Brookings' Anthony Downs Concludes That Traffic Congestion Is Our Inevitable Future

By Anthony Downs,

Senior Fellow, The Brookings Institution*

March 21, 2001

Accommodating Future Population Growth

The first crucial consideration is the likely future growth of U.S. population. Even before receiving results from the 2000 Census, the Census Bureau estimated that, from 2000 to 2020, the total population of the U.S. will rise by 48.2 million persons, or by about 12 million every five years. That is a 17.6% increase in 20 years. And the first results of the 2000 Census indicate that future growth may be even greater than that. Somehow, U.S. ground transportation systems must expand their capacity to cope with this large increase in persons and households and goods.

This significant growth will occur mostly in the West and South, and in a small percentage of our metropolitan areas (MSAs). Average MSA growth will be slightly over one percent per year, but many MSAs will grow faster .

From 1980 to 1997, we added about 1.2 cars or trucks to our nation's vehicle population for every additional person in our human population. That implies that, from 2000 to 2020, there will be 48 to 62 million more vehicles in the U.S. That would be a 24-28% rise in vehicles over the 2000 total of 214 million.

Many existing residents facing greater congestion want to "limit future growth." But these sentiments are delusions. Existing residents in any region cannot stop either domestic or foreign immigration into it by adopting anti-growth policies. A region's growth rate is determined by such basic traits as its climate, its location in the nation, its topography, its natural resources, its demography, and past investments made in it by governments and businesses. These traits cannot be changed by local or even statewide policies. Our challenge is to accommodate growth, not prevent it.

Residents of a specific locality can limit future growth within its own boundaries, but that merely moves the region's growth to other localities therein-farther out or in overcrowded city slums. So local anti-sprawl policies make sprawl worse.

Over a longer period, future population growth will be even more enormous. The Census Bureau projects a mid-series estimate of 393.9 million residents by 2050-a gain of about 119 million over 2000, or 43.4%. Imagine the road traffic if we keep adding more than one vehicle for each added person in our population!

The Continuing Dominance of Private Automotive Vehicles

Privately-owned automotive vehicles will remain the dominant form of ground transportation for at least the near future, and probably longer. Attempts to cope with rising traffic congestion by luring more people to public transit will not work well enough to change this situation. During the next 20 years, the automobile will remain a preferred form of movement for most people in spite of worsening congestion. Under most circumstances, cars and small trucks are faster, safer, more comfortable, more flexible in timing and in linking scattered origins and destinations, and often cheaper. Hence private automotive vehicles will remain dominant even though they have significant social costs in deaths, injuries, and pollution.

Improving the quality, quantity, or frequency of public transit may be desirable in many MSAs, but doing so would not attract any notable proportion of present auto-driving persons into using transit. Only 3.5% of work trips in 1995 were on transit, compared to 90.7% in private vehicles. Even if the total percentage of persons commuting by public transit tripled, that would reduce the percentage using private vehicles by only 11.6%. Any reduction in congestion achieved through that reduction would be more than overcome by sheer population growth.

The only way to substantially increase the percentage of trips done on public transit would be to make use of automotive vehicles far less convenient or far more costly. But doing so by greatly raising gas taxes as in Europe, or escalating license fees, would be strongly opposed by most Americans. So these policies will not be adopted. Eventually, rising congestion may require such steps, but probably not by 2020.

Public transit proponents complain that automotive vehicles get large public subsidies. But transit now receives over 20% of all public transportation spending, but provides under 2% of all person trips per year, and only 10% in large urban areas. So transit is even more heavily subsidized per trip or per capita than private vehicles.

Why Traffic Congestion Will Inevitably Get Worse Everywhere

Most Americans consider traffic congestion as one of the most serious social problems they face in daily life, especially in fast-growth suburban areas. No doubt, it has been getting worse. But there is no way to prevent traffic congestion from intensifying even more in the future. This is a problem without a solution-at least no solution the American people will accept.

[T]here is no doubt that congestion is impeding movement during much of the day in regions like the San Francisco Bay Area, affecting efficiency and the cost of living there. Nevertheless, the American people will not accept any of the tactics that might actually reduce peak-hour congestion. Hugely escalating the cost of driving through higher gas taxes or other fees is a political non-starter. Improving public transit will not lure enough Americans out of their cars to cut congestion much. Once peak-hour congestion has appeared on major roads, widening them only temporarily speeds traffic there.

Similarly, it may often be desirable to build more roads to cope with population increases. But in the long run, improving roads often attracts even more population and other growth that eventually tends again to overload those roads.

Economists have long recommended charging high peak-hour tolls to ration scarce highway space during rush hours. But U.S. politicians have unanimously rejected that strategy for two reasons. First, most drivers would consider such tolls as "just another tax" on something they can do without monetary cost now. Second, most drivers would regard high enough peak-hour tolls to reduce congestion to low levels as unfairly benefiting wealthy drivers who could always pay such tolls at the expense of poorer ones who would be forced to drive at other times. Perhaps "HOT lanes"-high occupancy toll lanes on major expressways onto which single drivers could buy their way during peak periods-could provide some high-speed channels even during peak hours without forcing all other drivers off those roads during those hours. But "HOT lanes" would not eliminate peak-hour traffic jams for the majority of drivers.

In reality, traffic congestion is essentially a balancing mechanism that enables people to pursue certain specific objectives they value other than minimizing commuting or driving time. Employers want most firms to use similar work periods during each day so workers can communicate with each other for economic efficiency, but that requires most people to travel to and from their jobs at the same times. Employers also want to operate mainly in low-density workplaces, widely scattered across each metro area. Most households want to (1) have access to a wide range of choices of where to work and live, especially in multi-earner households, (2) combine multiple purposes into individual trips, (3) live in relatively low-density communities, and (4) separate their own dwellings spatially-and within public school districts-from families with lesser incomes and lower social status.

It is not possible for people to pursue all of these objectives effectively without generating a lot of traffic congestion, especially during peak travel times. Yet most of us will endure a lot of congestion before giving up any of these objectives. The congestion we encounter is bad enough to make us complain loudly, but not bad enough to make us change our behavior.

Thus, increasing traffic congestion is an inescapable part of living in modern metropolitan areas everywhere. Peak-hour congestion is actually worse in most other parts of the world than in America. It is a mark of rising prosperity around the globe. If congestion gets bad enough, more people will react by relocating their homes closer to their jobs or vice-versa, or by moving to smaller metropolitan areas. To believe that future congestion will be remedied by adopting more public transit or any other set of remedies is a myth.

Changing Land-Use Patterns in Order to Reduce Future Congestion

In the long run, it would be possible-at least in theory-to influence traffic congestion by changing future land-use patterns to improve accessibility and ease of movement. Two basic approaches to changing existing patterns are being promoted as means of decreasing infrastructure costs, reducing congestion, and increasing accessibility. Unfortunately, both approaches are partly confounded by the fact that about 85% of the developed portions of the nation that will exist in 2020 were already in place as of 2000. Even if radical changes in the form of the to-be-added 15% could be achieved-which is not likely-that would not substantially change the patterns already in place today, which will necessarily dominate the overall picture in 2020. Even in fast-growing regions, well over half of all settlements in 2020 are already here today.

Over a longer run, existing settlements will comprise a smaller percentage of the total built environment, so more change can be achieved. By 2050, 30% of the then-existing population will live in settlements built after 2000. But that means 70% of those future settlements are already here. Nevertheless, some discussion of these two approaches is desirable.

• Adopting more Pedestrian Oriented Developments (PODs). The "New Urbanists" and others claim that adopting pedestrian-friendly and transit-oriented residential development can help remedy regional traffic congestion by causing much higher percentages of future residents to walk or use public transit. Their pedestrian-friendly design concepts, and grid street patterns, are genuine social contributions. But I do not believe their predictions of much higher future transit use are justified.

[P]icture an MSA initially containing one million residents, 5% of whom commute by transit, with an annual population growth rate of 1.5% from all sources. Assume half of its new residents are housed in circular PODs. Further assume that 50% of those POD residents commute by public transit. Then at the end of 10 years, total commuting by public transit would have risen from 5.0% of all workers to 8.1%. If all the rest commuted by private vehicles, that would leave 91.9% .

• Using urban growth boundaries to limit outward expansion. The second basic approach to reducing future congestion by changing land-use patterns is adopting urban growth boundaries to encourage higher-density new development. This would presumably make use of public transit more feasible in new-growth areas because they would have higher average residential densities than present fringe areas. This approach would also reduce the costs of building infrastructures to serve large low-density areas, and would rely heavily on in-fill development to slow future sprawl. But it has three major problems. In this scenario too, 85% of all 2020 settlements already exist now .

Moreover, the only rational way to limit outward sprawl is through regional growth boundaries. But American political allegiance to the sovereignty of local governments over land use policies is rooted in the desire of the home-owning majority in each suburb to insure that property values keep rising, or at least do not fall. This pressures politicians at all levels both to reject regional land-use planning, and to permit exclusionary zoning that hurts renters and the poor. We now have regional planning only as a last resort in a few MSAs where crises have arisen. Such crises have occurred where development threatened environmentally-sensitive and highly-valued areas, such as the Everglades in Florida and the Willamette River Valley in Oregon, or where the Environmental Protection Administration cut off federal highway funds due to high air pollution, as in Atlanta, or where state courts threatened to curtail local zoning unless more affordable housing was built, as in New Jersey. Almost everywhere else low-density zoning and sprawl prevail. True, sentiments regarding some regional arrangements are likely to change over the next two decades and beyond when the need for area-wide planning becomes more evident to voters. Traffic gridlock may create pressures for more rational planning, but only in some of the largest MSAs where the worst traffic congestion exists.

Another difficulty with this approach is that limiting outward expansion of growth requires raising densities in existing built-up areas. Yet residents of almost every existing neighborhood resist any increases in density, even near mass transit stations. It seems that Americans oppose both sprawl and higher density.

Over the very long run, it is hard to see how we can cope with huge population increases without some type of regional land-use planning coordinated with regional ground and air transportation facilities. But such regional arrangements have little popular support today.

Two Types of Dysfunctional Institutional Structures Concerning Transportation

To a great extent, approaching future ground transportation rationally and efficiently is hampered by two types of archaic institutional structures. First, existing means of governance in most metro areas are not capable of managing regional growth so as to create consistently higher densities in new-growth areas. Some type of regional planning and authority over land-use and transportation actions of local governments could create such a major change in existing development patterns. But only when traffic congestion reaches much worse crisis proportions are politicians likely to give up their loyalty to the concept of local autonomy, which portends that future policies will not work. Local governments seek to benefit only their own residents by shoving off all costs possible onto others. No one has any strong political incentive to focus on the well being of the entire region, so it is not well served.

The second major institutional roadblock lies in the regulations that govern public transit. Existing authorities bolstered by transit unions want to maintain monopolies of very inefficient large-scale systems that cannot achieve flexible approaches to serving low-density residential areas. Yet such areas will comprise the vast majority of all the new areas we are likely to build in the next two decades. We need to deregulate or even privatize public transit and allow small-scale operators that will serve low-density and low-income areas on demand. The construction of the Los Angeles subway is an example of legislative arrogance willing to spend billions on an approach guaranteed not to meet the real needs of future growth.

Imaginative management of public transit funds would encourage bidding for new types of services by private entrepreneurs. But the political power of transit unions and established institutions makes that unlikely. There is no need for both the funding of transportation and its production to be carried out by the same organizations. Public funds could support privately-run transit systems, including highly decentralized systems that could possibly provide effective service in low-density residential areas. There is no doubt that the nation needs widespread public transit services to provide mobility to persons who cannot drive because of old age, youth, infirmity, disabilities, or poverty. But continuing to focus the provision of such services in large-scale, high-cost public monopolies is not likely to work in the future any better than it does now.

Will the U.S. Resolve These Ground Transportation Difficulties Soon?

How likely is America to resolve the many problems with its future ground transportation described above?

Regarding traffic congestion, I do not believe there is any such thing as a "solution" or a "remedy" that will stop congestion from getting worse. We can and probably should build more roads to accommodate new growth areas, and better repair the roads we already have. We should also develop more effective means of public transit. But the desires of the American public for low-density living served by private transport and the immense flexibility it provides will not be diverted into any huge shift into mass transit. Moreover, all the added public transit we build will not really reduce future traffic congestion much, as our experience to date so clearly demonstrates. Some of the worst traffic congestion in the nation occurs in those MSAs best served by large-scale public transit systems, such as those in New York, San Francisco, and Boston.

Regarding land-use planning to reduce movement needs and emphasize public transit, only regional governance arrangements of some type can make a dent in our present infatuation with further outward sprawl-and even that may not work. However, anti-sprawl sentiment is rising around the nation, and it may eventually lead to willingness to accept some type of regional planning or coordination of growth closely tied to ground transportation facilities.

I realize that my assessment of the future of ground transportation may sound very pessimistic. But I have not mentioned one very positive factor. It is the adaptability of our population if given enough freedom from government rules and regulations. As congestion and other undesirable conditions worsen, people and firms will react by moving their homes, their jobs, their offices and other workplaces, and even their areas of residence so as to minimize the worst impacts of those undesirable conditions. That may take a long time, because people moved into congested areas in the first place because those areas were more attractive than elsewhere. But such adjustments will gradually occur.

The key goals of public policies should be to remove the political and institutional barriers to this adjustment process that now block it at so many turns. These include local zoning barriers to new housing development, unwillingness to consider region-wide planning and decision powers, and excessive regulation of public transit. When we are willing to break down these obstacles, we will do a better job of coping with the problems I have described.

Conclusion

You may be surprised that I have not predicted more radical technical and other changes in ground transportation over the next 20 years. But when I look back 20 years to 1980, I do not see evidence of many radical changes from then until now. And high-tech ideas like high-speed automated highways being proposed by many futurists frankly strike me as ludicrously costly and ineffective. My only radical conclusion is that, in spite of all our problems, I optimistically believe the nation will continue to be able to move around well enough during the next 20 years to maintain a rising standard of living for most Americans.

*The views expressed in this testimony are solely those of the author, and not necessarily those of the Brookings Institution, its Trustees, or its other staff members.