"Bell the Hybrid Act" -- Freedom of Info. Act

Bob, I'll post my two cents.

The decision to "bell the hybrid" is moving forward along two tracks.

The first is Congressional, via the legislation noted above. There, the Congress, based on no information whatsoever, would require DOT to implement some fix for the assumed problem of low-noise vehicles such as hybrids.

The second is within DOT itself. The lawyer at yesterday's meeting made it clear that DOT has the authority to move forward on this issue without a Congressional mandate. DOT is actively pursuing several issues, mainly focused on the "how" of fixing the assumed problem, but also on trying to get a handle on whether or not there appears to be any excess risk of reportable low-speed collisions.

To that end, my take on it is that they are doing the right thing. They are looking for reported low-speed collisions involving hybrids. Then they are examining those in a multivariate (regression) analysis, to see whether, after accounting for factors such as visibility, alcohol, and such, there appears to be any excess risk of collision that is purely associated with being a hybrid. (Because, as we all know, if the car goes fast enough, the ICE comes on).

The problem is that there are so few such collisions that 1) they cannot work in their sample data and have to look at 100% of reported accidents in roughly 13 states, and 2) even looking at 100% of accidents, they can't find enough low-speed hybrid-pedestrian collisions to make any valid determination of whether or not there is excess risk associated with hybrids. They are, however, looking at older data ending in 2005, which would tend to keep the number of hybrids in the data low.

After the meeting, I wrote an email suggesting that they put aside their regression analysis (which seeks to get at the reason for any excess risk that might occur), and just publish the number of such incidents that occurred in 2005 (in the states from which they have data). My guess is that the number of such incidents involving a blind pedestrian would be zero. Whereas they see that as "not enough information to analyze", I see that fact in and of itself as an extremely powerful piece of information. And, because the state datasets are not samples but are the universe of incidents, there's no "statistical" issue involved in reporting the number -- it's not a statistical estimate of the number of such events, it's a count, and has no associated statistical uncertainty about it.

As far as I can tell, I can't get my hands on those state databases, or I'd do the analysis myself.

At any rate, my conclusions are the following:
DOT said there were not enough cases of low-speed collisions between hybrids and pedestrians to meet their standards for determining whether or not hybrids pose excess risk, within their regression analysis approach.

DOT didn't say that what I did (using a different database, and just calculating averages) was not legitimate. As far as I can tell, they calculated the same numbers I did and gave them to the National Federation for the Blind, and gave my report to the National Federation for the Blind. As I asserted before, there's no reason not to use the statistics I generated. They don't address the core issue of low-speed collisions, but they are helpful in framing any rational debate by showing that serious injury from vehicle-blind pedestrian impacts is rare.

It would be nice if DOT would show how few reported low-speed collisions there are between pedestrians and hybrids, but I doubt that they will go to the trouble.

There is still the potential for the Congress to force this issue by passing legislation.

My guess is that DOT is on track to issue some type of regulation in roughly two years. My experience is that if somebody in the bureaucracy has make his or her mind up that this is a good idea, then absent any pushback from a powerful entity with some standing on this issue (e.g., from the car makers), it'll proceed. I temper that with the sense the DOT/NHTSA is more fact oriented than most, and they'll probably require that there be some evidence of excess risk and some reliable way to eliminate that risk before proceeding. For example, they said that they would have to do a cost-benefit analysis before proceeding with any regulation.

My final upshot is this: on the Federal level, the pending legislation is a wild card that could be played at any time. It's a feel-good bill with lots of sponsors that would cost essentially no tax dollars. If the Congress picks up this issue, it would be worth trying to get the ear of somebody would pay attention to the facts. (For example, my back of the envelope tells me that, based on the observable data and on some plausible upper bound on excess mortality risk from hybrid-bline pedestrian collisions, even if the "bell" costs just $10 per vehicle, it saves so few lives that it would cost about $100 million per life saved. That's an unreasonably high cost-benefit tradeoff by traditional standards.)

However, DOT will proceed with this apace. My guess is that it they issue regulations, it'll take about two years to do that.

States may do whatever they want to do. So the state situation is essentially 50 wildcards. If you have an issue percolating in your state, you might want to use some of my data to get across the idea that hybrid-blind pedestrian collisions are not going to be a source of significant excess deaths or hospitalizations any time on the near future.

 

Bob,

I think that you are doing a terrific -- and I suspect very time consuming
-- job on behalf of all present and future hybrid, PHEV, EV, etc vehicle
owners. Thank you for your efforts on all our behalfs.

Esssentially the "need" for this legislation will be based on which ever side
presents the most pressing case. It will be statistically based, and "There
are three kinds of lies; little white lies, damned lies, and statistics." It is
also necessary to demonstrate both a history, and a "benefit" that
outweighs the costs.

I know that for emphasis you are catagorizing Prius/pedestrian injuries as
"bodies." But I hope that in searcjhing the data bases the approach is
more subtle.

Personally, I don't think it has been rigorously established that at speeds
typical in residential areas the Prius is significantly more quiet than any or
most well maintained cars. And from what you and NHTSA are saying,
there either aren't enough statistics, or the casualty data is unreliable or
all mucked up by other relevant issues at the time and place of the
casualty.

Anecdotally, since this foolishness has begun I have been listening to cars
that pass my house in a 25 MPH zone. In reality, at low speeds my and
the other Prius in my neighborhood make make just as much noise from the
tires/road interface as most other vehicles that are rolling along. From this
not-quite scientific survey, if the Prius is "dangerously quite" below 25
MPH, then so too are most cars. In turn, all cars should be required
to have whatever the warning device will be, which should be
automatically activated at less that 25 MPH.

Now isn't that just the most ridiculous thing that you've ever heard?

I keep wondering why it is that the contemplated "fix" does not consider
that the special and very small class of persons allegedly put at risk are not
required to wear some high visibility clothing or visible warning device --
probably a strobe light. In the long run, those persons would be safer,
no matter what the propulsion of the vehicle(s) they might encounter.

 

I wonder what the Japanese engineers at Toyota think about this. They have been looking for ways to reduce the amount of noise created by the Prius and other Toyotas for years. They must think we (in the US) are completely insane.

 

After carefully reading Dr. Hogan's post, I have to conclude we are in the
same camp.

Some comment on these points:

Whether a push to "bell the hybrid" comes from a congressional mandate or
internally in a federal agency, there is still a need for the agency to
rigorosly follow the "regulatory process" which involves the opportunuty to
make comments, perhaps public hearings, cost-benefit analyses, and access
to any statistical information or data bases used for analysis/support -- this
latter can be quite costly.

Once proposed regulations have been drawn up, they have to be published
and there is another period for comment and rebuttal.

The regulatory process was designed to be immune to hyperbole, press
hype, and emotionalism, and the Congressional Budget Office and other
internal and public watchdogs are constantly vigilant to assure that the
process isn't abused. Sometimes, proposed regulations are withdrawn
because sufficient need or benefit cannot be demonstrated by the agency.

I would guess that absent a body of well documented case histories and/or
iron-clad statistical analysis, there is no pressing need for action. I suspect
that proposed regs are at least two years off -- whatever the proposed fix
is, it will have to be very artfully crafted not to awaken the sleeping dogs of
the U.S. auto makers who are now clambering all over each other in
announcing their PHEV, EV, whatever, to arrive at a dealership near you in
the next two to three years.

Of course, there's always the possibility of a last minute, going-out-the-door
Executive Order... !!??