The Dangerous Street Was Designed “Correctly”

Somewhere there lies a set of construction drawings for a street where a crash turned fatal. If we pulled those drawings and analyzed them against municipal standards, we would likely find nothing wrong. Lane widths, curb radii, crosswalk placement and design speed all meet standards. Although the plans are right by every professional measure, the resulting street fails the people who use it. At first glance we might blame the driver, then perhaps the designer. But take it a step further: why was the design allowed to pass, to begin with?

In Confessions of a Recovering Engineer, Chuck Marohn confessed to once building streets exactly like that one. Not out of carelessness, but out of compliance. Like most engineers early in their careers, he worked within a system where the manuals defined what “safe design” looked like, even when real-world outcomes suggested otherwise. This is not unique to one engineer or one agency. It is how the system functions. Design manuals exist to make practice uniform and to reduce engineer liability. “How do we keep cars moving” is the basis standards were built on, and which placed less emphasis on “how do we keep a person on foot alive.” So a street can satisfy every rule in the book and still be a place where a bad crash is not a fluke but a matter of time. 

So what are “the standards,” exactly? Most people have never seen them. We cannot fix a system we cannot name. So let’s name them.

Two books shape nearly every street in the United States. The AASHTO (American Association of State Highway and Transportation Officials) Green Book sets the physical layout: shoulder and lane widths, curbs and sidewalks, and sight distances and medians. The MUTCD (Manual of Uniform Traffic Control Devices) governs the signs and signals. Inside both is one decision that shapes everything else: the design speed. The number itself rarely looks alarming, but even a modest design speed widens the lanes and flattens the curves, and a wider, straighter street invites a heavier foot, no matter what the posted sign says. The driver is responding to what the street invites. This is the ranking Chuck has described: speed first, then traffic volume, then safety, and cost often last. Ask the people who live on the street to rank those same priorities and you usually get something closer to the reverse. 

Once you start looking, the pattern is everywhere. Start with something almost everyone has felt and never questioned: the crossing signal that starts counting down before you're halfway across. That countdown isn't arbitrary. It comes from a formula that assumes a walking speed of about three and a half feet per second. That is brisk enough that plenty of real people can't hit it. An older adult, someone pushing a stroller, anyone with a knee that doesn't bend the way it used to, all of them are, by design, given less time than they need. Nobody set out to leave people behind. The number in the manual just assumed a walker who doesn't fully exist.

Lane width tells a subtler version of the same story. Ask an engineer why a lane is 12 feet instead of 10, and you may hear about "operational comfort," giving drivers a little more room for driver error. What that room actually does is tell the driver that it's safe to relax and go a little faster, yet the speed limit sign here wasn’t changed. Ask about the wide, generous curve at the corner and you'll hear that a fire truck needs to make the turn without climbing the curb, which is true, but the corner doesn't know it was built for the fire truck. Every car takes it a little faster, too. The clear zone requirement works the same way from the other direction. The manual calls for open, obstruction-free space alongside the road so a driver who runs off the pavement can have space to recover. 

None of these choices look reckless by themselves. A countdown, a lane, a corner — each one has a memo behind it. But add them up on the same street and the message comes through as: go ahead, you have room.

In many states, these manuals are adopted into law. The MUTCD (the Manual on Uniform Traffic Control Devices) is required under federal regulation and enforced through every state's traffic code, and the Green Book is written into state design manuals as the default. New editions come out roughly once a decade, so a formula built on one era’s assumptions can stay in force as law for a generation before it gets revisited. That means an engineer who wants to build something narrower or slower often can't. Even when flexibility exists, there is unspoken pressure at work. Build to the standard, and a crash becomes something that happened on a road that met every requirement. Step outside the standard to make something safer, and if anything goes wrong, that decision is the one that gets questioned. So the standard becomes the safe choice, for almost everyone except the person out on the street.

And who writes the manuals? Not Congress. AASHTO is an organization made up of state transportation agencies. Its technical committees include state DOT engineers, researchers, federal representatives and other transportation professionals who draft and revise the Green Book over long periods of review. Once the review process is completed, AASHTO publishes the new edition and it becomes guidance rather than a law on its own. States then choose to adopt it, usually by referencing it as the default standard or by writing it directly into their own design manuals. 

The MUTCD path is different as it is a federal regulation. The FHWA (Federal Highway Administration) proposes updates and publishes them for public comment before adopting a final rule. Though the FHWA ultimately decides what makes it into the final version, local government or an individual citizen can weigh in. Once adopted, the MUTCD is incorporated by reference into federal code and that makes it binding. By the time it reaches the engineer at their desk, the decisions inside it were created long ago and far away from the street where they will be applied. Most people will never read these documents, but they shape every street outside our front door. 

None of this is one street’s bad luck. At Strong Towns, we have spent years looking closely at real crashes through our Crash Analysis Studio, and the same handful of features keep turning up. When we gathered the findings from two dozen of those studies into the Beyond Blame report, we weren't describing two dozen unlucky streets. We were describing one design, repeated two dozen times. 

What matters is that none of this is fixed in stone. The manuals get rewritten, and cities are not stuck waiting for that to happen through the usual channels. One of the alternatives already on the table is the NACTO (National Association of City Transportation Officials) Urban Street Design Guide. It was put together by professionals working directly with streets that serve people walking and biking, and it treats a street as somewhere people live, not just a route for cars to get through. For a long time, a city that wanted to use NACTO ran into a wall: the state Department of Transportations controlled the rules, even on a city's own streets. Today, cities have more ability than before to choose their own path, as federal law now lets a city adopt the guide on its own streets, and the process is genuinely low-lift. A standalone memo from the city's chief transportation engineer can do it. So can a council resolution, or a line added to the city's existing design manual naming NACTO as its official design standard. Dozens of cities and nine state DOTs have already gone through some version of this. 

So when we find an unsafe street, the instinct is to ask who failed. Was it the driver or the engineer? But trace it back far enough and you arrive somewhere different. The driver behaved the way the road invited them to. The engineer designed the way the standards required them to. Each one did the job in front of them. The failure is further up, in a system that was built to produce this and keeps being renewed. Which means the work isn't finding the person who got it wrong. It's changing the rules that make "right" so dangerous in the first place. That sounds huge. It starts small.

Pick a street that feels uncomfortable to you. Walk it or slow down driving it, and watch the street itself. How far is it from one corner to the other on foot? Time the crossing signal against your own steps and see how much room it actually gives you. Notice the width and number of the lanes and the way the corners are shaped. You start to see the design language behind the behavior. 

If you want to go further, run a Crash Analysis Studio. It's a process any of us can lead, not just professionals, to look at a crash and ask what in the design made it likely. That's how the conversation shifts: one street, one crash, one question at a time.