29.9.06

SEPARATING EQUILIBRIUM. There's an instructive bull session in progress at 11-D on the merits, or lack thereof, of high-occupancy-toll lanes, inspired by a Dissent article suggesting that those lanes don't really solve the metropolitan transportation problem.
What to do about traffic is a political issue of growing importance. The cost of road-building is increasing, and a growing body of research shows that sprawl development stimulated by new highways quickly makes new roads just as congested as the old ones. Should congestion nevertheless be relieved by continuing to add more road capacity? Or should investment be shifted toward mass transit?
There's something called the Law of Peak Expressway Congestion that suggests road improvements divert traffic from other roads, until travel times are the same on the improved road as they are on the unimproved roads. (Indifference at the margin, anyone?) An elaboration of the law suggests that additional capacity shortens the duration of the peak subject to the same indifference condition, which makes sense as long as the total volume of trips stays the same, but a shorter crush hour serves as an inducement for more people to relocate.

The next few paragraphs of the article summarize a number of possible policy contradictions that affect both road advocates and transit advocates. I want to focus on the efficient pricing of high occupancy toll lanes.
The toll roads that Samuel champions have been heavily promoted by free-market transportation thinkers, and they have attracted growing interest from policy-makers. A confluence of forces puts tolls at the center of the current transportation agenda. As suburbia spreads out, people must drive farther to reach their destinations, so cars spend more time on the roads. The cost of highway construction has exploded as suburbs spread, with price tags often approaching two or three billion dollars even for roads that serve only one sector of a metropolis. State highway departments, squeezed between the growing costs of maintaining existing highways and the difficulty of raising taxes in a conservative political environment, find their ability to add road capacity falling farther and farther behind traffic congestion. Toll roads seem like an obvious solution, one consistent with the broader political climate that prefers user fees to taxes and favors privatization of public services.
The article neglects to note: and the reluctance of state or county highway commissions to use their own money to expand the road networks, and the reluctance of Congress to release moneys from the Highway Trust Fund in order to make the national government's fiscal house look less disorderly, but let's abstract from that and focus on efficiently self-financed toll roads.
In principle, the new emphasis on tolls is a long-overdue move in the direction of a more rational transportation system. The concept is unassailable; if drivers pay the full cost of the roads they use, environmentally damaging and economically costly overuse of the automobile will be discouraged. Funding better mass transit with toll revenues can advance both social equity and environment protection. These ideas have long been close to the hearts of transit advocates and environmentalists. But toll road proposals do not always accomplish in practice what tolls promise in theory. Many tolling schemes now under discussion would preserve and expand the highway lobby's subsidies rather than curbing them.
Here's a related point from testimony by Douglas Holtz-Eakin as director of the Congressional Budget Office.
If a good or service is provided free of charge, people tend to demand more of it--and use it more wastefully--than they would if they had to pay a price that reflected its cost. Hence, congestion pricing is premised on a basic economic concept: charge a price in order to allocate a scarce resource to its most valuable use, as evidenced by users' willingness to pay for the resource.(4)

Introducing congestion pricing on a crowded highway--that is, charging tolls that are higher during peak times of the day and lower during off-peak ones--has two economic effects. First, it dampens demand for the highway during the most congested periods by inducing some motorists to alter their travel plans. Some drivers will be able to modify their schedules so they use the road at less busy times. Others will find alternative routes or switch to public transit. Second, continued demand in the face of appropriate congestion pricing serves as a signal for additional investment in road capacity.(5)
But there are efficient and inefficient ways of so doing. Oliver Williamson's 1966 peak load pricing article (JSTOR) lays out the argument with some glorious Lagrangians and some intuitive diagrams. The high-occupancy toll problem (and the related problem of selling cuts in the roller-coaster line) is one in which the provider is simultaneously providing a premium service at a higher price and a congested service at a lower cash outlay, but with the people who choose the cheaper congested service incurring disutility. As a modeling exercise, the marginal commuter is indifferent between the marginal utility adjusted by the higher price of the premium service and the marginal utility adjusted by the lower price of the congested service (there are some additional subtleties involved in avoiding division by zero.) The Dissent article correctly flags a potential inefficiency.
This is especially true of the latest fad among the free marketeers, what are known as express toll lanes. These are pay lanes added to existing highways that currently don't charge tolls. Toll rates vary from hour to hour, increased at times of heavy traffic in such a way that the toll lanes never back up. The main advantage of this procedure is that the driver who pays the toll is guaranteed a fast trip; on the busy suburban highways where these lanes are under consideration, there is so much traffic that simply widening the road would not get rid of congestion. Proponents argue that express toll lanes give the consumer more choice than building additional free lanes — when you need to get somewhere in a hurry, you pay the toll; when your time is less valuable, you don't.
"More choice" is not equivalent to raising sufficient moneys over the cycle of use to cover the incremental cost of expanding the network, which is one of the constraints in Professor Williamson's Lagrangian. And thus Dissent claims a flaw in the pricing system.
These survey results suggest that the "Lexus lanes" moniker is well deserved. Who uses pay lanes is mostly determined by income. For most of the people in the free lanes, consumer sovereignty is a fiction. They haven't made a voluntary decision that their time isn't worth the price of a quicker commute. They are sitting in traffic jams because the toll exceeds what they can afford to pay.
Consumers optimize subject to budget constraints, indeed.
Even if this is true, toll lane supporters respond, the lanes still benefit lower-income drivers. Those who can't afford to use the new lanes benefit from the added road capacity tolls pay for. With the wealthy on the new lanes, fewer drivers are squeezed into the free lanes, and everyone has a faster commute. The argument is logical, but it does not fit the facts. It turns out, excepting rare circumstances, that express toll lanes added to existing highways cannot raise anywhere near enough revenue to pay for their construction cost.
That, however, means the tolls have not been designed in such a way as to satisfy the incremental cost constraint. (It does not follow that the tolls are too low. A lower toll and a somewhat lower speed in the carpool lanes may be the efficient outcome.)
In such circumstances, express toll lanes only modify the failed policy of subsidized highways so as better to preserve it. When it becomes impossible to keep traffic moving freely on all lanes, express lanes give an affluent minority the open roads that can no longer be provided for everyone. Tolls in this scheme are primarily an allocation mechanism, and only incidentally a source of revenue. Their purpose is to deter those less able to pay from using the new lanes. Only those wealthy enough to afford the tolls bypass the traffic jams, but everyone who is backed up on the free lanes gets to pay the bills.
In other words, Dissent concurs with the Paul Weyrich wing of "conservative" transportation policy that sees highway projects as a form of socialism, complete with shortages and shoddy services. The article goes on to advocate different forms of congestion charges, although it omits mention of efficient peak-load pricing of the entire road network as an alternative.

No comments: