CATHARINE JAAP IS A SECRETARY for a lawyer with a law firm in Los Angeles, California. Her workday begins at 9:00 every morning, but she gets herself to the office by 7:30 or 7:40. An eager beaver? No. In fact, she reads the paper until 9:OO. But to arrive at 9:00, she says, she would have to leave her home-“10 miles on the nose” from her office-even before 8:OO because of rush-hour traffic. Instead, she rises early, leaves at 7:15, and spends only 15 to 20 minutes door to door. Leaving work, she must wait until after 5:30 to begin the drive home.
“It’s inconvenient,” she admits. Stores aren’t open in the morning and are often closed by the time she’s going home, so she can’t run errands along the way. Every day, five days a week, she spends at least 10 hours at the office-nearly 11 hours away from home. But “I decided years ago,” she explains, “that that’s how I was going to deal with” rush-hour traffic in Los Angeles.
Across the country, Joseph Dikdan is a vice-president of sales for a large computer services firm in New York City. He lives in Pine Brook, New,Jersey, 26 miles due west of the Lincoln Tunnel. He could take a bus, but it’s cheaper, he’s found, to commute with another person, splitting the cost of gas and a toll into the city every day. “And the bus isn’t any faster,” he adds.
So he leaves Pine Brook a little before 8:OO in the morning. Like others who work in cities and drive during rush hour, he’s resigned himself to express- way congestion as a way of life. “I know how to work around it” to reduce the problem to a minimum, he says. Still on the best of days, it takes him 40 minutes to make it into New York, and all too often travel time mounts to as much as an hour and 10 minutes, which works out to a “leisurely” drive averaging about 22 miles an hour largely on 55-mph roads.
Every workday, stories like these two commuters’ are repeated all across the country. Millions of Americans work in major metropolitan areas. Some of these people take a train, some a bus. A few can walk or bike to work. But most of them, as anyone who drives can attest, take their cars.
The result is that the country’s urban “expressways” have become anything but an express mode of travel. Originally designed to enhance the convenience and mobility of city life, they are increasingly a frustrating source of inconvenience and limitations. Nor is congestion the only problem; urban sprawl, pollution, and noise are also linked with the proliferation of expressways, or freeways, in the last few decades.
Politicians’ and urban planners’ attempts to solve the problem of urban expressway congestion have been a dismal failure. Traditional responses to congestion have been aimed at increasing the supply of roadway space or reducing the demand for it by providing alternative means of transportation. Neither approach has worked.
Until about 15 years ago, the knee-jerk “solution” to freeway congestion was to build more and bigger expressways. The most notable effect of this policy was a dramatic rise in the number of vehicles snarled in rush-hour tie-ups, leading economist Anthony Downs to formulate the Law of Peak-Hour Expressway Congestion: On urban commuter expressways, peak-hour traffic congestion rises to meet maximum capacity.
The reasoning behind this is not complex. Additional roadway capacity adds to the attractiveness of driving relative to other means of getting about the city. New drivers, and those who previously took other routes, are attracted to expressways if their travel time is lowered by using the new facilities. In the longer run, improved road systems, by providing ready access to the city, encourage more urban residents to settle in outlying areas. So in almost every case,
carving out a new expressway or widening an existing one has been found to generate a greater volume of traffic than the planners had projected from prior trends. Sometimes congestion would be shifted from one site to another, but it would never be overcome.
Faced with the failure of the more-is-better approach-and with widespread dismay at “the paving of America”-transportation planners switched gears. Instead of increasing roadway supply, they now try to reduce roadway demand by diverting commuters to fixed-rail rapid transit systems. These projects require enormous capital outlays and heavy operating subsidies, but they are thought to be the only alternative to additional road construction and the attendant increase in traffic, pollution, ‘and noise. The rationale for rapid transit rests on the belief that frustrated commuters would prefer a safe, quiet, relatively cheap ride to work instead of the hassle and lost time of battling the freeways. Unfortunately for the taxpayers in areas that have gone ahead with such plans, this supposedly commonsense belief has turned out to be false.
The Bay Area Rapid Transit District (BART), serving the San Francisco area, was the first modern fixed-rail system designed specifically to lure commuters from the roads. After an investment of over a billion dollars, analysis of traffic flows indicates that BART has succeeded
mainly in drawing riders from buses and other preexisting public transit, without significantly affecting automobile traffic or congestion. The system now requires an operating budget of about $100 million a year, two-thirds of which is drawn from state and local taxpayers’ pockets.
BART is often cited favorably by those seeking to bring rapid transit to other cities. The truth is, however, that it has been a monumental flop in improving the overall efficiency of transportation in the San Francisco region. In fact, the inflexible fixed-route layout is so ill-designed to meet the needs of a modern metropolis that regional planners have proposed land-use controls to force the population to live and work closer to the BART stations!
In the view of such planners, and many others concerned about traffic congestion and automobile pollution, nothing short of a comprehensive network of government controls can solve urban transportation problems. It is the private automobile that is the problem, they charge. If enough of us will not voluntarily renounce its use, then our freedom to choose times, places, and modes of travel-and even to choose where we will live and work-must be restricted.
In the face of the massive and costly failure of the traditional political “fixes” and the drastic implications of envisioned control of people’s lives, it seems warranted to step back and consider an alternative. This alternative has played well in academic journals for years but has been steadfastly ignored by politicians and planners. .It relies on some basic facts about people’s use of any resource-including roadway space.
WHAT IS CONGESTION?
Congestion is a form of queuing, an economic phenomenon that can occur whenever the demand for a good, at a given time and place, exceeds the quantity that is supplied. Some expressway congestion is transitory, of course, arising from interruptions of the traffic flow due to accidents,
road maintenance, or other temporary conditions. Some is site-specific, associated with approaches to airports, shopping centers, and the like. But the most typical and recurring variety is caused by weekday commuters, whose demand for road space regularly outstrips the available supply, often congesting whole metropolitan areas.
On any expressway, only so many vehicles, traveling at 50-55 miles per hour, can pass by a given point in an hour’s time. When more than that optimal number attempt to use a particular stretch of roadway, average speeds are reduced, increasing the time it takes to drive that stretch. In other words, the average cost of roadway use, in time and operating expense, rises when congestion sets in.
There’s a crucial point about the way these costs rise, though, and it can be clarified by a simple example. Suppose that any number of cars up to 999 per hour can drive a particular roadway segment in 20 minutes at normal speed. If an additional driver, car number 1,000, enters the roadway, it will take each car 21 minutes to pass through this stretch. Because the added vehicle causes the roadway to become slightly congested, its driver incurs a time-cost one minute
greater than if he had driven at some other time. One minute is negligible to this driver, of course, and to each of the other 999 drivers whose travel time he has caused to rise by the same amount. Yet the addition to the total time-cost generated by this one additional vehicle is 1,000 minutes (one minute times 1,000 drivers). That’s 16.67 hours!
Enter now drivers number 1,001, 1,002, and so on, and the same thing happens. And the situation gets worse in another way: each additional vehicle causes a greater and greater increase in the time it takes each driver to travel this road segment. The more crowded the roadway already is, the greater the cost engendered by one more driver.
The actual costs, though, are not visible to the additional drivers. Place yourself in the shoes of driver number 1,000. For you, the decision say to drive. now or at some other time when the road space is less in demand is a choice between 20 and 21 minutes. Since you don’t bear the actual additional cost of your decision-the 16.67 hours-there’s no way you can even perceive what it is.
The costs of goods and services are normally reflected in their prices (otherwise the providers would go out of business), and purchases are made only when buyers believe that the benefits to them are at least equal to the prices. The virtue of this system is that resources will be devoted to providing a good or service only so long as the benefits to be gained, as evidenced by consumers’ willingness to pay, are not outstripped by the cost. The happy result, in economists’ talk, is an efficient allocation of resources.
Roadway space, though, is generally provided as a free good. Construction and maintenance, of course, are paid for through taxes and registration fees, but the use of roads is not priced. While there can be little doubt that the benefits gained by drivers 1,001, 1,002, and so on from using the expressway are nowhere near as high as the actual costs generated by that use, those are not the costs they face. The inevitable result is heavier- than-optimal traffic at the times and places of greatest demand-or an inefficient use of this particular resource, roadway space.
The consequences of this inefficiency can be staggering. Daily congestion costs for Los Angeles-area freeways have been conservatively estimated at 100,000 hours. If average driving time is valued at $4.50 per hour, at a typical vehicle occupancy of 1.1, these direct congestion costs come to $118.8 million per year. The greatest share of these costs is concentrated in the weekday morning and afternoon peaks, when traffic is 50 percent greater than during other daylight hours.
Obviously, increased travel time is by no means the only negative effect of congestion. Much air pollution (particularly airborne carbon monoxide) occurs primarily in areas of high traffic density and follows the same spatial and temporal patterns as expressway congestion. In Los Angeles, 40 percent of all vehicular smog originates during the morning rush hour, and it has been estimated that cutting peak-hour traffic by one-fourth would lead to a seven percent reduction in total smog. Noise pollution, too, is closely associated with heavy traffic and is known to reduce property values near congested routes. The total of these direct and indirect costs has not been, and probably cannot be, measured but could be safely said to amount to several billions of dollars annually, nationwide.
WHY NOT PRICING?
Once the problem of congestion is seen in an economic light-at the times and places of greatest demand, roadway space is underpriced-the sensible solution, no matter how difficult to make popular, is obvious: put a price on peak-hour use of this scarce resource. With nearly every other resource we have found pricing to be an admirable way of smoothly parceling out goods and services among competing users. But with few exceptions decisionmakers remain oblivious to the merits of pricing as a way of reducing congestion.
And it’s not as though no one has been trying to point them in that direction. The pricing approach has been recommended by economists for more than 20 years. William Vickrey of Columbia University, in a 1963 article in the American Economic Review, argued that “in no other major area are pricing practices so irrational, so out of date, and so conducive to waste as in urban transportation.’’ Vickrey noted that private firms commonly charge different amounts for peak and off-peak services (examples: resort hotels, theaters, airlines), and that this principle could be applied to road- ways.
If potential drivers faced not just their own time-costs but prices reflecting the actual costs of their decisions, they would have an informed basis for deciding whether the benefits of a given trip warranted using a central expressway at rush hour. Some automobile trips would be rescheduled, others rerouted, and some forgone altogether as car users found that alternative transit modes were a better buy. And with concrete knowledge of the benefits of various services-evidenced, as in any other market, by what people are willing to pay for alternatives-decisionmakers could establish rational spending priorities for maintenance and new construction and, assuming mass transit were priced-in line with its costs, could guide future investments in alternative transit modes.
Since Vickrey took up the subject, roadway pricing has drawn keen interest from economists at Harvard, Chicago, UCLA, Stanford, and other academic centers. Writers in professional and scholarly publications like the Journal of Transport Economics and Policy have been debating the best means of implementing pricing for several years, having long since agreed on the necessity of such an approach. The Washington, D.C.-based Urban Institute has published nearly a score of studies, dating back to 1973, taking up various aspects of pricing as a solution to congestion.
Yet the steadily mounting theoretical support for roadway pricing has led to no significant applications in this country. This seems especially curious in view of the eagerness with which authorities have pumped billions of tax dollars into poorly conceived boondoggles like BART, which rested on little or no theoretical foundation.
In the few cases where pricing proposals have been put to transit bureaucracies, they have managed to squelch them through Catch-22 reasoning: congestion pricing is only an idle theory, they say, because it has never been applied; and it can never be applied, because it is only an idle theory. This argument not only relies on faulty logic; it conveniently ignores a conclusive demonstration of the utility of pricing in a city half-way around the globe.
Roadway pricing during the morning rush hour has been in effect in Singapore since 1975. A bustling metropolis of 2.2 million people and a quarter of a million cars, Singapore was plagued by chronic congestion in its downtown area when officials there decided to try a pricing system seven years ago. Access to the central city between 7:30 and 10:15 A.M was restricted to vehicles displaying special stickers, priced at about $1.67per day. The response was swift and dramatic. A study conducted by the World Bank reveals that the number of autos using the congested zone at rush hour dropped by 65 percent, reducing overall traffic by 40 percent.
What about the people for whom it was not worth $1.67 to drive in that time and place? Many rescheduled their trips, spreading roadway demand over a wider time span. Cars carrying at least four people were exempt from the charge, and the number of such car pools doubled after congestion pricing went into effect. Others formed smaller car pools to spread the cost of the fee over several riders. The city also encouraged commuters to switch to bus service by setting up park-and-ride shuttle services at various outlying posts.
The results in Singapore surpassed even the most optimistic predictions. But could rush-hour pricing work in America to relieve congestion on urban expressways? Skeptics reply that this isn’t Singapore; Americans love their cars, and a pay-as-you-drive plan wouldn’t achieve anything but political unpopularity here.
Back in 1973 the experts said that the American people wouldn’t respond to rising gasoline prices, either. Only government rationing would get them to cut back on the use of this increasingly scarce resource. But an “amazing” thing happened on the way to the gas pumps:
between 1973 and 1978, the annual increase in gas consumption slowed from 5 percent to 3 percent, and by 1979 consumption was actually falling, a trend that continues today. These figures, it has been found, represent not only a switch to more fuel-efficient cars but
fewer and shorter automobile trips. “The experts” have had to eat humble pie.
Another criticism has been that the technology applied in Singapore-daily or monthly stickers-is too primitive for the United States. Congestion costs in American cities change rapidly with respect to time and place, and the pricing mechanism should be flexible enough to match these changes. Ideally, roadway prices should be continuously variable, so that a vehicle entering a roadway segment would be charged according to how crowded the road is. At a minimum, prices should vary with the congestion that is characteristic of particular roads or road segments and times. Because Singapore did not use such a flexible system, it is sometimes assumed that none exists.
In fact, the literature on roadway pricing is replete with more sophisticated methods of chargi.ig users. Coming up with a workable scheme is probably the least significant obstacle to road pricing. Even a licensing scheme can be made more sophisticated than Singapore’s. In any such arrangement, drivers must purchase in advance and display a sticker in order to travel in congested priced zones. With variations in the color, number and letter codes, and shape, such stickers could allow differential pricing for various typical levels of congestion associated with different zones and times of day.
Such an arrangement has certain advantages compared to other proposed methods: the stickers are easy to use; the system does not require a large initial investment; and its overall costs are estimated to be low. On the other hand, the more it allowed for differential pricing, the more complicated it would be to administer. Also, enforcement, via monitoring of moving vehicles, would pose difficulties. (With 22 points of entry into its priced zone, Singapore found that one or two extra policemen were required at each point, but this was for a simple two-sticker [daily and monthly] system.)
Another method already familiar to drivers is the use of conventional toll booths. In order to enter a priced zone, vehicles have to come to a stop and pay a fee. Disadvantages are the delays associated with stopping and the large initial investment for toll plazas. There are also difficulties with charging by time-of-day: drivers anticipating an impending decrease in charges might slow down or stop and wait, thus causing congestion; but if this obstacle is met by gradual changes in the toll, administration and payment of tolls become cumbersome.
A modified toll-booth system could overcome this problem by using magnetic cards and monthly billing, much like telephone or utility bills. But the system would still be saddled with high initial capital costs and the need for drivers to stop, both disadvantages counting heavily against its feasibility for a modern urban expressway system with a huge number of entry points.
Less familiar to the public but well analyzed by those who have studied road pricing are various metering options by which charges can be applied for driving at specific times and locations. These include on-vehicle meters and Automatic Vehicle Identification systems. On-vehicle meters operate much like a taxi meter. One type of device would require the driver to activate the meter upon entering a priced zone by “plugging in” a timer purchased in advance and good for so many hours of peak-time travel. Another type would be activated automatically as the driver entered a priced zone, via a signal from an electronic device embedded in the road. Payment would be similar to the system used for metered mail, with the meter periodically being presented for payment and adjustment. In both systems, enforcement could be effected through pilot lights on the outside of the car indicating that the meter is running.
The plug-in meters pose potential safety problems because they would have to be activated by drivers of moving vehicles. And while the automatically activated meters have the major advantage of allowing pricing flexibility along one continuous trip, the method of payment would be inconvenient for the road user and there is a potential problem with fraud because of the possibility of drivers tampering with the meters.
Automatic Vehicle Identification (AVI) systems use off-vehicle monitors at pricing points to identify cars as they enter and leave zones. A central computer records use by vehicle, time, and location. Itemized billing can then be issued on a monthly basis. Technologies pro- posed include roadside camera scanning of optical windshield labels (such a system has been used on railroad cars), laser beams, microwaves, and scanners embedded in the road to monitor low-frequency signaling devices mounted on the underside of automobiles.
DRIVING WITH AVI
The latter AVI system is considered the most promising technology for congestion pricing. Electrical induction loops in the road would pick up a signal from a small radio transmitter placed underneath the vehicle. The signal would give the vehicle identification code and this, along with the time and location, would be transmitted to a computer and user charges run up on a monthly basis. It would thus permit instantaneous identification and charging of moving vehicles without disrupting the flow of traffic and with minimal involvement bythe user once the transmitter is installed.
Although there have been no tests to evaluate a full-blown AVI system in use, the technology has worked well for automatic truck and bus toll collection in New York and New Jersey. And tests
have shown this AVI system to be highly reliable and unaffected by weather conditions, lateral movement of vehicles across lanes, and other circumstances that can impair some types of metering. Because AVI would allow for many variations in rates by time and location, it is superior to stickers, toll collection, or other methods of charging at fixed rates. Aside from the advantages, from the point of view of pricing theory, of being able to relate charges to different conditions, there is another plus: as economist Ward Elliott, of the Rose Institute of State and Local Government of Clarement McKenna College, has pointed out, “Many regular road users, if they have to pay a congestion charge, would prefer a precise one to a rough one for the same reasons that they prefer itemized tax deductions to the standard deduction. Greater precision could save them money with lower charges for near-peak use, minor direction changes, shorter or more peripheral trips, separate charges for morning and afternoon peaks, and so on.”
The economics of AVI compare favorably to other pricing methods and suggest a very positive ratio of benefits to costs. Prototype signal units, available from several manufacturers, have cost from $75 to $120. Unit costs in large-scale production have been estimated at $25 to $45. The cost could be charged directly to motorists at the time of installation or recovered from revenues.
A scanner system costs about $32,000 per four-lane pricing point. Overall, an extensive AVI network of 200 pricing points serving 500,000 vehicles would entail an initial investment of perhaps $60 dollars per car, with operating costs of five cents per vehicle trip. The exact structure of charges would depend on demand patterns in specific locations, but the optimal price would undoubtedly exceed operating costs and should yield a reasonable return on the capital investment.
Implementing AVI would involve some administrative and enforcement problems, mainly in providing for road use by unequipped vehicles and in broadcasting the rate structure. Signal units should be readily available, perhaps at service stations, so that newcomers to the area could have one installed without delay and worn-out units could be easily replaced. A supplementary sticker system could handle out-of-area motorists, infrequent drivers, and those who, for reasons of privacy, preferred not to have their trips recorded on a computer. Rate schedules could be published in newspapers and posted along routes.
Sudden changes in the rate structure at peak hours would have to be avoided, or congestion would simply shift to the times just before the price goes up. Similarly, the value of efficient pricing on expressways would be limited if alternate routes on city streets remained unpriced. (An excessive diversion of traffic to surface routes might actually decrease the overall efficiency of the transit network.) The most practical solution in many cases would be to limit peak-hour street access to vehicles bearing specially coded stickers. This was the method successfully used in Singapore and could be price-coordinated with an AVI expressway system.
In no case would these technical details surpass the complexity of operating a rapid transit system, rationing roadway use, or adopting other bureaucratic approaches being seriously advanced today. AS pointed out by economist William Vickrey, who kicked off the debate on road pricing back in 1963 with his discussion in the American Economic Review, the obstacles to road pricing raised by skeptics “can be overcome by the exercise of a modicum of ingenuity such as would be normal for those genuinely interested in developing the possibilities . . .rather than merely defending the status quo.”