Non-Pedestrian Deaths Fell 10% in 23 Years. Pedestrian Deaths Rose 54%. The Fix Has Never Been on the Road.

Between 2000 and 2023, the number of Americans killed inside vehicles dropped from 37,182 to 33,587. Down 9.7%. Airbags matured, electronic stability control became mandatory in 2012, and crumple zones evolved through three generations of increasingly sophisticated energy management that earned their keep.

In the same 23 years, the number killed outside vehicles went the other direction. Pedestrian fatalities rose from 4,763 to 7,314, a 53.6% increase.^[1] Their share of all traffic deaths climbed from 11% to 18%. One in five people killed on American roads in 2023 was walking.

Everyone knows the standard explanations: phones, SUV hood heights, darkness, speed. We have covered all of them on this site. None of them explain this. A new study from Florida Atlantic University, published this month in the Journal of the American Planning Association, argues they are all downstream of something more basic: where we put the Walgreens.^[2]

Eric Dumbaugh and his team analyzed 334 road segments and 489 signalized intersections across 222 miles of Florida arterials in Miami, Orlando, and Tampa. What matters: grocery stores, pharmacies, gas stations, and fast-food restaurants located on high-speed, multi-lane arterials are a primary predictor of pedestrian crash severity, not a background correlate or a statistical footnote but the variable that moves the needle.^[2]

Sweden, Germany, and the United Kingdom do not allow this. They separate commercial destinations from high-speed traffic by design, using zoning codes that predate the word "Vision Zero" by decades. Not mysterious. American pedestrians die at four times the rate of Brits, three times the rate of Canadians, and more than thirteen times the rate of Norwegians.^[3]

Run the arithmetic on that British comparison. The U.S. pedestrian fatality rate in 2023 was approximately 2.18 per 100,000 residents. Britain's rate sits at roughly one-quarter of that: 0.55 per 100,000. Apply the UK rate to the American population of 335 million and you get 1,826 pedestrian deaths against an actual count of 7,314. That gap is 5,488 people per year who would still be alive if Americans walked to destinations in environments designed like British ones. That is fifteen people every single day, killed not by any vehicle defect or driver impairment, but by a land-use pattern that forces them across arterials to reach places that should never have been sited there.

And yet. NHTSA just announced that total traffic deaths fell to an estimated 36,640 in 2025, down 6.7% from 2024, with a fatality rate of 1.10 per 100 million VMT, the second lowest in recorded history.^[4] The GHSA reported a 10.9% drop in pedestrian deaths in the first half of 2025, the largest decrease in 15 years of tracking.^[5] The headlines wrote themselves: progress. Look closer.

Zoom out. That "historic" pedestrian decline still leaves H1 2025 deaths 2.5% above pre-pandemic 2019 levels.^[5] We are celebrating a return toward a number that was already catastrophic. Pandemic-era reckless driving created an anomalous spike from 2020 to 2022; the regression to the mean looks like a victory only if you never look at the baseline.

Fifty-plus U.S. jurisdictions have adopted Vision Zero, and only New York City has achieved a net reduction in traffic deaths.^[2] The FAU study offers an uncomfortable explanation for that batting average. Vision Zero focuses on road engineering: narrower lanes, better signal timing, refuge islands, all interventions worth pursuing. All insufficient when the fundamental problem is that your zoning code requires a human being to cross six lanes of 45 mph traffic to buy milk.

Limitations

The international comparison above uses population-adjusted rates, not exposure-adjusted rates. Americans drive far more per capita than Europeans, which inflates the pedestrian encounter rate. Some of the 53.6% increase since 2000 may reflect more Americans walking and cycling, not just more danger per trip. Dumbaugh's team covered only Florida arterials; generalizing to the full U.S. assumes similar land-use patterns nationwide (a reasonable but untested assumption). FARS records fatal crashes, not the land-use context surrounding them; the causal chain from zoning to fatality runs through the FAU study, not directly through FARS.

The Counterargument

Road design does matter. Dumbaugh's own data shows that road characteristics, including speed and lane count, contribute to pedestrian crash severity alongside land use. Europe has lower pedestrian deaths for multiple reasons beyond zoning: lower speed limits, more robust enforcement, less distracted driving, denser transit networks that reduce exposure. Isolating land use as "the" cause overstates a multifactorial problem.

That counterargument is correct and incomplete, because Dumbaugh's data shows that land use is the variable most strongly associated with crash severity on the corridors studied, even after controlling for road geometry. Speed kills, certainly, but it kills with particular efficiency when your pharmacy is built on a stroad. That compounding is the point: you cannot engineer your way out of a land-use problem, and the 50-city Vision Zero failure rate suggests that is exactly what America has been trying to do.

What You Can Do

Check your own commute. If you walk to any destination that sits on a multi-lane arterial with no median refuge, no signalized crosswalk within 300 feet, and a speed limit above 35 mph, you are occupying the exact environment the FAU study identifies as high-risk. Use crosswalks where they exist, and lobby your city council for median refuges and signalized mid-block crossings. Better yet, attend your next zoning board meeting. Oppose commercial variances that put high-traffic retail on arterials without pedestrian infrastructure. No amount of engineering will save you if the zoning already chose to endanger you.