Speed limits don’t matter

Back in July, AAA released the findings of a multi-year study it conducted of several speed limit update projects across the United States. The study included cases where speed limits were both raised and lowered, and touched virtually all road types. You can read my summary and find a link to AAA’s discussion in the link above, but the salient points are these: While crash frequency and severity did correlate with higher speeds, the most statistically reliable trends that emerged had nothing to do with property damage or personal injury. It was enforcement outcomes that were most demonstratively impacted. In other words, when speed limits go up, fewer people get tickets; when they go down, the opposite happens. 

Many weren’t happy with AAA’s conclusions (or lack thereof) and it precipitated some rather heated social media discussion, with much of the thermal overload coming from those who took issue with the study’s fundamental validity. But those complaining the loudest weren’t citing other studies like AAA’s, instead they were citing something far more fundamental: physics. And they’re not wrong. Energy increases with the square of velocity; the quicker a car moves, the more exponentially dangerous it becomes to crash it, especially into something (or somebody) else. But AAA wasn’t performing crash tests in a controlled environment; it was studying the behavior that resulted from changing the rules

Inherently, rules exist to impart order to a system that doesn’t inherently provide it. If everybody followed the rules — and I mean every single one of them, from precise adherence to manufacturing specifications and tolerances all the way down to wearing seatbelts, adjusting our mirrors and signaling properly — crashworthiness would be irrelevant. Defective parts? Eliminated. Driver mistakes? Gone. But if I need to explain to you why that’s a pipe dream, then you should probably bail out now and queue up another episode of “Care Bears.” 

It follows then that the more intuitive a given system is, the fewer rules you need in order for people to navigate it. Look no further than Germany — a country known for its fondness for order, yet its open stretches of autobahn are unrestricted straight-line paradises. Highway speeds that would be considered straight-up antisocial here in the USA don’t raise an eyebrow in Bavaria, yet its road fatality rate hovers around 1/3 of America’s. Why? Because the other rules are all so super? Or is it possible that it’s not about the rules at all?

Behavioral science is hardly exact; it becomes far more fraught when we try to measure it using hard math. But quantifying data is the best (and only, really) way of objectively analyzing it. Admittedly, When I first reviewed AAA’s results, I questioned whether the study was worth releasing at all, but even inconclusive data can be useful. In this case, the fervor behind the discussion brought into relief a fact that plagues regulators the world over: at the end of the day, the rules just don’t matter. Whether we’re driving, walking or brushing our teeth, we instinctively operate within our comfort zone.

While it’s probably easiest to believe that speed limits are established by vigorous debate between hard-nosed safety advocates and people with somewhere to be, the reality is that they’re dictated by us. This is the basis for the 85th percentile speed. If you looked at AAA’s study, you saw this term referenced several times. It’s important that you understand the concept, because it upends the notion of speed limits being set by bureaucrats in a closed-door session. 

Its description may be verbose, but the concept is actually quite simple. 85th percentile speed is the average speed at which 85% of drivers travel on a given stretch of road (or one just like it, for the purposes of new construction). For our purposes, consider it synonymous with “the flow of traffic,” and it’s a necessary component in setting real-world speed limits. If 85% of drivers think a given speed is appropriate, it probably is. If it’s not, that means one of two things: either you have an epidemic of speeding on your hands, or the planners screwed the proverbial pooch. 

.embed-container { position: relative; padding-bottom: 56.25%; height: 0; overflow: hidden; max-width: 100%; } .embed-container iframe, .embed-container object, .embed-container embed { position: absolute; top: 0; left: 0; width: 100%; height: 100%; }

With that in mind, let’s take a look at one of AAA’s cases: SW Capitol Highway in Portland, Oregon (labeled C2 in the study). Along this 0.9-mile stretch of collector road, the speed limit was reduced from 35 mph to 25 (the largest swing among the projects AAA studied). It was also the most egregious example of the law of unintended consequences. The road was also widened to accommodate bike lanes and pedestrian sidewalks during the project, which might explain why things didn’t quite go to plan. You see, rather than decreasing by nearly 10 mph to match the change in speed limit, the observed 85th percentile speeds on this stretch went up by 0.5 mph after completion. Why? I already told you: Speed limits don’t matter

The new SW Capitol Highway is certainly more bike- and pedestrian-friendly than it was before (take a look at the before/after photos here), but in terms of managing driver speeds? Not so much. If the 85th percentile speeds increased, that means drivers are more comfortable traveling at — or even above — the speeds they managed back when it was a tight two-lane road with a dirt gutter and no bike lanes. Travel times along this stretch for vehicles also decreased while traffic citations increased. AAA says there are no data available for changes in fatality rates, but there were single-digit increases in accidents causing either injury or property damage during the study period. Remember, this is the slower, “safer” version of this road. 

Now, I want you to think about your daily driving routine. Are there stretches of surface street where a true majority of the traffic goes noticeably quicker than the posted limit? Picture those roads. Are they wide, with big medians and deep building setbacks, few crosswalks and perhaps even traffic lights without pedestrian controls at all? I’m betting yes. These are cues that tell drivers they can safely open it up. How do drivers behave when these same roads are clogged by commuters or event traffic? I’m betting on aggression ruling the day, because the conditions aren’t allowing traffic to move at the speed where the majority of drivers feel comfortable. 

Many urbanists will say that the only way to design roads safely is to force drivers to feel uncomfortable going faster than the target limit, but if you ask me, that’s looking at this from the wrong perspective. The trick is to design a 25 mph road, not design just any road and stick a 25-mph limit on it. Think about your daily driving habits again, but this time, visualize roads where you’ve caught yourself unconsciously driving slower than you might otherwise — those streets where your head tells you that you’re doing a pleasant 7 mph over the limit, but you look down to find you’re actually dead-on or even under. Now, think about what those roads have in common; I’m betting they’re nothing at all like what I described above. 

Roads exist to quickly and efficiently move traffic of all kinds and should be engineered to that purpose. If Portland’s experiment has demonstrated anything, it’s that people, cars and bikes can actually coexist without dragging each other down. The result of the overhaul is a road that carries vehicular traffic at the same exact rate as before without any penalty to travel times or driving speeds. If the city simply adhered to the 85th percentile rule and raised the limit back to 35 mph, citations should return to normal (something AAA’s study actually established to a statistically significant degree) without any detriment to driver, pedestrian or cyclist safety. 

No rule is going to eliminate speeding; as long as we live in a society, antisocial behavior will exist, but artificially low speed limits and aggressive and/or automated enforcement serves nobody except the bean counters in charge of police budgeting. Forget about correcting behavior through arbitrary limits and focus instead on building good infrastructure that meets the needs of the communities it serves. As Portland inadvertently demonstrated, a good road doesn’t have to be hostile to anybody.