And Then You're Dead

The Autopsy / Case File · Vehicular & Transit

After the Cameras

Red-light and speed cameras are one of the few traffic-safety interventions Americans consistently vote to remove — cited as revenue grabs, privacy intrusions, or simple annoyances, and repealed or defunded in dozens of cities over the past two decades. A study designed specifically to isolate what happens next found a real, measurable cost.

Researchers at the Insurance Institute for Highway Safety tracked 14 large U.S. cities that shut down their red-light camera programs between 2010 and 2014, comparing their crash trends against 29 similar cities that kept cameras running over the same years. After cameras came off, the rate of fatal red-light-running crashes was 30 percent higher than it would have been had the cameras stayed on — and the rate of all fatal crashes at signalized intersections, not just red-light-running ones specifically, was 16 percent higher. It was the first study built specifically to isolate the effect of turning cameras off, rather than just measuring the effect of turning them on.

And Then You're Dead · September 2026 · 9 min read

Join the discussion →
Red-light and speed enforcement cameras mounted on a traffic signal pole above a pedestrian crossing signal showing a red hand, against a clear blue sky
Automated enforcement cameras mounted alongside a traffic signal in San Francisco. A 2017 IIHS study found that removing cameras like these measurably raised fatal crash rates at the intersections that lost them. Photo: Pi.1415926535, CC BY-SA 4.0, via Wikimedia Commons.

Automated red-light enforcement has followed roughly the same political arc in dozens of American cities since the early 2000s: a program launches, citations and revenue flow, public complaints accumulate — about tickets arriving by mail for a rolling right turn, about a private vendor taking a cut of every fine, about surveillance — and eventually a city council vote, a ballot referendum, or a state legislative ban shuts the cameras off. Cities that have deactivated programs over the years include Los Angeles, Houston, and dozens of smaller municipalities, driven by a mix of genuine privacy and fairness concerns and, in some documented cases, straightforward political backlash against being ticketed. What's been harder to isolate, until relatively recently, is what actually happens to the underlying crash and fatality numbers once the cameras are gone — as opposed to what happens when they're installed, which has been studied far more often.

A Study Built to Answer the Removal Question Specifically

The Insurance Institute for Highway Safety, an industry-funded but methodologically independent research organization whose crash studies are widely cited in transportation policy, published the relevant analysis in the Journal of Safety Research in 2017. Researchers Wen Hu and Jessica Cicchino examined 117 large U.S. cities with populations over 200,000, tracking 57 cities that activated red-light camera programs between 1992 and 2014 and, separately, 19 cities that later deactivated theirs. The core comparison used Poisson regression — a statistical method suited to modeling relatively rare events like fatal crashes — controlling for underlying trends in crash rates over time, population density, and local unemployment, to isolate the camera program's own effect from everything else changing in a city simultaneously.

For the narrower group of 14 cities that turned off their cameras specifically during 2010 through 2014 — a tighter time window chosen because it let researchers compare them against a matched set of 29 cities in the same regions that kept their programs running over those same years — the results were both statistically significant and substantial: fatal red-light-running crashes ran 30 percent higher than expected after camera removal, and all fatal crashes at signalized intersections, a broader category capturing collisions beyond the red-light-running mechanism specifically, ran 16 percent higher. Across the full 19-city group of all cities that deactivated at any point in the study period, the increases were smaller — 18 percent and 8 percent respectively — and not statistically significant on their own, a gap the researchers attributed to the wider group's noisier mix of city sizes, timing, and local conditions diluting a real effect visible more clearly in the tighter, better-matched 14-city cohort.

0
percent higher fatal red-light-running crash rate after 14 cities turned off red-light cameras, 2010–2014, vs. matched cities that kept cameras on
0
percent lower fatal red-light-running crash rate in cities with active camera programs, compared to similar cities that never installed them

The Activation Side of the Same Study

The same research also quantified the other half of the picture — what cameras do while they're running, not just when they're removed. Comparing 57 cities that had installed camera programs against 33 similar cities that never had, IIHS found the fatal red-light-running crash rate was 21 percent lower in camera cities, and the rate of all fatal crashes at signalized intersections was 14 percent lower. Those figures track closely with the removal-side findings — activation lowers the same crash categories that deactivation raises, by a roughly comparable magnitude, which is exactly the pattern a genuinely causal safety effect should produce rather than one driven by some other citywide trend that happened to move at the same time as the camera decision.

The researchers weren't just asking whether cameras work. They built the study specifically to answer what a city gives up, in fatal crashes, the moment it decides to take them down.

Why "Turned Off" Isn't a Neutral Choice

The mechanism behind both halves of the finding is straightforward from a driver-behavior standpoint: red-light cameras work primarily through deterrence, not enforcement after the fact — a driver who believes a camera is watching a given intersection is measurably less likely to run the light there, and that reduced-risk-taking effect appears to fade relatively quickly once local drivers learn, through word of mouth or simple repeated observation, that the camera is gone. That's part of why IIHS and other safety researchers have pushed back specifically against the framing that automated enforcement is primarily a revenue tool — a genuine revenue-focused program would have little reason to actually change driver behavior, whereas the deterrence-driven crash reduction IIHS documented, and the corresponding increase after removal, is the signature of an intervention actually altering how people drive through an intersection, not just generating tickets after the fact.

None of that resolves the legitimate concerns that have driven camera repeals in the first place — due-process objections to citations issued to a vehicle's registered owner rather than a proven driver, concerns about private vendors profiting from enforcement, and documented cases of cameras deployed more heavily in lower-income neighborhoods than wealthier ones. IIHS's more recent guidance for automated enforcement programs has specifically emphasized addressing those equity and due-process design questions rather than treating them as unrelated to the safety case, aiming at running a program that's both effective and fair rather than treating "keep the cameras" and "fix the program's flaws" as mutually exclusive options. What the removal-side data adds to that debate isn't an argument that every concern about camera programs is invalid — it's a specific, quantified answer to the question of what a city is trading away, in additional fatal crashes at intersections that used to have them, the moment it decides removal is the fix.

A Patchwork That Predates Any One City's Vote

Camera removal in the United States has never happened through one uniform mechanism, which is part of why isolating its effect required the kind of matched-city statistical design IIHS used rather than a simple before-and-after count in a single place. Some cities have ended programs through direct city council votes after a change in local political leadership. Others have lost their cameras to statewide legislative bans — several states have passed laws prohibiting or sharply restricting automated traffic enforcement outright, overriding any individual city's preference to keep a program running. Still others have let vendor contracts lapse without renewal, a quieter form of the same outcome that draws less public attention than an outright repeal vote but produces an identical result at the intersection: a camera that used to be watching is no longer there. IIHS's 14-city cohort deliberately spanned this variety, which is part of why the researchers treated it as a real test of what removal does regardless of the political mechanism that produced it, rather than a study of any single city's particular controversy.

And Then You're Dead

A city council vote to shut off a red-light camera reads, on the surface, like a small administrative decision — one line item, one contract ended. The intersection doesn't know the camera is gone until a driver runs the light and nobody, and nothing, was watching to make them think twice.

That's it. That's the whole thing.

Comments

Loading…

Comments are anonymous and moderated.

Sources