How to Fix Traffic Forever

 In 2004 the American Highway Users Alliance published an article claiming that the Katy freeway in Houston, Texas was the second most congested road in America. Drivers spent a combined 25.4 million hours every year sitting in traffic on that road. That’s 36 lifetimes worth. It was an absolute embarrassment of a road for Texas so they decided to spend $2.8 billion to expand it to 28 lanes making it the widest highway in the world. All this extra capacity would surely fix the problem—more lanes means more room for more cars which means shorter travel times... right? Unfortunately not. Between 2011 and 2014 alone, travel times on the Katy freeway increased by up to 55%. It now takes an average of 64 minutes to drive the 28 miles between downtown Houston and Katyland during the afternoon rush hour. That’s up from 41 minutes in 2011. The problem with this project was that the solution violated the fundamental law of road congestion—more lanes mean more traffic. This probably seems counterintuitive. The more lanes there are the higher the capacity a road has so cars should be able to drive faster... but that’s a very narrow analysis of the problem. You can’t just think about how this would work on one road, you have to think about it in the context of a whole city. Many people choose not to drive places because of how long it takes. If the traffic is bad, they can take public transit or wait until a less busy time, or just not travel at all. When a road is expanded, travel times initially decrease so all those people who chose not to drive or to take an alternate route or whatever decide to switch to using that newly expanded road. What’s fascinating about roads is that this happens at a perfect 1 to 1 ratio. If the capacity of a road doubles, the amount of people using that road will also double. If it doubles again, the amount will again double. Of course if you kept doing this over and over again you would eventually build a big enough road that there wouldn’t be any more cars to fill the road, but in the real world where demand for roads far outstrips supply, drivers will adjust to any change in road capacity. So does that mean that it’s hopeless? Is there no way to solve traffic? No way to make our roads better and safer and more efficient? Well... no. There’s plenty we can do. Here’s the thing about traffic—it slows down exponentially. The 20,000 car on the road slows down traffic overall significantly more than the 5,000 car. This is a major driver for a lot of jams—a small addition of cars leads to a large addition in congestion—but it also makes solving traffic a bit easier since you only need to remove a small amount of cars from the road and that’s just what ramp meters do. Ramp meters are set up on the on-ramps of highways to restrict the amount of people getting on the highway. They usually let one car on every five or six seconds. Since the amount of cars actually on the highway is kept down, the highway stays at its most efficient speed. Minnesota did an experiment where they shut down their long-used ramp meters for eight weeks in order to see if they actually worked and they found that the highway capacity decreased by 9%, travel times increased by 22%, speeds dropped by 7%, and crashes increased by 26%. Stockholm, Sweden used that exponential nature of traffic to decrease travel times by up to 40% in 2006. Stockholm as a city lies across 14 islands which means that all the bridges act as huge chokepoints. Traffic, therefore, was historically horrible for the relatively small city. On January 3rd, 2006, Stockholm started to charge drivers who entered this central perimeter—the busiest area. The charge wasn’t much—between 10 and 20 krona, the equivalent of 1 and 2 US dollars— but it was enough to persuade 20% of drivers to not enter the central perimeter. They either went downtown on public transport or walked or didn’t go at all. These are the amounts of daily drivers in the perimeter in the years leading up to the charge. As soon as the charge was implemented in 2006, the daily amount dropped down to here. It wasn’t a fluke. After the 6 month initial trial period driving in the central core became free again and the amount of daily drivers increased to nearly the level it was before. When the charge became permanent in 2007, daily numbers once again plummeted. Even though the charge was minuscule, it was enough to dissuade 10s of thousands of people from using those roads. There are really two costs of driving—the money and the time. When the time it takes to drive isn’t enough of a cost to prevent people from driving, these charges increase the overall cost to a level where some people will decide not to drive. But what about safety? Roads are still unbelievably dangerous. In any given year, 1 out of every 10,000 people in the US die in a car accident. Just think about how high of a proportion that is. If you go to a Redskins game at FedEx field near Washington, DC, eight of the people sitting in the stands with you will die in the next year in a car accident. It turns out one of the best ways to prevent accidents is with something you’ve almost certainly already seen or used—the roundabout. There’s a reason you see these more and more. Roundabouts reduce deaths and serious injuries by 90%. That is not an error. With roundabouts, there’s almost no opportunity for the worst type of collision—the head on full speed crash. In a traditional intersection, cars come within feet of each other while going at a relative speed of up to 100 mph. A head-on crash at that speed is undoubtedly catastrophic. With roundabouts, cars naturally slow down to about 15-25 miles per hour since they’re going around a curve. Also, if there were to be a collision, it would either be a side-impact collision if a car failed to turn into the circle or a side-to-side collision if a car misjudged the curve. Both of these collisions happen at a low relative speed so fatalities are low. But what about capacity? Surely the lower-speed roundabouts cause horrible traffic problems. Well... they don’t. A single lane roundabout can handle a maximum of 1800 vehicles per hour which is exactly the same as a traditional two-lane signaled intersection. While cars will move through a signaled intersection at a much higher speed, they have to wait both for the light to change and left-turning cars. With roundabouts, you have a smooth, consistent, albeit slower, flow. So what’s the problem? Why haven’t we replaced every intersection with a roundabout? Well there are disadvantages—they’re more difficult for pedestrians, especially those who are deaf or blind, they require a larger footprint, they’re more expensive to maintain—but the real reason roundabouts are not ubiquitous nowadays is because of the biggest fallacy in road design—that drivers need rules. Poynton, just outside of Manchester, UK, used to have a typical, rather dreary intersection and nobody really liked it. Cars would back up for miles, pedestrians had to wait forever for the light to change, and it essentially split the town apart. So someone had the idea to remove the traffic lights, remove the zebra crossings, the curbs, remove almost every safety device in the intersection and just set up two adjoining roundabouts. Surely this would wreak havoc, but it didn’t. Turns out, when people are uncomfortable, when people aren’t really sure what’s going on, they pay more attention. The green light was a signal to people that the road was clear, that it was safe to speed, that they could let their guard down, but after the change the cars were able to flow freely, albeit at a slow pace, instead of waiting for the lights to change. Pedestrian incidents went down, collisions went down, traffic flowed faster, and the city center finally had some character. So, all around the world cities are replicating what Poynton did. They’re removing curbs, traffic lights, and pedestrian crossings to make one shared space. All around the world, these streets are resulting in fewer accidents and more pedestrian space. Discomfort is saving lives. On a larger scale, there’s one more innovative intersection design that’s beginning to save lives—the diverging diamond interchange. This interchange is designed as a way to get more cars on and off highways faster. After the on-ramp to the right side, the road crosses over so cars never have to traverse active lanes to get onto the highway. A car heading north can effortlessly join the on-ramp without crossing traffic, and a car heading south will cross over so it drives on the left side and can effortlessly join the on-ramp to head south. Not only is this easier for drivers, it improves safety. The dangerousness of an intersection is often rated by determining the number of conflict points—possible points where accidents could happen under normal circumstances. With a traditional on-ramp intersection there are 26. With a diverging diamond intersection, only 14. And they’re faster too. The US Department of Transportation found in a study that universally, whether the traffic was light or heavy, diverging diamond interchanges let more cars through faster. It costs less too. A traditional on-ramp intersection requires $11.3 million to build; a diverging diamond intersection, only $5.7 million. There are really no major disadvantages to this intersection so nearly 100 of them have been built to date and more and more are being installed each month. As good as these solutions sound, there’s no one way to solve traffic. The difference between cities with chronic traffic problems and those without is a combination of smart policies and designs that mitigate the effects of having more road demand than supply. But traffic won’t just fix itself so until cities at least experiment with solutions we’re all condemned to traffic, forever.