Reviewed by Edward L. Glaeser
The New Republic Online
Few if any man-made contraptions have shaped America as much as the internal combustion engine. Eighty-seven percent of American workers drive to their jobs, either alone or in a carpool. Eighteen percent of the average American family's spending goes to transportation, mostly for cars and gasoline. While Homo sapiens has spent most of its existence fighting to fill its belly, Americans today spend a lot more on cars than on food.
Just a century ago, before the rise of cars and trucks, American cities huddled around ports and rail yards. New York's needle trades grew up in the shadow of the clipper ships that brought cotton; Chicago's stockyards slaughtered cows by the millions a few feet from Armour's refrigerated rail cars. People mostly walked to work, and lived in small, often self-built homes close to the firms that employed them. Then, in the twentieth century, we rebuilt our urban landscapes around a new transportation technology.
Cars turned the early trickle of streetcar suburbs into an exurban torrent. Shopping malls with plentiful parking replaced downtowns as places to play and to purchase. Trucks enabled companies to give up the older transport infrastructure in the urban core; today more than three-quarters of the jobs in most metropolitan areas are more than five miles away from the city center. America built new cities with multi-lane highways and broad boulevards that suit the car the way the tangled streets of Jerusalem's Old City suit foot traffic. Cheap transportation enabled tens of millions of Americans to move to once-inaccessible places, such as California and Texas, with splendid climates, or at least with affordable snow-less living.
The car's weak heel, of course, is its dependence on petroleum, which is neither infinitely abundant nor environmentally innocent. Approximately one- fifth of America's carbon dioxide emissions is associated with private driving. In the years ahead, environmentalists could quite possibly get their wish that Americans pay for the ecological damage caused by the car through gas taxes as high as those that have long reduced European driving. Even without a carbon tax, gas prices have soared as demand from vast and growing regions such as China and India has collided with static oil supply. If our society decides to act on the threat of global warming with a carbon tax, or some other carbon control system, then gas prices will rise even further.
What would happen to our car-based world if the price of gasoline hits $10 per gallon? That number may be extreme, but it is not all that fanciful. A few weeks ago I paid $9 per gallon to tote my toddlers around Belgium. (Thank goodness it's a small country.) Some environmentally inclined urbanists have rushed to the view that rising gas prices will take us back to nineteenth- century urban densities, hopefully with a little less cholera. Since the belief that people respond to incentives is at the heart of the economist's catechism, I certainly share their view that rising gas prices will do something to behavior. But I am less confident that even extreme prices will mean an end to cars and car-based living.
There is much wisdom in Tom Vanderbilt's assertion in Traffic, a smart and comprehensive analysis of the everyday act of driving, that "it will be easier to remove the internal combustion engine from the car than it will to remove the driver." While rising gas prices will surely reduce miles traveled and increase the appeal of high-density dwellings, the main effect of expensive petroleum will no doubt be to increase automotive fuel efficiency.
A massive rise in gas prices would seem to create an enormous incentive to give up driving altogether. The average American family earns about $60,000 per year and buys about 1,000 gallons of gas to drive about 20,000 miles. At $2 per gallon, this was almost 4 percent of such a family's average expenditures. At $4 per gallon, they are out an extra $2,000 per year, which is plenty but perhaps not enough to uproot and move to midtown. At $10 per gallon, by contrast, they would be spending more than 20 percent of their after-tax income on gasoline -- a truly mind-boggling figure. This family, which has plenty of other expenses, will have to do something.
The easiest and most likely response to high gas prices is to buy a high mileage car. If that family switches to a fifty-mile-per-gallon Prius, then a fivefold increase in gas prices will mean only a $2,000 increase in gas spending. Moreover, as Vanderbilt's comment suggests, the automobile industry will continue to innovate as gas prices rise. The potential for more fuel-efficient cars is one of the most important reasons why public policy should not try to reduce energy costs: high prices ensure that the engineers at Honda stay busy building a better hybrid.
Middle-class Americans will drive a little less and the prices of houses far out on the exurban frontier will fall a bit relative to central-city residences. Public-transit usage is already up in many metropolitan areas. But that does not mean that Americans will desert their suburbs. The housing stock is durable and valuable, and all those shiny new homes on the edges of Atlanta and Phoenix are not going to be abandoned anytime soon. As long as those homes are occupied, the people living in them will drive. They have to drive. Attempts to service these low-density areas with rail are enormously expensive undertakings that can never be justified on the basis of ridership.
Americans, after all, have good reasons to like their cars. The average public transit commute in the United States is forty-eight minutes, about double the average car commute. We might be able to reduce emissions and highway deaths by moving back toward nineteenth-century transport modes, but we will do so only with a huge increase in travel times. So Americans are likely to stay behind the wheels of their cars -- and they were joined by fifteen million more Chinese cars in 2007 alone. And all those drivers mean, among other things, that Vanderbilt's book is likely to remain relevant well into the new century. The book is a shrewd tour of the much-experienced but little-understood world of driving. For almost three hundred well-researched pages, Vanderbilt takes us through the automotive insights of social scientists and transportation engineers.
Driving is such a common, mundane activity that most of us forget how strange it is to get into a two-ton hulk of steel and high technology and hurtle down ribbons of highway. After all, most of our day-to-day experiences involve decisions made over long periods of time with stakes that are often no higher than where we want to eat lunch. Driving, by contrast, involves split-second decision-making in which our lives are almost constantly on the line. And the semi-subconscious nature of driving leads Vanderbilt to the counterintuitive conclusion that dangerous roads may in fact be safer. When roads are straight and simple, then drivers get complacent, which actually makes things more dangerous. But a slight bend in the road or the right road hazard can make drivers pay more attention.
Vanderbilt's views are a more moderate take on the Peltzman effect. Thirty years ago, Sam Peltzman, a former teacher of mine at the University of Chicago, achieved academic immortality for presenting evidence suggesting that seatbelts cost lives. Peltzman argued that seatbelts cause drivers to take more risks, and that those risks more than offset the direct effect of seatbelts on lives lost. While later evidence has challenged Peltzman's take on seatbelts, his basic intuition about safety is surely correct. Whenever some aspect of driving gets safer, drivers somewhat offset this safety by taking more risks.
Another unusual aspect of driving is its anonymity. In most of our social interactions, our identities are clear and obvious, and this keeps our darker side in check. In non-anonymous settings, people behave with more civility and generosity. I wouldn't want to have a rant in a restaurant that could show up on YouTube. But our cars mask us as well as anything the Lone Ranger ever wore, and our highways involve millions of one-shot anonymous interactions. No wonder highways sometimes feel like a Hobbesian free-for-all.
The fact that civil society ends at the on-ramp creates challenges for highway design. Vanderbilt illustrates these problems with an extensive discussion of lane merges. In polite society, it would be relatively easy to punish people who wait until the last minute to get out of a lane that is coming to an end, but on the highway last-minute movers get away scot-free. Vanderbilt teaches us about the "late merge" system, which respects the human tendency to behave badly, and asks everyone to stay in their lane until it closes. At that point the two lanes merge and drivers are instructed to take their turn merging. Apparently this works better than traditional lane-merge systems in many settings.
There are other significant anomalies about driving. Our relationship with the government changes drastically when we take to the roads. In the rest of our lives, we live in a world where property is private and the police are a benign force that protects us from rowdy teenagers. But a trip down our highways, by contrast, can seem like a visit to the old Soviet Union. Everything is owned by the government, and speed limits are set so that pretty much everyone is always guilty, just like back in the USSR. It is certainly the only circumstance in my life in which the sight of a police car fills me with dread.
Vanderbilt's book reminds us, though, that the public servants who oversee the roads are hardly as hapless or malignant as the old Soviet apparatchiks. Transport engineers come off as smart, innovative people who are constantly trying to make the roads a little bit safer. Vanderbilt provides a wonderful series of vignettes about the Dutch transportation guru Hans Monderman, a father of "psychological traffic calming," which is the art of making drivers in crowded areas slow down. Monderman eschewed speed bumps and street signs in favor of the stuff of neighborhood life--sculptures, children's bicycles, and unmarked roundabouts that forced drivers to be on guard. The result was a drastic reduction in accidents.
But just because there are plenty of competent highway experts out there does not mean that our transport policy is in good shape. It is an odd feature of American politics that we are going to spend a lot of time over the next six months talking about health care and almost no time talking about transportation. Most younger people spend much more money and time on transportation than they do on their health. The government has been deeply involved in the transit sector since George Washington was president of the Potomack Canal Company. Driving involves social costs -- congestion, road accidents, pollution, road depreciation -- that are not automatically internalized by drivers. Unless we move to a world of completely private roads, which is awfully hard to imagine, this means that there will be plenty of scope for good government policy.
Vanderbilt gives short shrift to the pollution problems of the car, which seems fairly reasonable, since oceans of ink have been spilled on how to clean up our air and fix the world's fever. The most natural means of reducing carbon emissions from driving is to have an appropriate carbon tax on petroleum (and on every other form of carbon-emitting energy, for that matter). If we want to make sure that this tax does not turn into just another revenue source for the state, the taxes can be returned to taxpayers on a per capita basis.
The thornier problem is how to reduce the environmental costs of pollution in places where the American government's writ doesn't run, most notably India and China. One approach to their carbon emissions is to offer incentives to these places to impose their own carbon taxes. A second approach is to subsidize research in fuel efficiency and just give away the results of that research. A multibillion-dollar program that gives cash rewards to big energy innovation offers a way of using the human tendency to respond to incentives to help reduce energy use.
Vanderbilt is more interested in the problems of traffic congestion and accidents. Unfortunately, reducing one of these problems can exacerbate the other. Twenty years ago, the Harvard economist Julio Rotemberg wrote an elegant paper showing that drivers increase congestion by leaving too much space between themselves and the car ahead of them. But while less space between cars speeds up commutes, it also increases the rate of accidents. Conversely, Vanderbilt tells us about a lot of neat tricks that reduce accidents in urban areas by slowing cars down. This saves lives, but it increases congestion.
One way of reducing traffic is to lay down more asphalt, but Vanderbilt concludes that this is a false panacea. One of the iron laws of transportation is that "more roads lead to more traffic." We currently ration our roads with traffic delays, so as we add more roads, people will drive more. Vanderbilt also reminds us that "adding more land is a process of diminishing returns," because highways become less efficient as they grow. We have built more and more roads, and still travel times increase. The Texas Transport Institute estimates that Americans are spending 4.2 billion hours a year in traffic delays. An average American spends 438 hours a year, more than ten normal work weeks, in travel time.
Vanderbilt touts freeway entrance-ramp metering as a means of congestion control. It is almost paradoxical that we can speed up highways by slowing down the rate of entry onto them, but it is true. A famous Minnesota experiment turning on and off the freeway ramp meters showed that with the meters, freeways were able to handle more cars at higher speeds with substantially fewer accidents.
As an economist, I am inevitably drawn to using the price mechanism to regulate the flow of traffic. In the half-century since William Vickrey first pushed congestion pricing, the idea has moved into the mainstream. Americans seem to think that the right to drive at no cost was slipped into the Bill of Rights, perhaps as part of the penumbra of the Second Amendment, but this makes no sense. If you do not ration something by price, then it gets rationed by queue. All that time spent slowly trudging in traffic is pure social waste. It would be far better to follow London and Singapore and embrace the principle that people should pay for the costs that their driving imposes on other drivers, especially in big cities.
The enemies of congestion pricing like to argue that it is regressive, but a die-hard socialist such as "Red" Ken Livingstone would hardly put forward a policy aimed at hurting London's poorest. In that city, as in most of the world, it is the rich who drive and the poor who take buses. By charging drivers, the poorer bus-riders gained twice. They got faster commutes, and Livingstone used the congestion charge revenues to improve public transit. If there is a single American metropolis where people should face an added charge for driving, it is New York, but Mayor Bloomberg's valiant fight for congestion pricing died in the committee rooms of the State Assembly in Albany. There is always going to be plenty of political opposition to charging for something that now seems to be free. The irony, of course, is that congestion pricing promises to reduce the now-exorbitant time costs of driving in downtown Manhattan. The London experience was that congestion pricing was politically unpopular right up to the day it was imposed, and then drivers suddenly realized that they liked facing less traffic.
Spreading traffic more evenly over commute times offers some of the biggest gains in driving efficiency. If we charged higher tolls for driving during peak hours, when freeways are most crowded, and lower tolls for driving off-peak, then some people would be induced to commute during times when there are fewer drivers. It is better to use our existing roads more efficiently than to just mindlessly build more highways.
Vanderbilt is particularly exercised by the almost forty thousand people who die in traffic fatalities each year in the United States alone. Few wars over the past thirty years have been able to match the car in lethality. Pro-safety interventions that lull us into a false sense of security are the mindless approach to traffic safety -- analogous to fighting congestion by building more highways. Vanderbilt makes a strong case that increasing attention on the road is a smarter path toward safer streets. In the safety arena, using stronger incentives means harsher penalties for drunk or risky driving. We have made great strides in fighting drunk driving since the 1950s, when the sight of a drunk behind the wheel of a Buick seemed like a great joke. The continuing flow of deaths on the road, however, reminds us that there is much left to be done.
America is a car-based society. There is no point in pretending otherwise, even with the price of gas soaring. All those cars are unlikely to disappear any time soon. If anything, the vast increases in car ownership in India and China are just making it more important that we continue to address the large problems that accompany the car-given gifts of speed and autonomy.
I grew up on the subway, and like many urbanists I hope that more people will take public transportation. Yet it is pure fantasy to think that the majority of Americans are likely to give up their cars anytime soon. Higher gas prices will lead to more fuel-efficient cars long before we see a wholesale abandonment of car-based suburban living. The car is here to stay, and transportation policy should try to make driving faster, safer, and less environmentally harmful. Tom Vanderbilt's book is a balanced and instructive discussion on how to improve our policies toward the inexorable car.
Edward L. Glaeser is Fred and Eleanor Glimp Professor of Economics at Harvard and director of the Taubman Center for State and Local Government.
Books mentioned in this post