The Camera Tax Project

The reframe

A camera fine is a tax with a safety story attached.

Call something a safety device and it slips past the questions we ask about any tax. Who imposed it, who voted on it, who pays it, and what happens to the people who cannot.

Traffic fines are not assessed on ability to pay. They are flat. A $100 citation is an annoyance to one household and a crisis in another, and the machine that issued it cannot tell the difference.

That is what a regressive tax is. And in most places it is one that was never put to a vote, gets decided outside a courtroom, is run by a private contractor, and ends in a boot, an impound lot, or a suspended license when it goes unpaid.

Vision Zero Network, which supports camera enforcement, concedes the point directly: flat fines "are inherently regressive" and can "reinforce harmful poverty traps."

The same fine. Wildly different burdens.

A single $100 camera citation, expressed as a share of one week's gross pay.

$25,000 / yr20.8%

$50,000 / yr10.4%

$100,000 / yr5.2%

$250,000 / yr2.1%

Most of Europe solved this decades ago. Finland, Germany, Austria, France, and Switzerland scale traffic fines to income. Chicago piloted a 50 percent reduction for low-income residents. The tools exist. American cities have chosen not to use them.

Where this came from

A rally driver built it to make himself faster.

Maurice Gatsonides won the 1953 Monte Carlo Rally. He built his first speed-measuring rig to time his own cornering so he could improve his driving. Only later did anyone realize the same device could be pointed at everyone else.

1958

Gatsometer BV is founded in Haarlem

Two rubber tubes across the road wired to a stopwatch. The company Gatsonides founded still supplies enforcement cameras today.

1965

The first red light camera

Deployed in the Netherlands. Radar for road traffic follows around 1971, the first mobile speed camera around 1982.

1986

The technology reaches America, and Americans reject it immediately

Friendswood, Texas runs the first US photo-radar program in 1986. La Marque follows in 1987. Neither lasts more than a few months before public pressure ends it. Resistance to automated ticketing in this country is as old as the technology's arrival here.

1987

Paradise Valley, Arizona makes it permanent

The first US photo-radar program that sticks.

1988

Nottingham installs the first UK red light camera

Put in after a collision that killed three people. First prosecutions that July.

1992

New York City turns on the first US red light cameras

Authorized by a 1988 state demonstration program limited to cities over one million people, then extended again and again.

May 1992

Britain's first speed camera catches 22,939 drivers in 22 days

On the A316 near Twickenham Bridge. The volume was the point, and it was visible from the very first installation.

2012

Red light cameras peak at roughly 533 US communities

Then the decline begins. More programs get shut down than started, every year, for a decade.

2026

Speed cameras overtake red light cameras

The public beat the red light camera. The industry pivoted.

The pivot

Americans defeated one machine. A second one was already going up.

Red light camera programs have been shrinking since 2012. Speed camera programs more than doubled in about four and a half years. As of 2026 there are now more US communities running speed cameras than red light cameras, reversing two decades of the opposite pattern.

Red light camera communities Speed camera communities

Source: Insurance Institute for Highway Safety community trackers. IIHS does not publish a single continuous year-by-year table, so intermediate years are drawn from IIHS releases and reporting citing them. Figures are community counts, not camera counts.

The record

When Americans are allowed to vote on this, they vote it down.

Camera enforcement has been put to a public vote at least 46 times. It lost roughly 42 of them, often by margins that do not exist anywhere else in American politics. Sulphur, Louisiana rejected cameras 86 to 14. League City, Texas voted 77 percent to ban them.

These are not close calls, and they are not confined to one region or one party. Voters in Ohio, Texas, Washington, California, Louisiana, Maryland, Missouri, New Mexico, and Alaska have all reached the same conclusion when handed a ballot.

Jurisdiction	Date	Result	Outcome
Sulphur, Louisiana	2009	86%	Cameras out
League City, Texas	2013	77%	Cameras out
Cleveland, Ohio (Issue 35)	Nov 2014	~77%	Cameras out
Anaheim, California (Measure K)	Nov 2010	~73%	Banned preemptively
St. Charles County, Missouri	Nov 2014	~72%	Cameras out
Chillicothe, Ohio	Nov 2009	72%	Cameras out
Mukilteo, Washington	Nov 2010	~70%	Cameras out
Dayton, Texas	Nov 2011	70%	Cameras out
Bellingham, Washington	Nov 2011	~65%	Cameras out
Monroe, Washington	Nov 2011	~65%	Advisory
Sykesville, Maryland	2010	61%	Repealed pre-launch
Ashtabula, Ohio	Nov 2011	60%	Cameras out
Arlington, Texas	May 2015	59.5%	Cameras out
Longview, Washington	Nov 2011	59%	Advisory
Baytown, Texas	Nov 2010	58%	Cameras out
Newport Beach, California (Measure EE)	Nov 2012	57%	Cameras out
Murrieta, California (Measure N)	Nov 2012	57.2%	Cameras out
South Euclid, Ohio	Nov 2011	55%	Cameras out
Garfield Heights, Ohio	Nov 2011	54%	Cameras out
Houston, Texas (Prop 3)	Nov 2010	52.8%	Overturned in court
East Cleveland, Ohio	Nov 2011	>54%	Voted to keep

A partial list of votes with confirmed margins. Percentages are the share voting against camera enforcement, except the final row. Several additional votes are documented without published percentages.

The exception

East Cleveland voted to keep its cameras, and told you exactly why.

In the one clear case where a community voted to keep camera enforcement, the mayor had warned that roughly 68 city jobs, about a third of the workforce, depended on ticket revenue. That is not a safety argument. It is a budget argument, made out loud. We include this vote because leaving it out would be dishonest, and because it proves the point better than any of the wins.

Houston, 2011

A city sued to overturn its own voters, and won.

Houston voters repealed red light cameras in November 2010. The city and its camera vendor went to federal court, where a judge ruled the ballot measure was an untimely referendum and therefore invalid. The voters were overruled. The City Council shut the cameras off anyway in August 2011, but the precedent stands: when the public voted the machines out, the government litigated to keep them.

Georgia, 2008

One second of yellow did what the cameras claimed to do.

Georgia passed a law requiring yellow lights at camera intersections to run one second longer than the national minimum. It took effect at the end of 2008. What happened next is the most damaging fact in this entire debate.

1,468 → 313

Lilburn citations, January 2008 compared to January 2009

15 → 3

Norcross citations per day, before and after

~80%

decline in violations across affected programs

cities that ended their camera programs afterward

Duluth, Lilburn, Norcross, Snellville, and Suwanee shut their programs down. Not because the cameras stopped working. Because they stopped paying for themselves.

The cheap engineering fix worked. The programs died anyway, because there was no money in it.

The reverse experiment has also been run. A Chicago Inspector General investigation found the city quietly dropped its yellow-light floor from 3.0 seconds to 2.9 while switching camera vendors, producing 77,000 additional citations at $100 each. In Florida, the state transportation department removed language from its engineering manual that had kept yellows longer, and citation volume rose. Half a second can roughly double the number of tickets an intersection produces.

When the machine is wrong

Baltimore ticketed a minivan that was sitting still.

The city told the public its speed camera error rate was under one quarter of one percent. A confidential audit commissioned by the city sampled roughly 1,000 citations from a single day in 2012 and found the real error rate was above 10 percent, roughly 40 times what officials claimed.

One camera on Loch Raven Boulevard was wrong 58 percent of the time. Thirteen cameras had double-digit error rates. The city issued about 700,000 citations that year at $40 each, which puts roughly 70,000 bad tickets and $2.8 million in the hands of people who had done nothing.

The program shut down in 2013. When Baltimore relaunched automated enforcement in 2017, the new system issued $38,480 in duplicate tickets on its first day.

The error rate is not the whole problem. The appeal is.

A camera citation is typically a civil matter, not a criminal one. There is no officer to cross-examine, the registered owner is presumed to be the driver, and the hearing is administrative rather than judicial. Contesting a $40 ticket can cost more in lost wages than paying it.

Which means the true error rate does not really matter to the balance sheet. Most people pay regardless, because fighting it is not worth the day off work. That is the design.

Where a fine goes when you cannot pay it

A $100 citation is not a $100 problem.

Step one

It doubles

Unpaid tickets in Chicago double, then accrue collection fees of 22 percent. In one documented case, less than half of a $6,700 balance was the original fines.

Step two

The car goes

Booting and impoundment follow, with their own fees. Losing the car frequently means losing the job that would have paid the ticket.

Step three

The license goes

Roughly 11 million Americans hold licenses suspended over unpaid fines and fees. Chicago sought more than 21,000 suspensions in a single year, triple its 2010 level.

Chicago motorists collectively owe about $1.45 billion in ticket debt going back to 1990. Ticket revenue reached $264 million in one year, close to 7 percent of the city's operating budget. Personal bankruptcy filings that included city debt rose from roughly 1,000 a year to more than 10,000.

Eight of the ten ZIP codes with the highest ticket debt per adult are majority Black. Majority-Black areas hold about 40 percent of the debt while receiving about 22 percent of the tickets. The disparity is not in who gets photographed. It is in who can absorb the hit.

What we do not claim

The case against us, stated fairly.

Most organizations in this space lost credibility by overclaiming. We would rather concede the strong version of the other side's argument and win the part that actually matters.

Cameras do reduce serious crashes. We are not going to pretend otherwise.

The Cochrane systematic review found injury crashes fell between 8 and 50 percent near speed cameras. A study in Montgomery County, Maryland found a 19 percent reduction in the likelihood that a crash produced an incapacitating or fatal injury. New York City recorded speeding drops of up to 94 percent at camera locations. Research that corrects for the statistical objection most often raised against these studies still finds real reductions.

Anyone telling you camera enforcement does not work is not reading the evidence. Our argument does not require that claim and never will.

A camera is less racially biased at the moment of the citation than an officer is.

A 2024 study of roughly 46 million Chicago-area trips found camera citations tracked the racial makeup of drivers actually on the road, while police stops did not. Automated enforcement also removes the pretextual traffic stop, which is where a great deal of police violence against Black drivers begins. That is a serious argument made by serious people, and we will not claim the opposite.

Our objection lives on the back end, not the front end. Not who gets photographed. What happens to them afterward when the fine is flat, the debt compounds, and the license disappears.

Roughly 39,345 people died on American roads in 2024. That number is not an abstraction to us.

The United States has about 12 road deaths per 100,000 people, more than double most peer nations and roughly triple the median among wealthy countries. Between 1991 and 2021 American fatal crashes fell about 21 percent while comparable countries fell roughly 77 percent. This is a genuine crisis and we are not indifferent to it.

We support engineering that works. Longer yellow intervals, better signal timing, road design that slows traffic without a billing department. Georgia proved a one second change can cut violations 80 percent. Those fixes save lives and generate no revenue, which is why they are not the ones getting installed.

The standard

Five conditions. Any jurisdiction that meets them can keep its cameras.

We are not asking for a national ban. We are asking that a power to tax and surveil be treated like one.

Put it to a vote

No new automated enforcement program without direct voter approval, and periodic reauthorization after that. Georgia has already adopted a referendum requirement. Ohio is considering one. Arizona has a statewide measure on the ballot.

Engineering first

Before a camera goes up, the intersection must meet minimum yellow-interval standards and the jurisdiction must document that engineering remedies were tried. No jurisdiction should be permitted to shorten a yellow light at a camera location, ever.

Scale the fine to income

A flat fine is not a flat burden. Finland, Germany, Austria, France, and Switzerland scale traffic penalties to ability to pay. Any American jurisdiction can do the same, and none of them lose a single safety benefit by doing it.

Break the revenue link

No per-citation vendor compensation. No camera revenue to the general fund. Money above operating cost goes to road safety engineering or is returned. A program that cannot survive without the income it generates is not a safety program.

No license suspension for inability to pay

Roughly 11 million Americans have suspended licenses over unpaid fines and fees. Taking someone's license because they are poor does not make a road safer. It makes the debt permanent.

It's not a ticket.
It's a tax.