The dark paradox of traffic jams

The morning of December 15, 1973, was not an especially eventful date. But in New York City, a road collapsed, proving that the amount of traffic congestion in a city is only tangentially linked to the amount of road space.



The road was the West Side elevated highway, which was built between 1929 and 1951. When it finally opened, stretching from the southern tip of Manhattan all of the way up to West 72nd Street, it became a symbol of New York’s postwar vitality.



[Photo: Allan Tannenbaum/Getty Images]



Unfortunately, it wasn’t especially well built, and it was worse maintained. That day in December 1973, as a heavy dump truck overloaded with nine tons of asphalt went over it, several of the supports failed, and the truck plunged onto the street below. By sheer chance, nobody was killed—a rare miracle in the history of road collapses.



But it did cause an almighty traffic jam. The West Side Highway was not a lightly used route. Each day, it carried around 70,000 vehicles up and down Manhattan. Hundreds of vehicles had to reverse their way off the road, over a period of several hours. The city of New York closed the road, and for the next couple of decades it sat, derelict. Before it was finally replaced, in the 1990s, it was practically a perfect symbol of the Big Apple’s decline.



What it also could become, however, is a symbol of how more roads don’t solve congestion. When the road collapsed, one of the immediate fears of New York’s traffic engineers was that the vehicles it carried each day would reappear elsewhere in Manhattan, clogging up streets with more traffic. But in fact, after that initial jam, there were no more.



The reality was that “the predicted traffic disaster never appeared,” wrote former New York City Traffic Commissioner Sam Schwartz, who coined the term gridlock. The cars “went somewhere, but to this day we have no idea where.” Over the next few years, even though the road had disappeared, the number of people traveling into Manhattan’s southern business district actually increased. It just happened that fewer of them were doing so in cars.



This reveals something about traffic, and the problem of what is usually called “induced demand.” At the margin, whenever they set out on a journey, people choose between different forms of travel. In a city like New York, where walking, bicycling, or taking the subway or bus are all plausible options for getting somewhere, most people have alternatives to getting in their cars. If you think there will be a traffic jam, you might choose to take the subway instead. If you know that the road will be clear, you might decide to get in your car.



The result is that the level of congestion on a particular road is essentially priced in. How many lanes a road has doesn’t actually have that much of an impact on how jammed up it gets. Instead, what determines the traffic level is what other options drivers have. Traffic jams stop growing at the point at which the marginal driver decides it would be better to walk, or hop on a train, or simply go somewhere else.



So what happens if you expand a road by adding more lanes? Initially, people carry on with their old travel habits, and so traffic on the road speeds up a lot. But over time—perhaps just a few months—what happens is that people notice that they can get somewhere a lot faster on the new, wider road. Pretty quickly, congestion is just as bad as it was before.



Induced demand isn’t inherently a bad thing, obviously. If you build a road to, say, a rural village in Tanzania, you’d hope that more people would use it, and more residents of that village would be able to get jobs farther away from their homes, or more farmers would be able to sell their produce at higher prices in the city. But looking at congestion is a terrible way to measure the need for more roads, largely because using roads is generally free.



If you give away something that costs money to provide, you will almost always get more demand than you can ever possibly meet. And if driving is too cheap relative to other ways of getting around, there will always be more demand to drive. That’s why congestion is a problem in almost every big city.



And when we look at cities in rich countries, the reality is that overall, we have almost certainly oversupplied the amount of road needed already. To take another American example, look at Louisville, the biggest city in Kentucky. There’s a bridge there, called the John F. Kennedy Memorial Bridge, on the I-65 Interstate highway, connecting Louisville to Jeffersonville, a smaller city in Indiana. In 2010, around 120,000 vehicles crossed the bridge daily, which had six lanes of traffic (three in each direction), and routinely got clogged up.



So the highway engineers at the Kentucky and Indiana departments of transport decided to widen it. As reported by the City Observatory, together the departments of transport put up $1 billion to build a new, second bridge (this one named for Abraham Lincoln), adding six more lanes. In 2017, when the bridge opened, to pay back the cost they introduced a toll to cross both bridges—$2 for each crossing, with discounts for regular commuters.



And guess what happened? The number of drivers using the bridges fell, dramatically, to around 60,000 cars per day. It has never risen back to its previous level. What this meant was that the extra lanes proved completely pointless. As the City Observatory reported, “the two states spent $1 billion doubling the size of I-65, only to have half as many people use the bridge. That money was wasted.”



What happened to the extra cars? Much as with the collapse of the West Side Elevated Highway, it’s almost impossible to know. But the most plausible explanation is that most of them switched to using the Second Street Bridge, which was just a few hundred yards away but had no toll. That adds several minutes to the typical commute, and more at rush hour, but it saves drivers a few bucks.



The states thought they were helping drivers by speeding up their commutes, but the drivers actually didn’t care. When charged even a modest fee, traffic dropped away. There was already plenty of road to begin with.



Traffic engineers can’t resist trying to fix congestion with one more road. They tend to want to simplify things down to single measures that they can try to change. In America, the key measure used is typically the speed of a journey. The logic is that if people can get from A to B faster, then they have more time left over to work, or spend, or whatever. They can also go farther distances, meaning that more people can realistically reach any given job. When a lot of people already drive, this produces a logic that always favors widening roads, or building new ones, because it speeds up the time spent in traffic. And time spent in traffic is straightforwardly time wasted.



But the reaction to road tolling shows that these estimates don’t really stand up. The actual “cost” of congestion is what people would pay to avoid it. If so many people refuse to pay even a relatively small toll to avoid being in a traffic jam, that suggests that their productivity isn’t really being damaged anywhere near as much as the estimates suggest. Perhaps their journeys weren’t actually that important to begin with, or else there were perfectly good alternatives they were choosing not to use.



This is an example of what economists called “revealed preference.” It applies a lot to transport. Poll people and they say they want wider roads to reduce congestion. And yet if you actually build those roads, you discover that they won’t pay for them—their preference is revealed, and either people don’t actually mind congestion, or they’d prefer not to drive.



Trying to solve traffic congestion by endlessly building more roads isn’t a good use of resources. So when faced with congestion, what should we do about it? In general, the best way to reduce congestion is to charge a reasonable price for using the roads. If the price of using a road is put at a level where the people who need them most, and have no other realistic way to get around, can afford to reasonably use them, and there is still crippling congestion, then roads are probably in short supply.



But if, as with the bridge in Kentucky, the result of charging is that congestion falls significantly, then you already had enough road space to begin with, or possibly too much. And even if it doesn’t, you can invest the money raised in improving public transport options so that they don’t need cars anywhere near as much.



[Cover Image: Abrams Books]



Any sane transportation policy would think in these terms—what transport do people value, and what would they be willing to pay for? We know that access to good public transport is something that people are willing to pay money for. Roads, by contrast, are simply not that valuable—at least relative to their cost—because we already have so many.



Excerpt adapted from the new book Carmageddon: How Cars Make Life Worse and What to Do about It by Daniel Knowles, published by Abrams Press ©2023