Surface Transportation Innovations Newsletter

Surface Transportation Innovations #32

Topics include: Brookings' new study on highway spending vs. congestion; phasing out toll plazas; evacuations and toll roads; moderate growth projected in vehicle miles traveled; and news notes.

In this issue:

Why More Highway Spending Hasn’t Led to Reduced Congestion

I’m surprised that the anti-highway crowd hasn’t yet latched onto a new working paper from the AEI-Brookings Joint Center, since it can easily be read as supporting their contention that “we can’t build our way out of congestion.” Brookings fellow Cliff Winston and UC Berkeley economist Ashley Langer created an econometric model that uses data from 1982 to 1996 for 74 urban areas to see if there is any kind of statistical relationship between the level of highway spending and the cost of traffic congestion (as measured by the Texas Transportation Institute). Their conclusion is that, “on average, one dollar of highway spending in a given year reduced the congestion costs to road users only eleven cents in that year.” Therefore, they conclude that increasing highway spending is not a good way to reduce congestion, so we should support road pricing.

Well . . . several clichés leaped to mind as I read the working paper (#06-11, May 2006, soon to appear in the Journal of Urban Economics). The least critical of these is “the devil is in the details.” What the study measured and how it did so raise serious questions about the overall conclusion. And another finding of the study, far more important as I see it, got all too little emphasis.

Let me start with the huge, glaring flaw. The data on highway spending do not distinguish between capacity additions and other spending such as maintenance, repaving, repair, and replacement of existing roads and bridges. Without that information, it’s impossible to tell what the spending was used for, and hence whether it had any likelihood of reducing congestion to begin with. The paper makes a weak argument that repaving a road will reduce congestion somewhat, but that’s hardly likely to be noticeable on freeways where traffic is chronically snarled in congestion measured at Level of Service E or F.

Second, the statistical technique compares congestion reduction in year X with spending in year X. But if there is any relationship in the same year, it’s likely to be that spending in the year of construction leads to increased congestion that year, due to that very construction, other things being equal.

Third, even if the report were measuring the correct thing (amount of congestion several years after the completion of projects that add new capacity), it would still make no sense to simply look at reduced congestion costs for one year. The nature of capacity expansion is making large up-front investments to produce long-term, ongoing gains. So if a more sophisticated analysis found, say a 35 cent reduction in congestion per dollar spent in the first year after a new set of lanes (or a new toll road) opened, that would be the first of a stream of benefits extending for a number of years, until continued growth in travel eventually used up all the congestion reduction produced by that particular investment in new capacity. Nobody would judge a capital investment in long-lived infrastructure simply by what it produced in its first year of service.

The most useful part of this flawed paper is its final section, Discussion and Policy Implications. There the authors refer to a previous paper by Winston and Shirley, which argued that the reason for the declining (measured) rate of return on highway investment in recent decades is that “highway spending is compromised by inefficiencies related to pork barrel politics, by slow and inappropriate responses to demographic changes, by excessive maintenance expenditures caused by poor road design, and by inflated labor costs attributable to the Davis Bacon Act.” And in fact, even in the current paper, with its flawed analytical techniques, Winston and Langer report that when they modeled the reallocation of highway spending among and within states so as to minimize user costs (i.e., removing all kinds of political constraints), their flawed measure of same-year congestion cost reduction per dollar spent leaped from 11 cents to 25.3 cents.

So be prepared when the “don’t build more highways” crowd touts this paper as proving their point. It does no such thing.

» return to top

Phase Out-Rather than “Fixing”-Toll Plazas

In April the National Transportation Safety Board issued a report calling for national design standards to reduce accidents and deaths at toll plazas. The report stemmed from the agency’s investigation of a chain-reaction crash at an Illinois toll plaza that caused eight deaths. The NTSB report cites figures from Illinois, New Jersey, and Pennsylvania showing that one-third to one-half of accidents on toll road occur at the toll plazas.

There’s no question that traditional toll plazas are the source of many crashes. But given that the development of national safety standards takes years, and coming up with the funds to redesign and rebuild many hundreds of toll plazas could take decades, it’s not clear that such an effort makes sense. Not when electronic toll collection has made toll booths and toll plazas obsolete. Instead of wasting money on rebuilding an obsolete type of facility, we’d do far better to speed up the conversion of toll roads to fully electronic open-road tolling.

One of the leaders in going down that route is the Miami-Dade Expressway Authority (MDX). Last month its board approved an ambitious plan to phase out all its toll plazas over the next six years, replacing them with portal-to-portal automated tolling. The planned system is similar to that in use for years on Toronto’s Highway 407ETR, the Melbourne (Australia) CityLink, the Cross-Israel Highway, and the new urban toll roads in Santiago, Chile. (All of these, incidentally, are investor-owned toll roads developed under long-term concessions.) On these systems, customers can pay either via windshield-mounted transponders or via video billing (in which their license plate number is recorded on video and used to create a bill for the miles driven).

To implement the system in Miami, MDX will outfit every entrance and exit ramp in its system with overhead gantries, which will house both the antennas that communicate with SunPass transponders and video cameras. As in Toronto, the rates will be higher for video tolling, which costs the agency more per transaction. The six-year change-over will take place in stages, eventually converting all five of MDX’s toll roads.

In addition to eliminating the hazards and inconvenience of toll plazas, the plan will be fairer than the old system and also a financial boost for MDX. That’s because the old system is an “open” toll road design, in which tolls are collected only at barrier plazas. It’s perfectly legal to use various stretches of the MDX system and get off before a toll plaza, paying nothing. MDX figures show that out of the 685,000 vehicles that use some portion of their system on a typical weekday, only 28% pay a toll. The new system will charge lower rates per mile, but because everyone will pay, first-year revenue is projected to be 51% higher than today. This will help pay for the ambitious expansion of the MDX system that is already under way.

The Florida Turnpike Enterprise is also planning to move into open-road tolling, beginning several years from now with its Sawgrass Expressway in Broward County. I can’t think of any good reason for other urban/suburban toll road systems not to do likewise. Why should any of these 20th-century toll plazas still exist 10 years from now?

» return to top

Evacuation, Emergencies, and Toll Roads

Hurricane season began June 1, and that brings to mind last year’s devastating hurricanes in Florida, the Gulf Coast, and Houston. Evacuation of coastal areas is an important factor in transportation planning in these states, and both Florida and Texas have quite a lot of toll road capacity. Planners worry about being able to evacuate huge numbers of people relatively quickly, and increasingly point to the 14-hour traffic jams and gasoline shortages in Houston during the evacuation for Hurricane Katrina.

Here in Florida, I’ve discovered, it’s long been standard practice to suspend toll collections during hurricane evacuations. The rationale is that congestion at toll plazas would only add to the congestion and delays in a process that already takes far longer than desirable. Needless to say, this benefit comes at a significant cost. Last year, toll-suspension for Hurricane Wilma cost the Florida Turnpike Enterprise $15.4 million; add in the losses from several other hurricanes and the total reached $19.1 million.

And it’s not just hurricanes that lead to calls for suspending toll collection. Just last month state Sen. Jeff Klein in New York State introduced legislation to suspend tolls on all bridges and tunnels during major holidays and holiday weekends, to speed up traffic flow. Transportation planners in Melbourne, Australia are talking about suspending tolls during the Commonwealth Games next year.

It seems to me these people are stuck in a time warp. They have not yet realized that 21st-century tolling requires no toll booths, no toll plazas, and no need to slow down or stop to pay tolls. In fact, what we need during times of very high demand for limited road space is not zero tolls but higher tolls. Let me hasten to add that I’m not proposing such a change until we get rid of today’s toll booths and toll plazas. But let’s start thinking about how our policies need to change when we shift over fully to 21st-century tolling.

One of the reasons last year’s Houston evacuation was such a debacle was that multi-vehicle households evacuated with all their vehicles. That’s why the sheer number of cars on the freeways took planners by surprise and completely overwhelmed the system. Let’s do a thought experiment: Suppose Houston’s freeways were all toll roads, fully equipped for open-road tolling. Let’s further suppose they were operated on value-pricing principles—and that this fact was well-known to Houston residents. Might that not make a significant difference in how many vehicles a family puts onto the road during the evacuation? Incentives, as economists point out, matter.

» return to top

The Case for Moderate Growth in Vehicle Miles Traveled

The past 25 years have been an amazing period in U.S. surface transportation. While population increased 30% between 1977 and 2001, trips made per person increased by 49%, and those trips were longer than before. The net result was that what transportation researchers call total person vehicle miles of travel (VMT) increased 151% during this time period—five times the increase in population.

That unprecedented growth in travel certainly contributed to the massive congestion we experience today in large urban areas. But the key question for transportation planners is this: what will happen to VMT over the next 25 years? A new paper by Steve Polzin of the Center for Urban Transportation Research at the University of South Florida attempts to answer this question. Supported in part by the U.S.DOT, it is available at the CUTR website,

“The Case for Moderate Growth in Vehicle Miles of Travel: A Critical Juncture in U.S. Travel Behavioral Trends” offers both good news and bad news. The good news is that Polzin finds evidence that some of the key determinants of growing VMT have stabilized, suggesting that between 2001 and 2025, VMT may grow only another 50-60%, less than half the rate of the previous period. But the bad news is that even this much growth is likely to result in much worse congestion—unless there is a dramatic change in our approach to the demand/supply imbalance.

Polzin creates two scenarios for future travel growth, one based on assumptions about trends in travel time and the other based more on travel distance. Both use the Census Bureau’s estimate of 22 percent population growth through 2025. A number of trends of the past 25 years appear to be stabilizing: household size, female labor force participation, fraction of zero-vehicle households, mode-share for transit, walking, and bicycling, etc. Less certain are what will happen with the growth in trip length, trip frequency, and time spent on travel. The two scenarios lead to estimates of VMT increase by 2025 of 51% and 60%, respectively—or between 1.74% and 2% per year, both well below historical averages.

But then comes the bad news. Given that many urban roadways are close to the point on the speed-flow curve where traffic flow breaks down into stop-and-go conditions, “minor increases in flow or demand . . . may be resulting in more significant decreases in speed . . . [A]s the urban transportation system approaches high volumes, it is more fragile, and more minor increases in demand may be producing as large or larger responses in terms of increased congestion or slowed speed.”

And that will pose a challenge to transportation planners, who in most places have been de-emphasizing capacity expansion for the last decade or two. “The relatively modest scheduled increases in system capacity expansion may change if public will and financial support increase.” That is already happening in Texas, via explicit policy changes. And it seems close to happening in Georgia and Washington State, as I’ve reported in previous issues.

» return to top

News Notes

Court Upholds Outsourcing in California. Almost alone among high-growth states, California’s DOT (Caltrans) designs all its highways and bridges in-house, thanks to the power of its engineers’ union, Professional Engineers in California Government. In a series of court cases over the years, PECG eventually won a state supreme court decision giving the narrowest possible interpretation to the state constitution’s definition of civil service, such that it precluded outsourcing any function covered by civil service. In response, the Consulting Engineers and Land Surveyors of California proposed a constitutional amendment to explicitly allow the outsourcing of architectural and engineering services for public works. The voters approved it (Prop. 35) in November 2000. Yet Caltrans still does not outsource those services; under former Gov. Gray Davis, it signed a contract with PECG that essentially overrode the provisions of Prop. 35. But on June 14, the Third District Court of Appeal threw out those contractual provisions, on grounds that they are unconstitutional. Alas, it’s almost certain that PECG will appeal to the California Supreme Court, so this long-running battle is probably still not over.

British Rail Privatization Success. I’m sure you’ve read a number of articles on the “disastrous” privatization of British Rail a decade ago. It was certainly not a big success, being overly complex (dividing up the former state railway into far too many different business units) and failing to address a major underlying failure of state ownership (gross under-investment in the track and signaling systems). But by bringing in private-sector train operating companies, the program has led to a huge revival of rail travel in Britain. Over the past decade, passenger miles are up 41% and freight ton-miles by 58%. There are now 20% more daily trains than a decade ago, carrying 2.93 million passengers a day, versus 2.08 million in 1995. Surveys show that passenger satisfaction levels are about 10% higher than a decade ago, as well. More information is available from the Association of Train Operating Companies at

Job Opening at Reason Foundation. Reason is seeking a policy analyst in transportation. Qualified candidates should have a graduate degree in a transportation-related field, a solid understanding of free-market public policy, and an aptitude for written communication. Ideal candidates will have experience with transportation modeling. Work location is negotiable and salary will be commensurate with experience. Please send your resume to:

» return to top