices, however, make congestion pricing feasible. Cars can now be equipped with electronic devices that emit signals relayed to highway monitors, and car owners are billed for their highway usage. Highway and bridge tolls are now collected electronically in, for example, California, Florida, Oklahoma, Texas, France, Italy and Norway.

Even after clearing the practical hurdles to implementing congestion pricing, other obstacles hamper its acceptance. Some critics point out charging for highway use discriminates against low-income people, who spend a greater percentage of their income on transportation than high-income people, and often need to commute during rush hours when tolls would be imposed. Tax credits or other measures affecting the distribution of income could alleviate this hardship and still allow congestion pricing to improve the efficiency of highway usage.

People dislike tolls, and some argue that technology, instead of pricing, could relieve congestion. Technology can improve highway usage, but a phenomenon called latent demand will frustrate efforts to relieve congestion. A reduction in congestion—resulting from technological innovation or from new capacity—encourages others to drive during hours or on highways they normally avoid. In other words, adding more highway lanes, for example, will also add more drivers. Thus, even travel on congested highways does not represent the full, potential demand to use them. Therefore, asking people to pay the cost that their use imposes on others not only provides for rational use, but also eases the dilemma between those who want to expand facilities and those opposed to expansion.