Airplane flies over a drone during the Polar Bear Plunge on Coney Island in the Brooklyn borough of New York
Photograph by Carlo Allegri — Reuters

The FAA is Way Overstating the Risk Drones Pose to Airliners

Mar 16, 2016

In December, the Federal Aviation Administration issued a strict set of guidelines for small drones operating in U.S. airspace including a requirement that hobbyists register their drones with the government.

Any drone weighing more than 250 grams—or roughly half a pound—must be piloted by a registered drone operator. Additionally, the new rules restricted certain areas from drone flights and required that the pilots keep them in sight at all times.

The reason, according to the FAA: safety. Aside from the potential danger to people and property on the ground, small drones—even those weighing just a few pounds—are seen as a potential danger to civilian and commercial aviation. An uptick in reports by commercial airline pilots of “near-misses” with drones bolstered that argument.

But a study released Monday by the Mercatus Center at George Mason University suggests the FAA is overstating the risk by small drones to manned aircraft. In fact, it's way overblowing it.

The study found that, based on existing data, an incident in which an aircraft is damaged by a drone weighing 4.5 pounds should happen once every 1.87 million years of drone flight time. An injury or fatality? About 100 times less likely than that.

It’s only fair to note that there exists no real data on exactly how damaging a collision with a drone might be for, say, a 737 jetliner. But the FAA has decades of data on detailing aircraft collisions with birds, and it's from that data that the Mercatus researchers draw their conclusions.

Get Data Sheet, Fortune’s daily tech newsletter

There are roughly 10 billion birds zipping through the U.S. national airspace, and since 1990 there has been a sevenfold increase in reported bird strikes. That uptick is largely attributed to the Internet and the ease with which it allows pilots to report collisions with wildlife. But the larger point is that the national airspace is crowded with potentially dangerous large birds (1.9 million turkey vultures and 2 million to 3 million migrating snow geese annually versus roughly 1 million consumer drones). Or, as the authors of the report put it: “Contrary to sensational media headlines, the skies are crowded not by drones, but by fowl.”

Yet the number of dangerous bird strikes is still relatively low considering that 27,000 commercial flights take to the skies daily in the U.S. Of 160,000 reported wildlife strikes since 1990, just 14,314 have resulted in damage. Of those, 80% were caused by what the FAA defines as medium- to large-sized animals. Only 3% of small bird strikes cause damage, and the authors argue that because many small bird strikes likely don’t get reported (because they don’t cause damage) that small number is likely somewhat inflated.

How does that bird data translate to human impacts? Since 1990, the FAA records 238 wildlife strikes in which there were injuries or fatalities. Where commercial aviation is concerned—that is, excluding private aircraft, business jets, government aircraft, etc.—the total number of incidents falls to 37 (again: there are 27,000 commercial flights transiting U.S. skies per day).

MORE: Capt. ‘Sully’ on drone rules: 'We have a responsibility to do this right'

Of 398 people who have been injured, 100 of them stem from a single incident—the so-called “Miracle on the Hudson” in which Capt. Chesley “Sully” Sullenberger successfully landed a U.S. Airways Airbus A320 in the Hudson River in New York City after multiple geese were sucked into both engines of the aircraft, saving all 155 people aboard. (Incidentally, Capt. Sullenberger has been vocal about the need for more research into the damaging effects drone impacts might have on commercial airliners, and how those impacts might differ from bird strikes.)

Of the 12 recorded wildlife strikes resulting in fatalities only one involved a commercial jet, and the incident occurred on the ground when an Embraer EMB-120 hit a pair of deer on the runway while landing.

The Mercatus report contains a lot of numbers like these designed to drive home the point that the likelihood of any given commercial jet colliding with anything at all—drone, bird, or even an unlucky land mammal—is statistically very low. But to hammer the point further still, authors Eli Dourado and Samuel Hammond dig into the data further still by calculating how often we should reasonably expect a drone to damage an aircraft.

For more about drones, watch:

If drones spent as much time in the air on average as birds, we might reasonably expect one drone collision with one aircraft per year. Given that drones spend far less time in the air than birds (most consumer drones have a battery life of less than half an hour), that likelihood drops sharply. The study finds that a damaging drone collision will occur no more than once every 1.87 million years of 4.4-pound drone flight time. A fatality or injury will occur once every 187 million years of drone operation.

“This appears to be an acceptable risk to the airspace,” they conclude, though they admit their methodology comes with some caveats, mainly that we really don’t know much about the effects of drone collisions on aircraft. As Capt. Sullenberger told Fortune last year, the effects of hard composites, heavy batteries, and small drone motors impacting an aircraft engine or cockpit windshield are likely significantly different from the effects of striking a bird of the same weight. But as of right now, we just have no data to go on.

There’s two ways to look at that dearth of data. On the one hand, more data is needed to define just how big of a threat a drone impact with an airplane would be before small drones are given greater latitude within the national airspace. On the other, with more than one million consumer drones in the sky, there hasn’t yet been a single reported collision with a commercial aircraft that we can use as a measuring stick.

Correction (Mar. 17, 10:30 a.m. ET): An earlier version of this story incorrectly identified the location of the Mercatus Center. It is located at George Mason University, not George Washington University. The copy has been updated accordingly.

All products and services featured are based solely on editorial selection. FORTUNE may receive compensation for some links to products and services on this website.

Quotes delayed at least 15 minutes. Market data provided by Interactive Data. ETF and Mutual Fund data provided by Morningstar, Inc. Dow Jones Terms & Conditions: S&P Index data is the property of Chicago Mercantile Exchange Inc. and its licensors. All rights reserved. Terms & Conditions. Powered and implemented by Interactive Data Managed Solutions