No, millions of cars are not catching fire every year

[OVERKILL]

Car and Driver came across the AutoInsuranceEZ claim that produced this chart:

Which has been heavily referenced to promote that EV's catch fire far less frequently than ICE and Hybrid vehicles. The numbers to them just did not make any sense.

There are ~270 million registered vehicles in the US, most of them gas, so if 1,529.9 per hundred thousand (1.5%) gas vehicles were catching fire, that's 4.13 million vehicle fires every year.

AutoInsuranceEZ claims their data is from the NTSB. However, Car and Driver was informed by the NTSB that it does not collect this information. So, they reached out to the NHTSA to see if perhaps the wrong source was cited. Nope, they don't collect it in this manner either. But did note that only ~5% of fires are crash related. They get their fire data from the NTFIRS, which does not collect this data on a per powertrain basis.

According to NTFIRS, between 2013 and 2017, the average number of fires was 117,400 per year from a reference pool of 261 million vehicles (excluding commercial, motorcycles and busses). That's 44.98 fires per hundred thousand (0.045%).

Whoops!!!

Whole article can be found here:

No, Millions of Cars Are Not Catching Fire Every Year

A New York Times story about U.S. car fires references a study that gets the frequency wrong by a factor of at least 60.

www.caranddriver.com

 

[Pablo]

I really don't believe a lot of this stuff.

Especially when no one is clearly collecting data using a logical controlled and reviewable method - let alone a method without bias.

 

[fsdork]

The stats in the infographic appear to claim a given number of fires per 100k sold - presumably annual sales. That's a far smaller number than total vehicles currently registered. The numbers still look a bit out of whack at a glance, but not to such a severe degree.

 

[Imp4]

Simply a great big nothing burger.
An inaccurate internet 'news' article is hardly notable.

 

[OVERKILL]

The stats in the infographic appear to claim a given number of fires per 100k sold - presumably annual sales. That's a far smaller number than total vehicles currently registered. The numbers still look a bit out of whack at a glance, but not to such a severe degree.

But that still applies. if 1,529.9 vehicles per 100,000 sold are catching fire, that's 1.5% of new vehicles. New vehicle sales in the US are around 15 million/year, that means 225,000 just brand new vehicles are catching fire every year. But we know that TOTAL vehicle fires is only 117,400 vehicles per year, for ALL vehicles (of around 261 million registered). So there's absolutely no way that figure:

A) Makes sense
B) is right

 

[Barkleymut]

I work in the hood. I see plenty of vehicles every day that I am amazed have not yet caught fire.

While it would be nearly impossible, I would like to see a study on the rate of vehicles that are properly maintained that catch fire.

 

[OVERKILL]

I work in the hood. I see plenty of vehicles every day that I am amazed have not yet caught fire.

While it would be nearly impossible, I would like to see a study on the rate of vehicles that are properly maintained that catch fire.

And that's the thing. According to the NTFIRS, there are 117,400 vehicle fires a year. That includes everything from brand new ones to clapped-out 30-year old hoopties where somebody has used washer hose and screw clamps to connect a fuel line.

 

[fsdork]

But that still applies. if 1,529.9 vehicles per 100,000 sold are catching fire, that's 1.5% of new vehicles. New vehicle sales in the US are around 15 million/year, that means 225,000 just brand new vehicles are catching fire every year. But we know that TOTAL vehicle fires is only 117,400 vehicles per year, for ALL vehicles (of around 261 million registered). So there's absolutely no way that figure:

A) Makes sense
B) is right

Agreed. I was merely pointing out the difference in magnitude of the error. 225,000 is a far smaller number than ~4 million. Neither number is accurate, but one gets more clicks than the other.

 

[PimTac]

What is not mentioned here is the number of cars that are set on fire. This is a regular occurrence where I live and I suspect it happens everywhere. A car will be stolen and later found burning and abandoned.

These could be counted as vehicle fires by the fire dept.

 

[OVERKILL]

Agreed. I was merely pointing out the difference in magnitude of the error. 225,000 is a far smaller number than ~4 million. Neither number is accurate, but one gets more clicks than the other.

It's further confused by the fact that they use "Total fires".

Basically, it's just a dumpster fire, lol.

 

[OVERKILL]

What is not mentioned here is the number of cars that are set on fire. This is a regular occurrence where I live and I suspect it happens everywhere. A car will be stolen and later found burning and abandoned.

These could be counted as vehicle fires by the fire dept.

Yes, and those would be included in the NTFIRS figure.

 

[SubieRubyRoo]

And that's the thing. According to the NTFIRS, there are 117,400 vehicle fires a year. That includes everything from brand new ones to clapped-out 30-year old hoopties where somebody has used washer hose and screw clamps to connect a fuel line.

Hey hey hey, what’s wrong with washer hose & screw clamps? Some folks can’t afford new SRT Jeeps

There have been plenty of zip tie and duct-tape fixes I’ve seen on 8-second Chicagoland street racers that run indefinite periods. One story I couldn’t find the article for was back around 2002, a ‘Vette with a fogger system had a nitrous backfire during a race on Joe Orr road, and the racer was so obstinate about not letting people see his “secrets” that he literally pushed people with fire extinguishers away from the car and watched it burn to the ground. Wonder if that car was included in the totals!

 

[OVERKILL]

Hey hey hey, what’s wrong with washer hose & screw clamps? Some folks can’t afford new SRT Jeeps

There have been plenty of zip tie and duct-tape fixes I’ve seen on 8-second Chicagoland street racers that run indefinite periods. One story I couldn’t find the article for was back around 2002, a ‘Vette with a fogger system had a nitrous backfire during a race on Joe Orr road, and the racer was so obstinate about not letting people see his “secrets” that he literally pushed people with fire extinguishers away from the car and watched it burn to the ground. Wonder if that car was included in the totals!

I know of a Mustang where somebody used screw clamps and vacuum hose to fix a fuel line and the car burned to the ground, lol.

 

[SubieRubyRoo]

I know of a Mustang where somebody used screw clamps and vacuum hose to fix a fuel line and the car burned to the ground, lol.

Fire is a teacher who does not suffer mistakes well, that’s for sure…

 

[JeffKeryk]

Numbers without data are just numbers. Maybe they include house fires? Maybe someone fat fingered a keyboard key to jack up the numbers?

 

[dnewton3]

Junk science. It's everywhere these days. I question just about any claim I ever read or hear about, because it's highly likely it's uninformed or biased, or probably both.

 

[64bawagon]

I put out plenty of car fires in my day, theres a myriad of reasons they catch fire. The majority of the ones I saw had an electrical origin though.

 

[dishdude]

Came across this toasty Lexus a few months ago....at least he saved the bag of Big Berthas so nothing of value was lost.

 

[Pablo]

Came across this toasty Lexus a few months ago....at least he saved the bag of Big Berthas so nothing of value was lost.

Got his golf bags out!

 

[Warlord]

Turns out it’s trillions.