To truly prioritize safety, federal guidance must be grounded in evidence

Road safety is described as a top priority by transportation agencies and decision makers. However, the actions and regulations they produce often undermine this stated intention. Road design standards must be backed by concrete research to justify designs that prioritize safety ahead of speed.

T4’s policy proposal for improved safety

In our platform for reauthorization, under our first core principle of “Design for safety over speed,” we call for government agencies to “Conduct research and provide evidence-backed guidance for roadway design.” This article explains why this is critical, and some of the steps required to make it happen.

Current road design standards are NOT based on evidence and science about road safety

Roads are not naturally occurring features–they are designed and built by people. And these people are generally planners and engineers who follow certain rules and guidelines that are created by experts—particularly in the Federal Highway Administration (FHWA) of the U.S. Department of Transportation (USDOT). Sounds reasonable, right? The problem is that while everyone seems to agree that safety is a top priority, there are numerous instances in these rules and guidance where safety clearly takes a backseat to other objectives:

The Green Book and Highway Safety Manual: The main guide used by road designers is the Policy on Geometric Design of Highways and Streets, which is published by the American Association of State Highway and Transportation Officials (AASHTO) and informally known as the “Green Book.” AASHTO also publishes the Highway Safety Manual, which focuses more specifically on safety interventions and design. Both of these are adopted by FHWA, and they are often treated as gospel by transportation officials at lower levels. Unfortunately, these documents have justified road designs that prioritize vehicle speed over safety and access for everyone else. Tragically, and all too often, transportation professionals fail to take action until after a deadly threshold has been crossed. For example, residents in the Southwest neighborhood of Washington, DC identified an intersection as extremely dangerous with scores of crashes, dozens of injuries, and two deaths in just a few years. However, the local authorities failed to take action until another pedestrian was killed there in April 2025.

Intersection at 6th and M Street SW, Washington DC.

The same story played out in San Antonio, where it took dozens of crashes and multiple deaths at a dangerous intersection before the city finally decided that a traffic light was “warranted.”

The Manual on Uniform Traffic Control Devices for Streets and Highways (MUTCD) is a massively influential publication that sets standards for numerous aspects of road signs and is frequently invoked in the design and operation of roads across the country. A clear example of using the wrong evidence and prioritizing speed over safety that has dominated the MUTCD is the infamous 85 percent rule–a ridiculous practice that reads like an Onion headline. It basically means setting or adjusting speed limits on roads based on how fast cars are already driving—rather than designing speed limits based on the context and safety needs of the street. Because the majority of drivers tend to exceed speed limits, this creates a feedback loop that pushes speed limits up and makes roads more dangerous over time. And speed is a crucial factor in determining the likelihood and severity of car crashes. Simply put, high speeds and safety are incompatible when it comes to roads, especially in complex settings where there are multiple types of road users (e.g. pedestrians, transit, bicyclists, and motorists).

With the most recent MUTCD edition and in statements about the manual, the FHWA has insisted that the 85th percentile rule is just one factor among many that should be considered when setting speed limits. However, the term appears forty-four times in the document (and even more in a recent FHWA Speed Limit Setting Handbook) and thus remains an important standard that transportation professionals will most likely continue to view as a key input for determining road speeds.

By creating or adopting design guidance that prioritizes speed over safety, the FHWA reaffirms that the status quo is acceptable and provides legal cover for local agencies who follow their designs. Such standards can result in road designs that are clearly dangerous.

What can we do?

The federal government can play a significant role in determining the safety of our roads and should make changes to reverse the situation:

• Require science-based standards at FHWA that prioritize the safety of all road users, including within the MUTCD and any other guidance documents. These should be updated and improved regularly based on the evidence.
• Make clear that there is no legal shield for transportation agencies simply because they follow design standards, unless those standards are demonstrably tied to the safety of all users. Federal regulation should only include design standards that are backed by publicly available evidence showing that they reduce crashes that result in fatalities and injuries in the contexts where they’re applied.
• Provide guidance for context-sensitive speed limits (limits that are based on surrounding road and land use, not simply driver behavior) and require the posted speed and design speed to be the same. The 85th percentile rule is NOT the only way to establish speed limits.

There are a number of alternate proposed methods, including NACTO’s approach (which aligns with MUTCD guidance). Some states, such as Florida, have taken steps to create context-sensitive street design guidance. Under the Florida system, planners and engineers must consider existing and future characteristics to ensure that roads are designed for the right vehicle speeds, road users, and trip types. But the federal government should play a more active role in guiding and supporting states and localities. The U.S. can also learn from international best practices in this domain.