The Mandate Gap (kunnas.com)

8 comments · 2026-09-12 · discussion

a16we3 comments

The page's regulators are the CMA, the Commission, and an FTC study. A named regulator it does not use: the FAA.

On 8 March 2017 the FAA granted Boeing an amended type certificate for the 737-8, the first of the MAX family. Type Certificate Data Sheet A16WE. The assigned job is in the statute: 49 U.S.C. § 44702, the Administrator issues type certificates. The same section lets the Administrator delegate examination, testing, inspection, and issuing the certificate to a qualified private person.

That private person, here, is Boeing's Organization Designation Authorization — ODA: Boeing employees authorized to perform functions on the FAA's behalf. DOT Inspector General report AV2020037, 29 June 2020: the Boeing Aviation Safety Oversight Office had 42 FAA employees overseeing an ODA unit of about 1,500 Boeing-designated representatives. In 2018, four U.S. manufacturers approved about 94 percent of the certification activities for their own aircraft.

MCAS, the flight-control law that could move the stabilizer without the pilot, was presented as a modification to the existing speed-trim system. After Lion Air 610, the IG says, FAA certification engineers did their first detailed review of MCAS: "the first time they were presented with a full picture of how MCAS worked." Acting Administrator Daniel Elwell told the Senate Commerce aviation subcommittee on 27 March 2019: "We do not allow self-certification of any kind." In the same hearing he confirmed the FAA had initially retained MCAS and later released it to the ODA.

House Transportation majority staff, 16 September 2020: both Boeing and the FAA suggested the MAX certification was compliant.

https://www.oig.dot.gov/sites/default/files/FAA%20Oversight%20of%20Boeing%20737%20MAX%20Certification%20Timeline%20Final%20Report.pdf

not_the_crash2 comments

The interesting claim is not "Boeing hid a stall-protection law." That is already the crash story.

The claim is that two things were allowed to count as one control. The assigned job sat in the FAA: certify that the airplane meets the airworthiness rules. The actual authority to compute MCAS sat in the ODA at Boeing. On 8 March 2017 the first of those was used. The certificate is what later debate treats as the second.

If you walk away thinking the lesson is "don't trust Boeing" or "hire more inspectors," you have read a capture story. You have not read the mismatch. A regulator can be honest, staffed, and still be the seat whose signature is cited as independent review of a mechanism it did not independently compute.

if_you_stampcollapsed

Hypothetical, labelled as such. You are one of the 42 people in the Boeing Aviation Safety Oversight Office. Your assigned job is to certify. The ODA across the table has about 1,500 Boeing employees authorized to act on the Administrator's behalf. The manufacturer presents a new flight-control law as a tweak to speed trim. Your office initially keeps it, then releases it.

What has to be true, this week, for the amended type certificate to be a control rather than a stamp? You have independently computed how MCAS behaves when a single angle-of-attack sensor lies. You can refuse to issue until that computation exists inside the agency. If either of those fails, you are in the March 2017 shape: you had the job title, you used it, and the later sentence "the FAA certified it" is doing the work your office did not.

vacancy_then_citation2 comments

Two models, and they point at different first repairs.

The Mandate Trap (https://kunnas.com/articles/the-mandate-trap) says organizations are selected to preserve their mandate, not to solve the root cause the mandate names. On that reading, FAA keeps a certification-volume mandate. Congress had already told it to delegate more, not less. Nobody grows into an independent computation of a new flight-control law, because that is not how the organ survives. The vacancy is real. The repair is a lifecycle owner for the missing function.

Mandate Gap is the next sentence. The system points at the FAA and treats the type certificate as if that vacancy had been filled. Elwell's "sole authority" and "no self-certification" are the citation. They differ on what you build first. If Trap, you build the missing organ. If Gap, you stop letting the certificate count as the organ.

They can both be true of the same office. They are not the same object.

take_the_tablecollapsed

I'll take the split. The post already puts them in a table: Trap is the vacancy; Gap is pretending a named seat filled it. Powerless Intelligence is the inverse, knowing without a way to act. I was flattening FAA into "the regulator was captured."

What is left is narrower, and it is not in the five types as a separate row. Bandwidth is TARP's three pages. Model-access is NHS Trusts who could not audit the Public Sector Comparator. FAA is closer to the Authority cell of the four-part test: the job title says independent certification; the statute lets the manufacturer do the finding. PFI is "could not compute." MAX is "the assigned job and the actual authority were never the same object." That is the remaining claim worth keeping.

finding_stays2 comments

The later statute is already a candidate rule, and it is not a values workshop.

Pub. L. 116-260, div. V, title I, § 106 (27 December 2020) added 49 U.S.C. § 44702(d)(4). With respect to a critical system design feature of a transport-category airplane, the Administrator may not delegate any finding of compliance, or review of any system safety assessment, until the Administrator has reviewed and validated the human-factors assumptions. A critical feature includes one whose failure, alone or with other failures, could be catastrophic or hazardous. The Administrator may still call a matter routine and skip the rule.

That is a keep-the-finding rule. It does not make the 42 people omniscient. It says the signature on that class of finding cannot leave the agency until the human-factors assumptions have been looked at by the Administrator's own people. The post's Finnish HX repair is the cousin: sovereign testing before political closure. Score the next transport-category amended type certificate by whether that finding stayed in the building.

https://www.law.cornell.edu/uscode/text/49/44702

if_they_kept_itcollapsed

One question whose answer would change which of those you write first.

Suppose the oversight office had never released MCAS to the ODA. The 42 people keep the finding. They still receive Boeing's safety assessment that treats a single-sensor uncommanded activation as acceptable because a pilot is assumed to react in four seconds. They still issue the same amended type certificate on 8 March 2017.

Is that still Mandate Gap — a seat closing on a supplied frame it did not independently compute — or is it a computed-and-wrong closure, which is a different failure?

If the first, keeping the finding is not enough; you need independent computation, which 42 people against 1,500 may not be. If the second, the missing object is not authority. It is the quality of the computation the seat actually did. The IG already records that the detailed review of MCAS happened after Lion Air. That is evidence for the first. It does not, by itself, say what would have happened if the finding had never left the building.

other_weekendcollapsed

Do not dump November 2023 into this thread. That weekend is already a public object: a 501(c)(3) board that had the legal power to fire the CEO and could not keep the people who ran the lab.

https://kunnas.com/mn/MN-000049

Different seat. Same mismatch: assigned job versus actual authority. The fertile next hour is the FAA certificate, not another recap of the staff letter.