Coyote Blog

In the aftermath of the two Boeing 737MAX crashes:

For years, the FAA has allowed plane manufacturers to self-certify parts of the oversight process for new planes, called Organization Designation Authorization. This process, in which the aircraft manufacturer’s employees perform some of the safety tests and inspections with FAA oversight, reportedly saved the government body time and money.

That practice was examined at Wednesday’s Senate hearing.

Department of Transportation Inspector General Calvin Scovel III, who testified at the hearing, said the FAA will significantly change the oversight process for new aircraft by July. Speaking in vague terms, Scovel said that the changes would include new ways for the FAA to evaluate the self-certifying process.

Sen. Richard Blumenthal said that putting manufacturers in charge of their own safety audits was like putting “the fox in charge of the henhouse.” Saying he would introduce regulations to ban the practice of companies self-certifying, Blumenthal stated that “the fact is that the FAA decided to do safety on the cheap, which is neither safe nor cheap.”

A few reactions:

1. The fox in the henhouse analogy is not apt.  The fox wants to eat the chickens, whereas Boeing does not want to have airplane failures.  In fact Boeing is going to be paying out on a bunch of really big lawsuits, not only to families of the folks that died and the airlines that lost their planes but also to airlines that have had to change their flight schedules due to these issues.  Airbus sales people will use this story in their pitches until the end of time.  Regulation is not the only, or the most important, check on Boeing's behaviors.
2. That being said, aircraft regulation is a dumb hill for libertarians to die on.  This is just not that big of a deal.  Regulation and capital intensity has pretty much reduced choice in large aircraft to two companies and that will not likely change no matter what extra regulatory hoops are added.  Aircraft are a bit more expensive and spare parts are way more expensive due to our regulatory regime, but I don't think there is a public constituency for making a different trade-off.
3. Whatever the regulatory environment, it is unlikely to actually catch more failures of this sort in the future.  Regulators are notoriously bad at this sort of thing (see: US financial system).
4. I did engineering failure analysis early in my working career and my experience is that this sort of multiple stacked failure -- lack of pilot training for a bad software response based on a failed piece of instrumentation that was not reported as needing maintenance -- is hard to predict.  What will happen now in addition to some software fixes will be more mandatory training on this particular subsystem and likely a requirement that the specific piece of instrumentation involved needs to have redundancy.  At best we should hope they will also do a review of other instrumentation failures that might lead to a flight control issue and consider redundancy or software changes.  But there's always the problem of failure of imagination, the best dramatization of which is in the fabulous From the Earth to the Moon episode on Apollo 1.