PulsCar analysis of NHTSA public data, September 2026 Timing belts: 853 owner complaints, 39 manufacturer bulletins, 2 recall campaigns Licensed CC BY 4.0. Queries run 18 September 2026. -------------------------------------------------------------------------- SOURCE FILES -------------------------------------------------------------------------- NHTSA ODI complaint flat files COMPLAINTS_RECEIVED_2015-2019, _2020-2024, _2025-2026 -- https://static.nhtsa.gov/odi/ffdd/cmpl/ (dictionary CMPL.txt) 777,040 distinct vehicle complaints (ODINO, product type V) received 1 January 2015 to 8 September 2026. NHTSA recall flat file FLAT_RCL.zip, and the Part 573 report for 19V-298 fetched from static.nhtsa.gov/odi/rcl/. NHTSA manufacturer communication files TSBS_RECEIVED_2015-2019, _2020-2024, _2025-2026 -- https://static.nhtsa.gov/odi/ffdd/tsbs/ (dictionary TSBS.txt) Loaded into SQLite and queried with Python 3 and the sqlite3 module. -------------------------------------------------------------------------- CASE SENSITIVITY -- READ THIS BEFORE REUSING THE QUERIES -------------------------------------------------------------------------- NHTSA complaint narratives are ALL CAPITALS for complaints received through 2020 and mixed case from 2021 onward. A case-sensitive regex, or an upper-case SQLite GLOB, silently drops most of the post-2020 file while still returning hundreds of rows, so the query looks like it worked. Every query below is case-insensitive on purpose: - the SQL stage uses LIKE, which is case-insensitive for ASCII in SQLite; - every Python regular expression carries re.IGNORECASE; - no GLOB, and no upper()/lower() comparison against a cased literal. -------------------------------------------------------------------------- WHICH COMPLAINTS COUNT (the timing-belt set: 853) -------------------------------------------------------------------------- STAGE 1 -- SQL pull, case-insensitive LIKE on the narrative: descr LIKE '%timing belt%' -> 858 complaints. STAGE 2 -- makes excluded: motorcycle, heavy-truck, bus, trailer and RV makers, and rows where make is UNKNOWN (the same list our other analyses use). That removes five rows. RESULT: 853 complaints. No further keyword filtering is applied. The phrase "timing belt" is unambiguous as a phrase; the ambiguity, and it is a large one, is in what the owner means by it. See the next section. -------------------------------------------------------------------------- SAMPLE READING -- AND THE BELT/CHAIN PROBLEM, WHICH IS THE HEADLINE -------------------------------------------------------------------------- Thirty complaints were drawn at random (Python random.seed(1234) over the 853-row set) and read in full. For each one we looked up which timing drive the described vehicle's engine family uses. 15 of 30 describe a vehicle whose engine uses a TIMING BELT. 13 of 30 describe a vehicle whose engine uses a TIMING CHAIN. 2 of 30 are AMBIGUOUS, because that model year was sold with both a belt engine and a chain engine and the narrative does not say which. THIRTEEN IN THIRTY -- 43% -- OF COMPLAINTS THAT SAY "TIMING BELT" ARE ABOUT A CAR THAT DOES NOT HAVE ONE. Two of the thirteen say so themselves inside the same narrative: one writes "timing chain tensioner failure" and then calls the part a timing belt tensioner two sentences later, and one writes, in so many words, "we are dealing with a chain, not a timing belt." The full sample, with make, model, model year, the mileage the owner reported, our belt-or-chain verdict and a one-line note on how the verdict was reached, is published as belt-or-chain-read-sample.csv. Readers who disagree with a row can check it against the vehicle's own owner's manual. CONSEQUENCE FOR EVERY OTHER NUMBER IN THIS FILE: treat "853 timing-belt complaints" as 853 complaints that USE THE PHRASE, of which a large minority are chain vehicles. Where a figure would be distorted by that -- notably the mileage distribution -- the article says so out loud rather than quietly correcting it, because we cannot classify all 853 by hand and we are not going to guess at the other 823. This is not a criticism of the owners who filed. It is the strongest evidence we have found for the premise of the article itself: if 43% of people writing to the federal government about their timing belt do not have one, the two-minute belt-or-chain check is not a formality. -------------------------------------------------------------------------- FIELDS AND RESULTS -------------------------------------------------------------------------- Outcome flags (share answering Y, of 853): towed 8.32% (71), crash 1.76% (15), fire 1.06% (9), 4 injuries, 0 deaths. All-complaint baseline (777,040): towed 5.11%, crash 4.37%, fire 1.82%. The tow share is the one that matters: it is 63% above the all-complaint baseline, which is what a drivetrain failure that ends the trip looks like. Mileage: NHTSA MILES field, rows with 0 < miles < 400,000 (525 rows, 61.5% of the set). Median 93,000; quartiles 68,000 and 124,000. under 60,000 94 17.9% 60,000-89,999 143 27.2% 90,000-104,999 80 15.2% 105,000-119,999 55 10.5% 120,000 and above 153 29.1% All-complaint baseline median (392,935 usable rows): 70,000. WHAT THE NARRATIVE SAYS (categories overlap; shares do not sum to 100) Names a recall, extension, campaign or bulletin 277 32.47% (all-complaint baseline for the same regex: 32.49%) Vehicle stalled or lost power while driving 216 25.32% Engine damage: bent valves, pistons or a replaced engine 160 18.76% Water pump named 121 14.19% Tensioner, idler pulley or belt guide named 102 11.96% Belt broke, snapped, shredded, slipped or jumped 67 7.85% Says the belt or service had been done recently 13 1.52% Wet belt: belt in oil, belt debris, clogged pick-up screen 10 1.17% MAKES AND MODELS By make: Ford 134, Honda 124, Chevrolet 110, Volkswagen 93, Hyundai 45, Kia 43, GMC 33, BMW 27, Acura 26, Nissan 21, Subaru 21, Toyota 19. By model: Honda Pilot 52, Chevrolet Equinox 48, Honda Odyssey 27, Ford Focus 26, Ford F-150 26, Volkswagen Passat 24, GMC Terrain 24, Volkswagen Jetta 23, Ford Escape 23, Chevrolet Traverse 21, Ford EcoSport 21, Honda Accord 20, Kia Sorento 20, Acura MDX 16. READ THIS WITH THE BELT/CHAIN CAVEAT ABOVE. The Chevrolet Equinox, GMC Terrain, Chevrolet Traverse and Ford F-150 rows are chain vehicles whose owners wrote "timing belt"; the Honda Pilot, Odyssey, Accord V6 and Acura MDX rows are genuine belt engines. MANUFACTURER BULLETINS (39 distinct documents) Rows whose Summary contains "timing belt", case-insensitive, collapsed from one row per make/model/year to distinct NHTSA document IDs, after the heavy-truck and engine-maker exclusions. Thirty-nine documents, a very small body of writing compared with 2,467 for rattles or 2,190 for squeaks -- which is itself the point: the belt is a maintenance item, not a warranty dispute. Of the 39: 11 (28.2%) concern oil or a belt running in oil; 9 (23.1%) change or clarify an inspection or replacement interval; 5 (12.8%) name a tensioner or idler. By make: Volvo 12, Acura 6, Ford 5, Chevrolet 5, Honda 4, Volkswagen 3, Jeep 2, Subaru 1, Audi 1. RECALL CAMPAIGNS (2) Campaigns whose defect, consequence or remedy text contains "timing belt", case-insensitive: exactly two, covering 93,996 vehicles and parts. 19V-298 (Honda, 93,946 vehicles, report submitted 11 April 2019) -- "The teeth on an improperly hardened timing belt are susceptible to separation." Estimated percentage with defect: 1%. Supplier: Gates Unitta Asia Company. The Part 573 form's field "Identification of Any Warning that can Occur" is answered NR -- not reported. 14E-007 (The Gates Corporation, 50 aftermarket water pumps) -- a recalled REPLACEMENT part, where "a failure of the timing belt may cause the engine to shut down." PRIMARY DOCUMENTS OPENED FOR THE ARTICLE Part 573 Safety Recall Report 19V-298, submitted 11 April 2019, fetched from static.nhtsa.gov/odi/rcl/2019/RCLRPT-19V298-6327.PDF Ford Customer Satisfaction Program 26L02, 25 August 2026 (11036546) Ford Customer Satisfaction Program 26L01, 25 August 2026 (11036545) GM Service Update 17466, January 2018 (10137157) Volkswagen scheduled-maintenance amendment, 2018 (10130695) All fetched from static.nhtsa.gov and returning HTTP 200 on 18 September 2026. -------------------------------------------------------------------------- LIMITATIONS -- READ BEFORE QUOTING ANY NUMBER -------------------------------------------------------------------------- 1. Complaints are self-reported, unverified and voluntary. They are not failure rates and cannot be converted into one: NHTSA publishes no denominator, so nobody knows how many belt-driven engines are on the road, and the belt population is shrinking every year as new models ship with chains. 2. THE BELT/CHAIN CONTAMINATION IS THE LARGEST SOURCE OF ERROR IN THIS FILE. See SAMPLE READING. Every count here should be read as a count of complaints using the phrase. 3. A timing belt is a scheduled maintenance item. An owner who skipped the service and broke a belt often does not file a federal complaint, because they know whose fault it was. The set therefore over-represents belts that failed early, failed after a recent service, or failed in a way the owner believed was a defect. 4. Mileage is blank or unusable in 38.5% of the set, and is the owner's own entry. 5. Recall "potentially affected" is the manufacturer's estimate of vehicles in the population, not of vehicles that have the defect. 19V-298 states its own estimated defect rate: 1%. 6. Bulletin counts are counts of documents, not of vehicles or repairs. -------------------------------------------------------------------------- FILES -------------------------------------------------------------------------- belt-or-chain-read-sample.csv the 30 complaints read by hand, with our belt-or-chain verdict and the reason. timing-belt-what-the-narrative-says.csv category, complaints, share of the 853. timing-belt-mileage-distribution.csv mileage band, complaints, share of the 525 complaints reporting a usable figure. timing-belt-complaints-by-make.csv make, complaints, median miles, miles reported, percent towed, percent naming engine damage. timing-belt-top-models.csv same plus top model year, for the 45 most-reported models. No personal data is published: no VIN, no city, no dealer name, no owner free text beyond short quoted snippets in the article itself.