Four consecutive deferrals have kept the US airline wheelchair liability rule unenforced since it took effect, and no passenger can sue

What is happening?

The United States Department of Transportation issued a final rule on 2024-12-17 that makes an airline liable when it fails to return a wheelchair on time in the condition it was received. The rule took effect on 2025-01-16. That liability provision, 14 CFR 382.130(a), has not been enforced on any day since, because the Department has issued four consecutive enforcement-discretion notices, the most recent on 2026-08-04, which carries the pause to 2027-04-30.

The four notices are dated 2025-02-20, 2025-06-10, 2025-09-30 and 2026-08-04. Each was issued as the previous pause ran out, and each names the provisions it suspends in operative regulatory text. Alongside 382.130(a) the notices pause the pre-departure notice of the right to contact a complaint resolution official, the reimbursement of a fare difference when a device will not fit, and the annual frequency element of two training requirements.

The reasons given in the later notices are a replacement rulemaking, review by officials appointed or designated by the President, and the issues raised by a lawsuit that six carriers and their trade association brought against the rule in February 2025. The replacement rulemaking has not been proposed. It carries the regulation identifier 2105-AF35 and a target month of August 2026 that has appeared unchanged across two editions of the Unified Agenda with a blank citation field. A query of the Federal Register on 2026-08-08 returns zero documents under that identifier.

What the passenger cannot do is go around the Department. Five federal appellate circuits have held that the Air Carrier Access Act of 1986 creates a right without a private remedy, so the Secretary of Transportation is the only enforcer the statute names. In calendar year 2025 airlines reporting to the Bureau of Transportation Statistics enplaned 907,259 wheelchairs and scooters and mishandled 1.09 percent of them.

Whose problem is this?

The body that decides whether to enforce is the same body that wrote the rule and the same body that has now paused it four times. Nothing outside that body can start the enforcement, and nobody inside it has to explain a decision not to.

Where does this problem end?

The boundary here is unusual because nothing is missing at the level of law. A prohibition exists, a rule exists, an enforcer exists and a penalty schedule exists. What is missing is anyone besides the enforcer who can act on any of it.

What is the state now, and what should it be?

Now

Needs a new measurementthe target state: no source opened here names a goal mishandling rate, a goal complaint-resolution time, a goal number of enforcement actions, or a date on which the deferred provision becomes enforceable and stays that way. The 2024 rule describes the intended floor, which is that a device comes back on time in the condition it was received and that the airline pays when it does not, and every source opened is silent on when that floor starts to bind.

How big is it?

Between 4,900 and 9,900 people a year. Airlines reporting to the Bureau of Transportation Statistics enplaned 907,259 wheelchairs and scooters in calendar year 2025 and mishandled 1.09 percent of them, which is about 9,890 mishandling events. That product is arithmetic performed here and no source opened prints it.

Events are not people. An enplanement counts a device once per flight leg, so a round trip counts the same device twice, and one person can be mishandled more than once in a year. Deflating fully for round trips gives the low bound of about 4,900 people, and treating every event as a distinct person gives the high bound of about 9,900. Nothing published narrows the band, because no source reports unique passengers rather than devices. The count also leaves out foreign carriers, which the statute reaches and the reporting does not. It leaves out small US carriers below the revenue reporting threshold. It leaves out every violation that is not a mishandled device, and it leaves out the people who stopped flying, a group the final rule names in words and no source counts in numbers.

The wider stock is not-derivable from what was opened. About 5.5 million people in the United States use a wheelchair, but that is a stock and this problem is a flow of travel events. Bridging the two needs the number of unique passengers who fly with a device in a year, and no agency publishes it.

Under what conditions does it arise?

1. The right was created without a remedy for the right-holder. The statute of 1986 prohibits discrimination and then names one enforcer. Under the framework the Supreme Court set in Alexander v. Sandoval in 2001, a statute that builds an administrative scheme without saying that individuals may sue gets no implied right to sue, and the more complete the scheme looks the stronger that inference becomes. The investigation and reporting duties Congress added in 2000 to strengthen enforcement are read by the Fifth Circuit in 2018 as evidence that the administrative route was meant to be the only one.

2. The sole enforcer sets its threshold above the unit in which the harm occurs. Proposed guidance published on 2026-01-06 states that the focus is compliance rather than penalty, that a warning letter comes first, and that enforcement action may be appropriate when there is evidence of widespread, systemic, egregious or intentional violations. One damaged wheelchair on one trip is by construction none of those things.

3. The deferral renews itself and the thing it waits for does not arrive. Four notices have issued since 2025-02-20, each as the previous pause expired. The replacement rulemaking that the later notices give as the reason carries the identifier 2105-AF35 and a target month of August 2026, and as of 2026-08-08 the Federal Register contains zero documents under it.

4. The claim that survives is priced at the object. 14 CFR 382.131 lifts the domestic baggage liability caps for wheelchairs and sets compensation on the original purchase price of the device. It is a claim made to the airline rather than to a court, and a purchase price does not reach lost mobility, missed work, substitute transport, or the weeks a custom chair takes to rebuild.

What has been tried?

Three directions have been tried and none of them meets the individual. Congress built the right and left the remedy to an agency. The agency wrote a rule that would have paid the individual and then suspended it four times. The courts closed the implied route and preserved only a narrow state-law path that reaches injury from interaction rather than a destroyed device.

What was found?

Why is it still unsolved?

Enforcement-absent — the apparatus is fully built, described by the agency in its own words, and then stood down provision by provision.

Start from what exists. In January 2026 the Office of Aviation Consumer Protection listed its own toolkit, and the list is not short. It names consent orders, formal proceedings before an administrative law judge under 49 USC 46301, civil actions in United States district court under 49 USC 46106 and 46107, and warning letters. Everything needed to enforce this statute is in that list. What the same document then says is that the focus will be compliance rather than penalty, that a warning letter comes before an action, and that the office generally acts on a pattern or practice. The apparatus is not missing. It is set to a sensitivity at which a single passenger cannot register.

The second part is that nobody else can register the passenger either, and this is where the missing private right of action does its work. Under every comparable civil rights statute a person the agency declines to help can go to court. Here five circuits have held that route closed, and because they agree there is no split, nothing for the Supreme Court to resolve and no escalation path. The state-law opening the Ninth Circuit preserved in 2013 is real but narrow, because it covers personal injury arising from how airline agents interact with a passenger rather than the loss of a device, and the same opinion holds that an airport terminal is not a place of public accommodation. What is left to the individual is a claim to the airline priced at the original purchase price of the device, administered by the airline and policed by the office that has just said it prefers warning letters.

The third part is what keeps the pattern stable, and it is that a deferral produces no event. A repeal makes news. A rule paused to 2025-03-20, then to 2025-08-01, then to 2026-12-31, then to 2027-04-30 looks at every moment like a rule that is about to be enforced. The reason now given is a replacement rulemaking whose target month is the month underway and whose file is empty. The reason given before that was a lawsuit whose docket could not be confirmed from any accessible source. No document opened here names a date on which the pause fails, and nobody will have to justify a fifth notice, because the fourth already stands as precedent for it.

What observation would mean it is solved?

Candidates — (a) the Department lets the deferral lapse and enforces 14 CFR 382.130(a) against a carrier (b) Congress enacts a private right of action along the lines of H.R. 9373 (c) the published mishandling rate falls and keeps falling.

(a) alone is weaker than it looks. Enforcement can resume and still run at the stated threshold, which is a pattern or practice. One enforcement action for accessibility non-compliance since 2019 is what that threshold has produced so far, against complaint volumes in the four figures. A resumed provision that is invoked once every several years restores the rule on paper and changes little for the passenger whose device was destroyed last month.

(b) alone is weak for a different reason. A right to sue is not the same as a suit worth bringing. The recoverable amount for one device is small against the cost of federal litigation, which is why practitioners report that attorneys rarely take these cases on contingency. H.R. 9373 answers part of that with mandatory fees and punitive damages, but a statute can be enacted and still leave the passenger to find counsel, and a bill referred to committee on 2026-06-18 with no further action found is not yet law.

(c) alone counts the wrong thing. The rate fell from 1.38 percent in 2023 to 1.26 percent in 2024 to 1.09 percent in 2025 while the liability provision was unenforced, so the decline is evidence for neither side of the argument about the rule. It also measures devices rather than people, excludes foreign carriers and small carriers, and says nothing about what reaches the passenger whose device falls in the remaining 1.09 percent. The three have to be read together, and (a) has to be read against how many actions follow it.

What is it connected to?

Fills with researchliability limits on international itineraries under the Montreal Convention, the boundary of the Americans with Disabilities Act at the terminal door, the same enforcement-only design in other transport modes, and comparable air travel disability regimes outside the United States. Relation type and evidence grade were not confirmed in this round.

What these sources do not say

• What passengers actually recover under 14 CFR 382.131. No aggregate paid, no average, no denial rate, no dispute outcome, no processing time. The one remedy that faces the individual has zero published performance data, and it is the remedy every other source points to as the alternative to a lawsuit.
• Whether the Department meets its own 120-day deadline. Congress wrote that deadline into the statute in 2024. Two years later no source opened here gives a compliance rate, a median or a backlog figure, so the performance of the sole enforcement channel is unmeasured in public.
• Complaint volume after August 2022. The statute requires an annual report to Congress on disability complaints. No edition could be opened from any accessible host and no source names its most recent date, so counts for 2023, 2024 and 2025 are absent from everything read here.
• What the deferral costs. None of the four notices estimates mishandling events during the pause, counts passengers affected, or names forgone penalties. The 2026-08-04 notice quantifies exactly one thing, which is a carrier request describing almost 30,000 flight attendants removed from operations for one additional day each year. The burden on the regulated party is numbered and the burden on the protected class is not.
• A gap of about sixty days in the record. The second notice suspended enforcement until 2025-08-01 and the third issued on 2025-09-30. Nothing opened here says whether the provision was live in between, or whether anything was enforced if it was.
• What happens if August 2026 passes with no proposal. The agenda has carried that target across two editions with a blank citation. No source names a consequence, or says whether the pause lapses, extends a fifth time, or becomes indefinite.
• How many claims have been dismissed for want of a private right of action. The five appellate holdings are known and the district court volume beneath them is counted nowhere found here, so the number of people who went to a courthouse and were turned away is unknown even in order of magnitude.
• Why a rule with negative quantified net benefits was finalized and then not enforced. The regulatory impact analysis prints negative USD 89.5 million over twenty years, carried by unquantified impacts named as avoided loss of mobility, improved passenger dignity, and avoided fatalities and injuries. Nothing opened here mentions that figure again in either direction.

See the evidence

Six primary federal documents were read end to end. Four Federal Register documents were obtained as Government Publishing Office PDFs and text-extracted locally, because federalregister.gov and ecfr.gov HTML pages redirect to a challenge host. The bill text and the Ninth Circuit slip opinion were read directly, as was the GAO testimony. The Federal Register API was reachable and supplied the docket chronology and the two negative findings about the replacement rulemaking, and a negative result there is weaker than a positive one, since it shows only that the API index contains nothing under that identifier. Everything about mishandling volume is secondary. The primary consumer report could not be opened from any Departmental host, so the 907,259 and 1.09 percent figures rest on outlets citing it. Where sources disagree the disagreement is left visible rather than resolved. The same GAO document gives 1,394 disability complaints for 2021 in its text and 1,395 in a footnote, and gives 37 disability-related consent orders for 2008 through 2019 in its text against 33 of 380 in a footnote citing an earlier GAO report. The 2024 mishandling count of 11,357 does not reconcile exactly with a rate of 1.26 percent applied to nearly 900,000 devices, which implies about 11,340, and neither source shows its base. One practitioner source dates the Fifth Circuit decision to 2017 while the operative opinion is dated 2018-04-05, which the Ninth Circuit account of a reconsidered position may explain, though no source opened confirms that reading. Two things are asserted here without a verifiable document and are flagged as such rather than dropped: the case number of the carrier petition, which appears only in search listings, and the current status of that petition. This is a Path A output, so observation_refs is empty and provenance_mode is press-derived.

People affected

Estimated range 4,900–9,900 As of 2025 calendar year

Sensitivity The width of the band is entirely the events-to-persons conversion, because both inputs to the event count are single published figures with no margin attached. The high bound of about 9,900 treats every mishandling event as a distinct person. The low bound of about 4,900 assumes full round-trip deflation, that is, that every affected person is hit on both legs of one trip. The truth sits between them and nothing published narrows it, because no agency reports unique passengers who fly with a mobility device. What the number fails to count runs in one direction only, and it is large. It excludes foreign carriers, which 49 USC 41705 explicitly reaches and which do not report to this series. It excludes small US carriers below the revenue reporting threshold, a set the regulatory flexibility analysis of the 2024 rule puts at 29 carriers reporting traffic in 2024. It excludes every violation of the Act that is not a mishandled device, including denied boarding, failure to provide assistance, seating accommodation and injury during a transfer, which are precisely the categories the pending bill would attach penalties to. It excludes devices carried in the cabin and mishandlings never reported. And it excludes people who stopped flying, a group the final rule names in words and no source counts. The limit in the opposite direction is that this is a flow of annual travel events and not a stock of people at risk. About 5.5 million people in the United States use a wheelchair, but that figure cannot be joined to this one without the number of unique passengers who fly with a device in a year, which no agency publishes. Complaint counts are also a floor rather than a census, and the last one available from any source opened here is January through August 2022.

Regional breakdown The reported series is national and organized by carrier rather than by state or region. No source opened here gives device enplanements or mishandling counts by origin, destination or airport. Splitting the national figure by state population would be proportional allocation and would be wrong twice over, because wheelchair air travel concentrates at hub airports and at destinations with accessible tourism rather than distributing with residence, and because the enforcement gap being measured is federal and uniform across every state.