Feds told Tesla to stop making “misleading statements” on Model 3 safety

"Your company has issued a number of misleading statements," NHTSA wrote to Musk.

Read the whole story

 

[Tondelaya Shillyshally von Brainstorm]

The clear issue here is that Musk is attempting to use an official source as a citation while arguing that the official source is wrong.

He could have said, according to an independent analysis of NHTSA data, the Model 3 is the safest car in its class, and probably been okay.

Instead, he wanted to use the reputation of the NHTSA as a reliable agency, so he said it was according to the NHTSA, and then, he compared to all cars, not just cars in the same class.

When the agency told him he was using the data in an unsupported manner, he said they were wrong.

If they are wrong, why refer to their numbers at all? Are they right on everything else, and just wrong on how to read their own tests? That would be odd, would you agree?

Musk is just showing he is really not fit to be in the position he has.

 

[Tondelaya Shillyshally von Brainstorm]

The clear issue here is that Musk is attempting to use an official source as a citation while arguing that the official source is wrong.

He could have said, according to an independent analysis of NHTSA data, the Model 3 is the safest car in its class, and probably been okay.

Instead, he wanted to use the reputation of the NHTSA as a reliable agency, so he said it was according to the NHTSA, and then, he compared to all cars, not just cars in the same class.

When the agency told him he was using the data in an unsupported manner, he said they were wrong.

If they are wrong, why refer to their numbers at all? Are they right on everything else, and just wrong on how to read their own tests? That would be odd, would you agree?

Musk is just showing he is really not fit to be in the position he has.

I'm doubtful it would have been OK. If the highest rating is 5 stars and if there is more than one 5 star rated vehicle, then they can't claim their car is 'safer'. By the criteria used in the rating they can only claim their vehicle is 'as safe' as the others.

I believe the problematic term in the statement was 'achieves.'

Model 3 achieves the lowest probability of injury of any vehicle ever tested by NHTSA

The word 'achieves' indicates that the NHTSA itself acknowledged the injury probability. The entire structure of the headline is somewhat problematic, as it does not specify the source of the claim as someone other than the NHTSA. However, I believe if we replace the word 'achieves' with the two words 'data indicates' the entire thing would have just gone by barely noticed.

You can independently analyze government data improperly and then tell people the result of your improper analysis, and as far as I am aware, that is covered by freedom of expression. You can release bad science and stamp it with the Tesla symbol and you did not break a law. The problem comes when you start pretending the source of your improper analysis is some other entity, such as a government agency.

If you said, according to an analysis done by my seventh grader during recess, using NHTSA data, the Model 3 would win in a crash with a Cadillac Escalade, you would be okay. You labeled the source. You did not imply that the NHTSA endorsed the claim.

The content of his claim was fine, but the wording was inappropriate.

 

[Tondelaya Shillyshally von Brainstorm]

Tesla noticed that the Model 3 had a better VSS score than any other vehicle on the market. That, in Tesla's view, means that a Model 3 driver is less likely to be injured in a crash than a driver of any other vehicle.

So would I. If NHTSA thinks the VSS isn't sufficient for such comparisons then the solution isn't to hammer Tesla's use. NHTSA needs to fix VSS then publish the scores.

This sort of thinking led to the oft-mocked 100-point video game review scores and the like.of yesteryear. As if the subjective reviewer could really draw a meaningful distinction between, say, a 93-point score and a 92-point score.

The position of the NHTSA, which knows a lot more than you or me about crashworthiness and auto safety, is that crash dynamics are so complex that there is no utility in offering more than a five-level (six? I don't know if there's a zero-star rating) scoring system. More verbose information is not necessarily better, especially when attempting to communicate to non-experts.

"People are too stupid to understand more data" is a pretty poor excuse for not having a more informative safety scoring system than 1-5.

It is already pretty useless in a lot of ways anyway, but I am unsure a more descriptive scoring system would really help.

People often forget that humans differ from each other. Safe for you is not guaranteed to be safe for me. If Jamie is 400 lb 5'6" and Sandy is 150 lb 6'4" they will have different safest vehicles. They certainly have different vehicles they simply cannot safely operate.

Aside from the fact that obese people are eighty percent more likely to die in an accident and more likely to get into an accident at all, there are certain cars which will not safely seat a 400 lb person. Seat belt extensions reduce safety on three point systems and seat rail modification can change the occupant position from where the crash test devices were placed during testing. On the other hand, some convertibles have such low tops that in the event of a rollover, Sandy will have his head crushed instantly. They may be perfectly safe for someone else, but not for Sandy.

The idea of one safest car is actually a bit silly, which is why they stick with a vague system like this. Five stars means it is good enough for a safety conscious consumer to look at up close, not that it is a guarantee the human who decides to buy it will be safe in it. Some consideration needs to be given to the driver, the environment, the amount of maintenance funding available, and the activity expected of the car on a regular basis. The rating only counts for a small bit, so there is no point in making it super specific.

 

[Tondelaya Shillyshally von Brainstorm]

I have a couple of questions for the group:

1) Does violating 'Guidelines' in this case actually mean anything in a real and legal manner? I note that in the letter from the NHTSA they are forwarding their complaint to the FTC. Does the NHTSA have any legal standing itself?

2) The FTC received this ~10 months ago. I certainly understand government departments take a while to respond, but isn't this a little long to go without some response? Or is this one of those cases of a department being able to hold something over a company. "You know, we'e still reviewing that complaint we received."


{ And not to inflame the fire going on, but I'd love to see the upvote/downvote reasoning breakdown. I've seen some (what I consider) interesting posts downvoted to oblivion and some remarkably useless ones left completely unvoted. Maybe it's a case where coming back in 12 hours will generate a clearer picture. }

1. Their job is to create not to enforce. They do not have any real power over anybody. They can easily refer your case to whoever has relevant power, in this case the Federal Trade Commission, but also the Department of Transportation, the Federal Aviation Administration and the Federal Bureau of Investigation. There is no purpose to employing an enforcement body inside of the NHTSA.

2. Ten months is not super long for this kind of thing. The FTC takes time. If a punishment comes out of this it will be retroactive to the date of the issue. If it takes them two years and they come back and determine that Tesla must pay a fine to coincide with a percentage of earnings, the earnings number they use will be from now, not from two years from now. If Tesla is circling the drain in 2021, but in 2019 they were worth $60 billion, the FTC will calculate based on $60 billion and force Tesla to go bankrupt paying their fine.

 

[Tondelaya Shillyshally von Brainstorm]

Let's level set here. Tesla's Model 3 performed better on NHTSA's "Vehicle Safety Score" than any other car ever. NHTSA does not dispute that fact.

NHTSA disputes comparison claims using VSS when those claims can be interpreted to suggest that a better VSS for Vehicle A than Vehicle B supports the conclusion that Vehicle A will be safer than Vehicle B in all scenarios. Tesla's claims don't clarify that there are specific circumstances where comparison using VSS won't be valid.

The specific circumstance NHSTA cites as an example where VSS comparisons are dubious is where Vehicle A has a crash with a significantly heavier vehicle, Vehicle B.

Tesla counters with two solid points in their response to NHTSA:

(1) As a matter of law, Tesla's claims only make the generic statement about overall likelihood of injury - they make no claims (implied or otherwise) about performance in specific crash scenarios (with a heavier vehicle, rolling over, setting on fire, etc.), AND

(2) the basis for taking the overall VSS score and applying it to real-life crash scenarios OVERALL, is that most crashes involving death or serious injury are not the weight-mismatch scenario that NHTSA highlighted.

Something around half of serious injuries/deaths occur in crashes that are single-vehicle , and, among multi-vehicle crashes, a significant percentage will involve vehicles that are going to be of similar or lesser weight than the Model 3 (which is about average in weight among light duty cars and trucks).

I appreciate NHTSA wanting to police the use of their data for representations made to consumers. And I also think it's appropriate for a consumer protection agency to confirm the validity of the evidence Tesla uses to support, in particular, Point 2 above (that their VSS comparison is valid for the majority of crash-types that result in serious injury or death).

But this is not the same, for example, as Musk's statement about having secured the funding for Tesla to be bought out or making irresponsible claims about when FSD will be truly available. Teslas really are exceptionally safe vehicles, and at least on certain tests, have outperformed all others, and - frankly - I think leaders in safety should be able to take credit for having done so.

To be clear, Tesla's claims may not specify circumstances where VSS may be invalid, but an absence of such a clarification tends to make the reader think there are no exceptions to the claim, not that the claim is valid for specific circumstances.

Honestly, I would think that actual empiric data - namely IIHS type data - would be a better talking point than NHTSA testing.

Heck, rollover scores are done as a static tilt test. Low-slung sports cars do well on this test, but empiric data clearly demonstrates that sports cars frequently roll over due to driver behavior.

Fortunately, businesses are not required to enumerate all specific instances where the general rule (i.e. so-and-so product is the safest) does not hold, so long as the general rule is true and supported by evidence and there are no clearly implied specific applications. Else, we would have a very difficult time communicating generally applicable rules.

It would be different if, in the context of any claim of being generally safer than other vehicles, Tesla implied that Model 3s would fair better than a heavier vehicle in a head on collision between the two (for example, if a commercial showed a Model 3 about to get into a collision with a bigger car/SUV). But that didn't happen here.

And I certainly agree that there are likely other stats/tests that car companies may want to rely on other than NHTSA's VSS.

Except the claim was

"Model 3 achieves the lowest probability of injury of any vehicle tested by NHTSA", followed by

"But when a crash happens in real life , these test results show that if you are driving a Tesla, you have the best chance of avoiding serious injury."

(Emphasis mine - note, it does not specify "under these test circumstances")

Now, the lawyerese of the Tesla response said they did not "claim Model 3 would outperform substantially heavier vehicles in a head to head crash...", it does leave open how any individual reading the previous quotes would come to the conclusion of a Tesla driver being less safe than a heavier vehicle in those circumstances.

Because presumably, any given reader would assume the heavier vehicle was tested by NHTSA as well, and that vehicle had a lower score - and therefore is less safe.

Except that Tesla's are actually quite heavy cars and unusually well built, so they will do very well against other cars, and there simply aren't many cars on the road significantly heavier.

That actually is accurate, which I did not expect when I started to look into the claim you stated.

Model 3 weight is 3600-4200lbs

Audi A8 weight is 4300lbs

BMW 5 series weight is 3700-4400lbs

BMW 7 series weight is 4200-4800lbs

Ford Taurus weight is 4000-4400lbs

Hyundai Equas weight is 4600lbs

Toyota Camry weight is 3200-3600lbs

Volvo V90 weigght is 4200lbs

There are certainly some that are much heavier, but they are not likely owned by very many.

Rolls Royce Phantom - 5600lbs

Bentley Mulsane - 6000lbs

Actually, many Rolls Royce and Bentley models are over 5000, as well as some Jaguars and a few others, so if you are in a Model 3, try to avoid hitting the fancy cars even if they look the same size.