How Many 2026 Recalls Are for a Part That Can Fall Off Your Car? We Audited All 523

Key facts

• Dataset: NHTSA ODI Recalls flat file FLAT_RCL_POST_2010.txt, source file last-modified 2026-08-21 07:05 GMT, retrieved and processed the same day.
• Universe: 244,499 rows → 15,197 unique campaigns → 523 vehicle recalls filed in 2026, covering 29,799,606 potentially affected units; Part 573 reports received 2026-01-02 to 2026-08-14.
• Detachment slice: 80 campaigns keyword-matched, then hand-audited one by one — 56 confirmed (10.7% of 2026 vehicle recalls; 900,226 units, 3.02% of all 2026 recalled units), 5 borderline, 19 excluded.
• Split: 22 exterior / road-debris campaigns (594,059 units) vs 34 structural / running-gear campaigns (306,167 units).
• Match precision: 56/80 = 70% strict, 76% counting borderline cases. The keyword filter alone overstates the count by nearly a third.
• The finding that inverts: only 22 of 56 confirmed campaigns (39%) are passenger cars and light trucks — but they carry 89.2% of the affected units.
• Hero campaign: 26V528, 2021-2023 Dodge Durango rear spoiler, 74,578 units, Part 573 received 2026-08-13.

How many 2026 recalls are for a part that can detach?

About 1 in 10. Of the 523 vehicle recalls filed in 2026 in the file we processed, 56 describe a part that can physically separate from the vehicle — 10.7%, covering 900,226 potentially affected units.

We got there in two steps, and the second one is the one that matters. First, a regex over each campaign's defect, consequence and corrective-action text looking for separation language — detach, separate, come off, break off, fall off, dislodge, come loose, road hazard, road debris. That flagged 80 campaigns. Then we read all 80 Part 573 summaries individually and labelled each one, publishing the verdict and the one-line reason for every campaign in the CSV.

Nineteen were false positives and five were borderline, so the keyword match alone would have overstated the count by nearly a third — and it would have overstated the units far worse, because the single largest keyword hit is not a detachment recall at all. Honda's 1,048,613-unit campaign 26V366 matched on the phrase "the bottle cap to detach": it is about a tire repair kit sealant bottle whose cap can pop off and hit whoever is using it. Nothing leaves the car. That one campaign is larger than every confirmed detachment recall combined, and a keyword-only study would have led with it.

The other big exclusions follow the same principle — the part has to leave the vehicle:

• 26V218 (294,448 Hyundai/Genesis): seat-belt anchors "may detach," but inside the cabin. That is a restraint failure, not road debris.
• 26V268 (179,698 Ford Ranger/Bronco): a dislodged seat-frame pivot bolt. Same reasoning.
• 26V230 (47,928 Nissan Cube) and two GM roof-rail airbag campaigns: inflator parts that detach into the cabin.
• 26V024 (8 Ford Explorers): a windshield that can detach — but only in a crash. Crash-only separations are excluded throughout.

Because that judgment is ours and not NHTSA's, the honest way to cite this is "about 1 in 10 of 2026 vehicle recalls," not a precise figure. A second analyst working from the same 80 summaries would likely land within a few campaigns, not on the identical 56.

Which parts fall off — and how dangerous is it?

Not all detachments are equal, so we split the 56 into two buckets.

Exterior / road-debris (22 campaigns, 594,059 units): trim, spoilers, moonroof and window glass, roof-rail covers, fender flares, hardtop roof panels, mirrors, solar panels, steps and ladders. These endanger the vehicles behind you.

Structural / running-gear (34 campaigns, 306,167 units): coil springs, shock absorbers, driveshafts, control-arm ball joints, wheels, lug nuts, wheel studs, subframes, brake pedals and steering linkages. These threaten the driver's own control of the vehicle.

Note the inversion: the exterior bucket has fewer campaigns but nearly twice the units, because a handful of very large trim and glass recalls dominate it. The structural bucket is mostly small campaigns.

Top confirmed 2026 detachment recalls, by units

Units are the manufacturer's potentially affected population estimate, not confirmed defective vehicles — the Jeep coil-spring report itself estimates an actual defect rate near 0.5%.

The check that changed the story: most of these campaigns are not cars

The robustness check most likely to break a "1 in 10 recalls" claim is the composition of the NHTSA file itself. It counts RVs, travel trailers, school and transit buses, heavy trucks and motorcycles alongside passenger cars — so a study that quietly implies "your car" may actually be describing fifth-wheel trailers.

We ran it, and it does change the picture, in both directions at once.

Most confirmed detachment campaigns are not passenger vehicles — but almost nine in ten affected vehicles are. Counting campaigns and counting cars give opposite impressions of the same 56 recalls, and any single sentence about this data should say which one it means.

The reason is visible in the roster: six separate confirmed campaigns are the same defect — incorrectly tightened shock bolts, five of them explicitly naming Curt Touring Coil Suspensions (Alliance 26V337, Grand Design 26V354, K.Z. 26V322, Forest River 26V336, Ember 26V339) plus Brinkley's 26V334 — spread across six trailer brands and together covering 10,620 units. They inflate the campaign count without moving the vehicle count. Two of the confirmed campaigns are motorcycles (Honda's handlebar lock screw, 26V500 and 26V146, 63,319 units between them), which is why Honda appears high on the manufacturer list.

We are publishing the campaign-count version as the headline because the denominator — all 523 vehicle recalls — has the same mixed composition, so numerator and denominator are consistent. But the units figure is the one to use when the question is "how many vehicles."

By the numbers

Manufacturers with the most confirmed campaigns: Ford 8 (447,861 units), Chrysler/FCA 4 (167,606), Honda 3 (67,252), Grand Design RV 3 (2,840), General Motors 2 (27,493). Ford alone is roughly half the entire confirmed detachment population by units.

The recalls in the news this week — what the file actually says

A cluster of "a part can separate" headlines prompted this count. Two of the three are not what the headlines imply, which is why we went to the primary file.

The Durango spoiler is the cleanest fresh example: a distinct new campaign number, Part 573 received 2026-08-13, covering 2021-2023 model years. It repeats the failure mode of an earlier 2021 Durango spoiler recall, so treat the two as apparently related rather than a confirmed expansion — the flat file does not link campaigns to one another.

Method

Source file. NHTSA Office of Defects Investigation Recalls flat file, FLAT_RCL_POST_2010.zip → FLAT_RCL_POST_2010.txt, downloaded from static.nhtsa.gov on 2026-08-21; the server reported the file last-modified 2026-08-21 07:05:27 GMT. It is public and needs no key. Field positions follow NHTSA's published RCL.txt record layout.

• Universe. 244,499 tab-delimited rows. The file stores one row per campaign × model-year × component, so rows were collapsed to one record per campaign on the CAMPNO field, giving 15,197 unique campaigns. Restricting to recall type V (vehicle) and campaign numbers matching ^26V gives the study universe: 523 campaigns, 29,799,606 potentially affected units, with Part 573 report dates from 2026-01-02 to 2026-08-14.
• Keyword filter. Case-insensitive regex over the concatenated defect summary, consequence summary and corrective-action text: detach | separat | come off | comes off | came off | break off | breaks off | broke off | fall off | falls off | fell off | dislodg | come loose | comes loose | became loose | road hazard | road debris. Yields 80 campaigns / 2,553,328 units.
• Hand audit. All 80 were read individually and labelled confirmed (a part can separate from the vehicle in normal use), borderline, or excluded. Exclusion rules, applied consistently: a part that detaches but remains inside the cabin is out (seat-belt anchors, seat bolts, airbag inflator end caps); a separation that occurs only in a crash is out (windshield bonding, bus body-panel joints); and text that matches the regex without describing any separation at all is out. Every campaign's verdict and the reason for it is a column in the CSV, so any of them can be disputed individually.
• Categorisation. Confirmed campaigns were split into exterior / road-debris and structural / running-gear by what the loose part endangers — other road users, or the driver's own control.
• Vehicle class. Each confirmed campaign was also tagged car / truck / bus / RV / trailer / motorcycle / chassis for the composition check above.
• Units. Sum of NHTSA's POTAFF column, taken once per campaign.

Known limits.

• The count is a judgment, not a measurement. 56 rests on one analyst's reading of 80 summaries. Precision of the keyword stage is 70% strict (76% including borderline), and the classifier also has false negatives: text describing a functional consequence rather than a separation will not match. Cite "about 1 in 10."
• Snapshot-bound. The latest Part 573 date in this file is 2026-08-14, so recalls publicised between 15 and 21 August are not in these counts. A re-pull on a later date will not reproduce these exact figures.
• All vehicle types are included, and as shown above that materially changes the campaign-count reading. A passenger-car-only subset is 22 campaigns / 803,236 units.
• POTAFF is a manufacturer estimate of the potentially affected population, routinely larger than the number of actually defective vehicles.
• No completion data. This file carries no remedy or completion column, so nothing here says how many of these vehicles were repaired.

Reproduce it. The frozen CSV carries all 80 matched campaigns with the audit verdict, the reason, category, vehicle class, units, dates and the full NHTSA defect text: 2026-detachment-recalls-2026-08-21.csv.

What a VIN check can and can't tell you here

Every campaign above is an open recall you can screen by VIN before you buy — but be precise about what the screen proves. A VIN check surfaces recall presence and count, the same as NHTSA's free tool. It does not confirm the repair.

This matters most for the freshest campaigns, where a remedy may not exist yet for anyone — a VIN showing the recall is simply telling the truth about the car. To learn whether a previous owner already had the work done, look the VIN up on NHTSA or ask a franchise dealer. Here is what a "recall completed" claim actually means on a used car, and why the label can mislead.

How to check this by VIN

1. NHTSA first, free: enter the VIN at nhtsa.gov/recalls for open recalls and their remedy status, straight from the regulator.
2. Screen the car's history: a Zilocar VIN check shows open-recall presence and count alongside the records NHTSA's tool lacks — prior accident and airbag-deployment records, salvage and junk-auction records, odometer and rollback checks, ownership and sales-listing history, specs, safety ratings and a valuation.
3. Look at the part itself. Trim, spoilers, moonroof glass, fender flares and roof-rail covers are visible on a walkaround. Gaps, misalignment, fresh adhesive or mismatched fasteners around them are worth a question.

Related checks

• Used Ford Explorer roof-rail cover detachment recall — campaign 26V448, the largest confirmed detachment recall of 2026
• Used Subaru Forester moonroof detachment recall — campaign 26V346, 69,663 units
• Ford Bronco Sport and Maverick "do not drive" recall check — how a severe recall shows up on a VIN

Sources

• NHTSA ODI Recalls flat file (FLAT_RCL_POST_2010.zip) — the primary dataset behind every count in this piece
• NHTSA flat-file field layout (RCL.txt) — the tab-delimited record spec
• NHTSA recall lookup by VIN — the free regulator tool for open-recall and remedy status
• Frozen dataset: 2026-detachment-recalls-2026-08-21.csv — all 80 matched campaigns with audit verdict, reason, category, vehicle class, units and defect text

Last verified: 2026-08-21