"our rubber-shod sacred cow is a ravenously space-hungry monster"

Road congestion has been a major problem for the past half century, imposing significant costs on the world economy. Numerous strategies have been employed to try and address the problem, but many of these end up doing more harm than good. The mechanisms that cause traffic congestion are complex, leading to some surprising results, and going some way to explain the failure of most attempted solutions.

Costs and causes

Traffic congestion occurs when the number of vehicles attempting to use a road is sufficient to force them to slow down or 'wait in line' for road space. Economists consider a road to be an example of a congested variable-use public good and that the congestion caused by road users is an example of a negative externality. What this means is that while everyone is relatively free to use the road, if too many people use the road it will have an effect on others trying to use it. However, when deciding whether to use a particular road, a driver will pay little attention to the congestion that they will cause. It is this imbalance that leads to overuse.

Arnott and Small offer $48 billion as an estimate of the annual cost of driving delays in the United States. This is based on the amount road users would be willing to pay to avoid said delays. This figure excludes harder-to-quantify costs such as extra fuel, disruption, pollution and road traffic accidents. The total costs are notoriously difficult to calculate, but extend well beyond the road users individually or as a group. When a driver decides whether to make a journey, they will consider the cost it represents to them. This private cost will comprise elements such as fuel, wear and tear on the vehicle, and the opportunity cost of the time spent. The external costs to society, such as congestion, are not considered by the driver. The marginal social cost of a user of a given road is not constant: it varies considerably according to the number of vehicles already using the road. This relationship is shown by the use of a speed-flow curve.

 |         ***
S|            **
P|              *
E|               *
E|               *
D|              *
 |             *
 |           **
 |        ***
 |     ***
 |  ***

This shows the relationship between the average speed of vehicles on a road and the flow of traffic. On an uncongested road the congestion caused by a vehicle joining the road is negligible, and remains so until the road approaches capacity. At this point the average speed drops rapidly until a queue forms, at which point flow and speed both drop off suddenly, returning to the origin as the traffic is at standstill.

Potential remedies

Road building

Increasing the supply of road space by building more roads has been the most widely practiced policy against traffic congestion. Since 1970, the United States Federal Government has spent nearly half a trillion inflation-adjusted dollars on building highways, with $32 billion budgeted for 2005. This is in addition to expenditure by the individual states. This is despite the fact that both theory and empirical evidence demonstrate that "building our way out" of congestion simply doesn't work. The reasons behind this are related to the way drivers choose a route to drive, as well as to some unusual characteristics of the transport market. Wardrop demonstrated how, given a choice of alternative routes, drivers will distribute themselves in such a manner that no one driver can reduce his journey time by choosing a different route. Building on this premise, the Pigou-Knight-Downs Paradox shows that increasing capacity actually creates its own demand. Latent demand in the form of potential drivers who currently choose not to travel on a route consumes all new capacity. Drivers will just redistribute themselves and find a new Wardrop Equilibrium. Unless capacity is increased to more than double the total traffic flow, journey time will not be reduced at all.


An alternative to increasing supply is to control consumption through rationing. This attempts to hold usage of roads below the level at which they become congested. However, as the consumer's cost is unaffected by these measures there is an unsatisfied demand. This leads to non-compliance, as demonstrated by the response to schemes involving episodic bans preventing certain vehicles from entering a city. Variations on these schemes, usually based on a vehicle's number plate, have been employed in several cities, including Athens, Mexico City, Tehran and São Paulo. One survey showed that 22% of Mexico City's vehicle owners had purchased a second number plate or vehicle in order to enable them to drive on days when their primary vehicle was banned. Other schemes based on legislation include "ramp metering", which involves restricting the flow of traffic joining a major road, and variable speed limits which respond automatically to traffic levels. These may be effective at helping existing traffic move more efficiently, but do not address the underlying excess demand resulting from the underpricing of the congestion. High-occupancy vehicle (HOV) lanes are an attempt to increase the capacity of a road in terms of flow of passengers rather than vehicles. This is done by designating certain lanes for the use of vehicles with several passengers. This falls into the same Pigou-Knight-Downs trap attendant to any increase in capacity, but with the added risk that the HOV lane will be under-utilised, while the remaining traffic is restricted to fewer lanes, actually reducing total capacity and increasing congestion.


Subsidies on public transport are widespread. For example, there are very few cases worldwide where passenger rail networks are privately owned and operated without subsidy. Bus travel in the UK tends to be an inferior good, meaning that as people become richer they use it less, except in some cases involving short journeys where the person would otherwise walk. Because of this, subsidies normally aim to increase usage by improving quality, and to allow the operation of routes that would otherwise be unprofitable at any ticket price. Most bus subsidies in the UK are currently in the form of fuel rebates based on mileage. This primarily benefits under-used rural services that would not otherwise operate, but is not efficient in encouraging use in urban areas. These short-distance services are the ones most likely to replace car journeys with a high congestion cost. A change to a per-passenger subsidy, geared according to service quality metrics, would encourage bus operators to invest in improving services and reliability. This would increase passenger numbers, which would in turn allow an increase in service frequency. The resulting drop in waiting times would further stimulate demand, leading to a virtuous circle of improvement. Higher quality modes of public transportation are less subject to the income effects that limit substitution for car travel. For example, those in the higher income social groups ABC1 comprise 67% of London Underground users, as opposed to 50% of bus users. Low income DE groups make up 30% of bus users, but only 10% of the users of the Tube.

Although subsidies can increase usage of public transportation through improvement in service quality, the under-pricing of car travel is such that even at zero ticket cost, it is unlikely that users would switch from the car in sufficient numbers to relieve congestion.


I have demonstrated that traffic congestion (an externality) leads to a inefficient allocation of road resources. In order to efficiently use the roads, the road users must be made to internalise that externality. This can be achieved through the use of a Pigovian tax, a tax on consumption equivalent to the difference between the cost to the driver and the cost to society. This ensures that consumption decisions are made according to the true cost of the action. In our example, this would mean that a driver must take account of the congestion caused by his decision to drive on a road. Unlike most taxes, which impose a deadweight loss on society, a Pigovian tax merely corrects a pre-existing imbalance, and leads to an overall increase in efficiency.

A Pigovian tax to correct road congestion externalities would take the form of road pricing. The reduction in utility caused by the reduced consumption is offset by the revenue from the tax. In many implementations of road pricing, the revenue is earmarked for public transport subsidy, with the aim of providing a stronger effect. There have been three principal barriers to the use of road pricing. The first problem is in ascertaining the optimal rate at which to set the tax. This arises from the difficulty in estimating the true social cost of congestion. Nevertheless, even if a rough estimate is used, it will help move consumption to a more efficient point than would otherwise be the case, so it is a weak argument against road pricing per se. The second obstacle has been technological: the logistical problem of tracking road usage and issuing bills to users. This objection has been declining in strength in recent years, with advances in technologies such as global position systems, radio frequency identification tags and the tracking of number plates using closed-circuit television and optical character recognition allowing automatic monitoring of road usage. The final objection has been political: people don't like bearing the costs of goods that they see as free, and the regressive nature of road pricing causes these costs to fall disproportionately on those with low incomes. The problem had been exacerbated by poor information and a failure to explain the utility generated by the reallocation of the tax revenue. Recently, however, support has been growing for the idea, most probably encouraged by the costs that people can observe as congestion worsens. Ken Livingstone was elected Mayor of London in 2000, advocating "congestion charging" - road pricing for central London. His successful introduction of the charge has helped win support for similar schemes elsewhere.


Traffic congestion is a major deadweight loss to the economies of industrialised countries. It results from road users failing to absorb the full costs of their journey, and is best corrected through methods that internalise that externality. Existing remedies have tried to solve the problem by escalating it with road building. This has been ineffective as it creates its own demand. While structural remedies and targeted subsidies can be beneficial, the only way to address the root of the problem is by charging users for the congestion that they cause.

  • Arnott, R., & Small, K. (1994). "The Economics of Traffic Congestion." American Scientist, 82, 446-55.
  • Atkinson, A. B., & Stiglitz, J. E. (1980). Lectures on public economics. London ; New York: McGraw-Hill Book Co.
  • BBC News. (2000). "Candidates battle over transport." Retrieved 11th December 2004, from http://news.bbc.co.uk/1/hi/uk_politics/701292.stm
  • Boadway, R. W., Wildasin, D. E., Brown, C. V., & Jackson, P. M. (1984). Public sector economics. Boston: Little, Brown.
  • Bristol City Council. (2003). "Traffic and Transport Projects and Schemes: Congestion Charging Proposals." Retrieved 10th December 2004, from http://www.bristol-city.gov.uk/PageRedirector/redirect.html?DTT0082+BG
  • Commission for Integrated Transport. (2002). "Bus subsidy simulation study: Faber Maunsell research report." Retrieved 11th December 2004, from http://www.cfit.gov.uk/research/psbi/faber/
  • Downs, A. (2004). "Why Traffic Congestion Is Here to Stay... and Will Get Worse." ACCESS, 25, Fall 2004, 19-25.
  • Elsom, D. M. (1999). "Urban Air Quality Management: A Global Issue." Retrieved 11th December 2004, from http://www.eetpc.org/Air%20Quality/PAPERS.AIR/elsom.htm
  • Goodwin, P. (1983). Subsidised public transport and the demand for travel : the South Yorkshire example. Aldershot, Hampshire, England: Gower.
  • Goodwin, P. (2002). "Re-launching the 10 Year Plan for Transport (Draft text with permission)." Transport Planning Society Annual Lecture, 22nd July 2002.
  • Highways Agency. (2004a). "Car Sharing Lanes for M1." Retrieved 10th December 2004, from http://www.highways.gov.uk/news/articles/7285828.htm
  • Highways Agency. (2004b). "Active Traffic Management (ATM) Project M42 Junctions 3A-7." Retrieved 10th December 2004, from http://www.highways.gov.uk/knowledge/tcc/atm/index.htm
  • Leach, J. (2004). A course in public economics. Cambridge, UK New York, USA: Cambridge University Press.
  • Maddison, D. (1996). Blueprint 5: The true costs of road transport. London: Earthscan.
  • Pigou, A. C. (1932). "The economics of welfare(pp. II.X§I). London: Macmillan
  • The Budget of the U.S Government. (2005). "Historical Tables." Retrieved 11th December 2004, from http://www.whitehouse.gov/omb/budget/fy2005/pdf/hist.pdf
  • The Economist. (1998). "Commuting: The Unbridgeable Gap." Retrieved 5th December 2004, from http://economist.com/displaystory.cfm?story_id=163351
  • Transport for London. (2003). "London Travel Report 2003." Retrieved 11th December 2004, from http://www.tfl.gov.uk/tfl/ltr2003/
  • Vickery, W. (1963). "Pricing in Urban and Suburban Transport." American Economic Review, 53(2), 452-65.
  • Wardrop, J. G. (1952). "Some theoretical aspects of road traffic research." Proceedings of the Institution of Civil Engineers, Vol. 1, Part II, 325-78.
  • Weisbrod, G. E., Vary, D., & Treyz, G. (2001). Economic implications of congestion NCHRP Report 463. Washington, D.C: National Academy Press.

This is based on an paper I wrote for my microeconomics course at the University of Bath. I've substantially rewritten it to make it suitable for the E2 audience.