The Problem With Setting Speed Limits by How Fast People Drive

Q: Traffic engineers have lowered the speed limit on a road in my neighborhood to 25 mph, and a traffic study revealed the 85th percentile of 40. The speed limit should be 40 mph, since it punishes me with added time to drive 25. Is there no oversite for traffic engineers making speed limits to slow?

A: Let’s talk about the 85^th percentile. For anyone who’s asking, “The 85^th what?” here’s brief summary of this tool for setting speed limits. Measure how fast vehicles drive on a stretch of road, and whatever speed 85 percent of the vehicles are under becomes the speed limit. That’s only barely oversimplified.

What began as a guess became a rule for setting speed limits; one that is now being abandoned by many traffic engineers, and with good reason: The concept lacked evidence, was reinforced by a flawed interpretation of crash data, and evolved into something it was never intended to be.

In the mid-1920s the top cruising speed of a car was about 45 mph. By the mid-thirties that had increased to over 60 mph. Cities and states started thinking more about how to set safe speed limits. Working from the assumption that most people make safe choices, they measured the speeds of drivers and figured that 85 percent of them were probably driving at a safe speed. It was intended to be a starting point, not a rule.

The 85^th percentile concept got a boost from researcher David Solomon, who evaluated the speed of vehicles in crashes, and showed that crash frequency increased as drivers deviated from the average flow of traffic, either faster or slower. The sweet spot, if you misinterpret his research, matched up nicely with the 85^th percentile, and it soon became a rule.

But the Solomon Curve, as it is now known, included all kinds of crashes. Many of the slow-speed crashes involved vehicles turning in intersections. The problem wasn’t slow speeds; it was all the conflict points in intersections. When you take out the crashes from vehicles turning, the data show that the problem isn’t the slow vehicles, it’s the fast ones.

Many transportation departments are abandoning or at least minimizing the use of the 85^th percentile. A speed limit shouldn’t be set based just on how fast people want to drive, because it turns out that we’re not very good at assessing driving risk.

Instead, a comprehensive review of all the factors that affect the safety of a roadway should inform the setting of speed limits. We’re already familiar with this in school zones. Twice a day during the school year we drop speed limits near schools to 20 mph because of an abundance of kids in and around the road. The road hasn’t changed, but the circumstances have.

Similarly, an urban corridor might change from 35 mph to 25 mph as increasing businesses along the corridor means more driveways, more pedestrian traffic going to those businesses, and more traffic volume. Many neighborhood streets have shifted from 25 mph to 20 mph because neighborhoods are where people walk their dogs and kids learn to ride a bike.

The 85^th percentile isn’t designed to consider the safety of people who aren’t protected by vehicles, but a quarter of Washington’s traffic fatalities in 2025 were pedestrians and cyclists. Maybe you think a slower speed limit punishes you with added travel time (and I’ll counter that premise in a future column), but higher speeds punish vulnerable road users with added risk. Even if you start with the 85^th percentile, there’s a lot more to consider when setting speed limits.