Apex’s police chief told the Public Safety Committee that vehicle thefts in town are down 60% since the Flock cameras arrived. Vehicle thefts in Apex are down. They are also down in Wake Forest, which has had twenty-five cameras since 2023 and fell further than Apex, and in Knightdale, which runs more cameras than Apex and barely moved, and across North Carolina by 22.6% in the first half of 2026, and across the United States by 23.2% in 2025. Here is the full record, taken from the FBI’s agency files and the Police Department’s own slides.
On October 1 the Apex Public Safety Committee took up the camera program and recommended nothing. The one number that made the evening news was the Chief’s: a 60% reduction in vehicle theft. It is the first crime-reduction figure the Department has offered publicly as a reason to keep this system, and it will be the centerpiece of whatever comes back on December 3.
So it is worth checking. We pulled Apex’s vehicle theft counts from the FBI’s Crime Data Explorer, which publishes what Apex Police themselves submit, month by month, going back a decade. We pulled the same series for nine neighboring agencies. We read the Department’s own June 9 presentation for the camera dates. Everything below is checkable, and every figure has a link at the bottom of the page.
The decline is real. Apex vehicle thefts did fall, and they fell by roughly the amount the Chief described. We are not saying he made the number up, and we are not calling it a rounding error or statistical noise. A town of about 79,000 going from 36 thefts to 13 over nine months is a real drop in a real number.
What is unsupported is the attribution. The claim that the cameras caused the drop is the part with no evidence behind it, and the evidence against it is sitting in the FBI’s public files. Thefts fell in every comparable town in this region, by amounts ranging from 10% to 82%. Every one of those towns runs plate readers of some kind, the camera counts run from seven to more than thirty, and the deployment dates span three years. The size of the fall lines up with none of it.
We are not arguing that ALPRs never help anyone. They recover vehicles. The Department has described cases. Our position has never rested on the cameras being useless. It rests on a town of about 79,000 photographing roughly 60,000 plate reads a day into a database that roughly 3,000 outside organizations can query, with no ordinance, no hearing and no vote. A crime statistic does not answer that, and a contested crime statistic answers it even less.
“It’s not a huge, expansive network, right? But during that period of time, you had 60% reduction in vehicle theft in town.”
Chief Ryan Johansen, Apex Police Department, at the October 1 committee meeting. Quoted by ABC11 and CBS 17, reported October 2, 2026Read that sentence again and notice what is not in it. No start date. No end date. No comparison period. No counts, in either direction. No source. “That period of time” refers back to the reporter’s framing of roughly ten months, which places the start somewhere around December 2025.
A percentage with no numerator, no denominator and no dates is not a finding. It is a headline. The Department holds the underlying counts and could have brought them to the committee on a single slide. It did not, and nobody in the room asked.
These are Apex Police Department’s own submissions to the FBI, retrieved from the Crime Data Explorer on October 2, 2026. Years 2016 through 2025 have all twelve months reported. 2026 has eight, January through August.
| Year | Reported thefts | Cleared | Clearance rate | |
|---|---|---|---|---|
| 2016 | 17 | 6 | 35% | |
| 2017 | 14 | 4 | 29% | |
| 2018 | 12 | 5 | 42% | |
| 2019 | 19 | 1 | 5% | |
| 2020 | 18 | 2 | 11% | |
| 2021 | 24 | 4 | 17% | |
| 2022 | 39 | 4 | 10% | |
| 2023 | 54 | 2 | 4% | Highest in this series. No Apex cameras in service. |
| 2024 | 29 | 4 | 14% | Down 46% from 2023. Still no cameras in service. |
| 2025 | 41 | 4 | 10% | First cameras in use October. Alerts on in December. Thefts up 41%. |
| 2026 | 12 | 2 | 17% | January to August. Annualizes to about 18. |
Three things in that table deserve to be said out loud.
The biggest single drop in the eleven years the FBI publishes for Apex happened before the cameras existed. Thefts fell from 54 to 29 between 2023 and 2024, a 46% decline, in a town with no ALPR cameras in service. The Flock contract did not even take effect until February 1, 2025.
2025, the first camera year, was worse than the year before it. Cameras were collecting plate reads from October 2025 and Apex finished that year with 41 thefts against 29 in 2024. If the cameras drive the number, that year is unexplained.
Apex has not reached a new low. It has returned to its own normal. From 2016 through 2021, before the national theft wave, Apex averaged 17.3 vehicle thefts a year. The 2026 pace annualizes to about 18. January through July of 2026 produced 12 thefts, which is more than the 9 Apex recorded over the same months in 2019. The town is back where it was, not below it.
Scroll the chart sideways to read it.
Apex Police Department reported motor vehicle theft, 2016 to 2026. Camera dates are the Department’s own, from its June 9, 2026 presentation to Town Council: “First cameras in use as of October 2025, alerts activated December 2025.”
The Chief gave no dates, so we have to infer the window. The roughly-ten-months framing in the coverage puts the start around December 2025, which is the month APD’s own slide gives for alerts being activated. Run that window and the figure reproduces, which is some evidence the inference is right. Worth noting that the same slide describes the program as “less than 6 months of full use” as of June 2026, and that cameras were collecting reads from October 2025, which is twelve months rather than ten.
So the arithmetic holds. Now take one month, August, off both ends of that comparison, and watch what happens.
August 2025 was a single month with 10 thefts in it, almost a quarter of that year’s total. August 2026 had none. So one pair of months carries 10 of the 23 vehicles behind the headline, and the nine-month figure needs that pair inside it to reach 64%. Take August off both ends and the decline is 50%. When a result moves fourteen points on the strength of one month, the result is partly a property of the window.
There is a name for the pattern underneath this. Apex’s baseline was an anomalous spike, 54 thefts in 2023 against a long-run average near 17, and anomalous spikes come back down on their own. Criminologists call it regression to the mean, and it is a well-documented way a place-based crime program gets credited with a decline it did not produce. A 2026 working paper by McCrain, Adams and Nix, not yet peer reviewed, demonstrates the problem directly: they ran a placebo test in which the intervention was defined to have no effect by construction, and standard methods still generated a statistically significant “effect” comparable in size to the reductions in the published literature. Ian T. Adams is a co-author of both that paper and the pro-camera study discussed further down.
This is the part the committee should see. If cameras drive vehicle theft down, towns with more cameras should fall further, and towns with none should barely move. Here are ten Triangle agencies on identical data, identical months, identical source.
| Agency | Jan–Jul 2025 | Jan–Jul 2026 | Change | ALPR cameras |
|---|---|---|---|---|
| Holly Springs | 11 | 2 | −82% | Has ALPRs, plus mobile units |
| Wake Forest | 29 | 12 | −59% | 25, live since September 2023 |
| Apex | 24 | 12 | −50% | 7, live since October 2025 |
| Clayton | 14 | 8 | −43% | About 18 reported |
| Morrisville | 29 | 19 | −34% | Not established |
| Cary | 94 | 63 | −33% | Its own ALPR on Axis hardware, not Flock |
| Garner | 52 | 40 | −23% | About 20 reported |
| Fuquay-Varina | 12 | 10 | −17% | About 16 reported |
| Raleigh | 1,000 | 894 | −11% | 32 police-operated |
| Knightdale | 21 | 19 | −10% | 20 to 30 reported |
Scroll the chart sideways to read it.
Change in reported motor vehicle theft, January to July 2026 against the same months in 2025, across ten Triangle police agencies. FBI Crime Data Explorer. Every agency shown uses plate readers in some form; the deployment column records what we could establish about each one, and the notes below say how firmly.
Take the rows that matter most.
Across these ten agencies the decline runs from 10% to 82%. Every one of them uses plate readers in some form. The camera counts run from seven to more than thirty, the vendors differ, and the deployment dates span three years. None of those variables lines up with the size of the fall. That is what it looks like when a regional trend is moving everything and the local equipment is incidental to it.
There is a straightforward explanation for why vehicle thefts fell in Apex, Wake Forest, Knightdale, Cary, North Carolina and the rest of the country at the same time, and it has nothing to do with license plate readers. The 2022 and 2023 theft wave was driven substantially by a specific vulnerability in Hyundai and Kia vehicles built without engine immobilizers. Those manufacturers began pushing an anti-theft software upgrade, and the Highway Loss Data Institute measured a 46% lower theft claim frequency on upgraded vehicles than on comparable vehicles that had not been upgraded.
| Measure | Figure | Source |
|---|---|---|
| US vehicle thefts, 2023 | 1,020,729 | NICB, a forty-year peak |
| US vehicle thefts, 2024 | 850,708 | NICB, down 17%, largest drop in forty years |
| US vehicle thefts, 2025 | 659,880 | NICB, down 23.2%, 190,828 fewer vehicles |
| NC motor vehicle thefts, 2023 | 25,499 | NC SBI |
| NC motor vehicle thefts, 2024 | 24,775 | NC SBI, rate down 2.6% |
| NC change, first half 2025 | −17% | NC DPS, against first half 2024 |
| NC change, first half 2026 | −22.6% | NC DPS, against first half 2025. Steepest decline of any property offense. |
| Hyundai and Kia share of US thefts | 21% → 14% | NICB, 2023 to 2025 |
| Hyundai and Kia theft claim frequency, upgraded vehicles | −46% | HLDI, against non-upgraded |
| Whole-vehicle theft claim frequency, upgraded vehicles | −52% | HLDI |
North Carolina fell 22.6% in the first half of 2026 against the first half of 2025. That figure comes from a sample of 221 North Carolina law enforcement agencies covering about 60% of the state’s population, so it is a statewide estimate rather than a complete count, and Apex’s seven cameras did not produce it. Compounded with the 17% drop the year before, statewide first-half vehicle theft is down roughly 36% in two years.
Note also what the NICB credits for the national decline: law enforcement coordination, vehicle manufacturers, insurers, state auto theft prevention authorities and stronger penalties. We read its last two annual releases and automated license plate readers are not named in either one.
One honest caveat on our side. The coverage of the Hyundai and Kia fix was still only around 40% of eligible vehicles as of the end of 2024, which means the effect was still building through 2025 and 2026. That is the exact window the Chief is claiming credit for, and it cuts against the camera explanation rather than for it.
Set the prevention argument aside and look at the other thing cameras are supposed to do: help close cases. Apex’s own clearance figures do not move.
That last number needs care, because it will be offered as a rebuttal. In its June 9 presentation the Department reported more than 1,700 hot list hits, including 111 for stolen vehicles. A hot list hit is not a recovery and it is not an Apex theft. It is a plate read of a vehicle entered as stolen somewhere in the Flock network, passing an Apex camera. Apex recorded about 41 vehicle thefts in all of 2025, so 111 hits cannot be 111 Apex cases.
The gap is the question, and we are asking it rather than answering it. The two figures are not strictly comparable: a hot list hit is a cumulative count of network-stolen plates passing an Apex camera, while a clearance is an Apex offense closed, and a clearance can close a case from an earlier year. Even allowing for that, 111 against six is wide enough that somebody on the committee should ask what happened in between.
Atlanta is the natural experiment on this point. It is Flock’s headquarters city, with more than 28,000 integrated cameras. Its motor vehicle theft clearance rate went from 9.9% in 2021 to 10.2% in 2025, a change of three tenths of a percentage point, across a period in which its camera network grew enormously.
The United States Department of Justice maintains CrimeSolutions, a clearinghouse that rates crime prevention programs on the strength of the evidence behind them. Its rating for license plate recognition technology as a crime deterrent is one word.
“The program is rated Ineffective. There were no statistically significant program effects on general crime, auto-related crime, vehicle-theft calls for service, or auto-theft.”
US Department of Justice, National Institute of Justice, CrimeSolutions, License Plate Recognition Technology as a Crime Deterrent. Rated on more than one study, including at least one high-quality randomized controlled trial.We will state the weakness in our own citation before anyone else does: that profile was posted in 2013 and carries a notice that it is no longer updated. It predates fixed camera networks of the kind Apex has. It is still the federal government’s standing assessment, and the newer literature has not overturned it.
On September 30, the day before the Apex committee met, CBS News published an examination of Flock’s national crime-solving claims. It asked the company’s chief executive directly whether the claims rested on peer-reviewed research.
“No, there’s no peer-reviewed study, and we’re not articulating that it was peer-reviewed.”
Garrett Langley, chief executive, Flock Safety, to CBS News, September 30, 2026CBS found that Flock’s own footnotes describe its impact figures as “directional estimates rather than audited totals” and its methodology as “not a finished product.” The widely repeated claim that Flock helps solve 10% of reported crime in the United States is not actually the finding of the paper it is attributed to, and rests on usable data from 123 agencies out of roughly 18,000.
This is not the first time. In 2024 Forbes examined Flock’s marketing claims city by city. In San Marino, California, Flock advertised a 70% crime reduction; residential burglaries over a longer window went from 60 in 2019 to 63 in 2023, and total Part 1 crimes were flat. The city’s own police chief, asked about the figure still live on Flock’s website, said he needed to talk to their marketing people. In Dayton, Ohio, a claimed 43% neighborhood reduction turned out to compare 2018 against 2020 for a system not installed until March 2020.
The point is not that Chief Johansen is doing what Flock’s marketing department does. The point is that this specific category of claim, a short-window local percentage offered without counts or comparison periods, has a documented history of not holding up, and a committee reviewing a contract should treat it accordingly.
The committee meets again on December 3 and will then make recommendations to the full Council. These are process questions. They do not require anyone to take a position today, and all three can be answered from records the Town already holds.
We do not know what source the Chief used. He may be quoting an internal records system rather than the submissions Apex sends the FBI, and those can differ. That is precisely why we are asking for the counts rather than accusing anyone of error.
Deployment details for neighboring towns vary in quality, and this is the weakest part of the page. Wake Forest’s 25 cameras and the September 2023 activation date come from the Town of Wake Forest’s own announcement, and that one is solid. Cary runs its own plate readers on Axis hardware rather than Flock. That is our understanding locally and we have not yet obtained a Town of Cary record confirming it, so treat it as unconfirmed and hold us to it. Separately, Cary PD also has access to around a dozen privately owned Flock cameras under an agreement effective March 2025, which is sourced to Cary town staff quoted in Axios Raleigh and INDY Week. The Knightdale, Garner, Fuquay-Varina, Clayton and Holly Springs figures come from vendor transparency portals, news coverage and crowdsourced mapping, and we label them approximate for that reason. Morrisville we could not establish at all, and the table says so.
On the December 3 date. We were in the room on October 1. The committee said it would pick the ALPR item back up on December 3. ABC11 reported on October 2 that the committee “is scheduled to meet again Nov. 5 to further discuss Flock cameras before making recommendations to the Town Council,” which does not match what was said; CBS 17 reported only that the next discussion would be in December. We are going with December 3 because that is what the committee stated. As of this writing no date appears on the Town calendar or the Council Committees page. File a written notice request with the Clerk under G.S. 143-318.12(b)(2) so the Town has to tell you, and confirm before you plan around it.
Figures for 2026 are partial and submissions can be revised. Coverage differs by agency: Cary, Holly Springs and Raleigh have reported nine months, Apex and several others eight, and Fuquay-Varina and Garner seven. We used January to July because that is the longest window in which every agency in this group has reported every month.
If a number on this page is wrong, tell us which one and where, and we will correct it and say that we did. Email contact@deflockapex.org.