PULSCAR ANALYSIS — OXYGEN SENSORS, P0420 AND CATALYST CODES IN NHTSA PUBLIC DATA Pulscar analysis of NHTSA public data, September 2026. License: CC BY 4.0. Attribution: "Pulscar analysis of NHTSA public data, September 2026." Published alongside https://pulscar.io/blog/o2-sensor-replacement-cost ------------------------------------------------------------------------------ 0. WHY THIS ANALYSIS EXISTS ------------------------------------------------------------------------------ An earlier version of the article carried the claim that the downstream oxygen sensor is the real cause of a P0420 in "roughly 25-35% of cases". We could not find a source for that number, so we removed it and went to the public record instead. This file documents what we found and, just as importantly, what this data can and cannot tell you. It CANNOT tell you what share of P0420 codes are fixed by a sensor. NHTSA complaints are filed by people with expensive, unresolved or dangerous problems. A P0420 cleared by a $200 sensor almost never produces a federal complaint. What this data CAN tell you is what the expensive, unresolved P0420 stories are made of, and what manufacturers themselves write about the code. ------------------------------------------------------------------------------ 1. SOURCE FILES ------------------------------------------------------------------------------ COMPLAINTS — 777,040 owner complaints received 1 January 2015 to 8 September 2026. https://static.nhtsa.gov/odi/ffdd/cmpl/CMPL.txt RECALLS — one row per make/model/year per campaign; campaigns counted DISTINCT. https://static.nhtsa.gov/odi/ffdd/rcl/RCL.txt TSBs — manufacturer communications, counted DISTINCT by NHTSA document id. https://static.nhtsa.gov/odi/ffdd/tsbs/FLAT_TSBS.txt No third-party data was used. Complaints are self-reported, are not failure rates, and have no denominator. ------------------------------------------------------------------------------ 2. BUILDING THE COMPLAINT SET ------------------------------------------------------------------------------ Full-text search of the narrative field, union of these terms: "oxygen sensor", "oxygen sensors", "o2 sensor", "o2 sensors", P0420, P0430, P0135, P0141, P0131, P0136, P0137, P0138, P0140, P0150, P0151, P0155, P0161 -> 1,446 complaints. Kept only those whose narrative also contains an oxygen-sensor phrase (oxygen sensor, o2 sensor, air-fuel ratio sensor, A/F sensor, lambda sensor) or a catalyst / sensor DTC, AND a fault or money cue (fail, replace, bad, faulty, defect, went out, not working, malfunction, code, check engine, light, error, quote, cost, $, diagnose). -> 1,380 complaints. This is the set used for the by-make counts and the cost figures. The P0420/P0430 subset — complaints whose narrative contains the literal code — is 376 complaints. Median reported odometer 100,000 miles (n=121; quartiles 74,350 and 126,700). ------------------------------------------------------------------------------ 3. READING 100 P0420 / P0430 COMPLAINTS IN FULL ------------------------------------------------------------------------------ 100 complaints drawn at random from the 376 (Python random.seed(11), random.sample) and read end to end. Each assigned exactly one primary outcome — what the story turned out to be about: 36 an upstream ENGINE fault was the real cause — excessive oil consumption, misfires, failed fuel injectors, a stretched timing chain, coolant intrusion, leaking valve seals, an engine already being replaced 18 the catalytic converter itself was named or replaced as the answer 16 off target — the code is incidental to a complaint about a transmission, steering, brakes or an unrelated stall 8 software, a PCM/ECM calibration or an emissions campaign was the event or the remedy (Ford 24E12 and 23B11/24E04, Kia SC176, a Toyota ECM update) 7 a repeat failure — a converter and/or sensors had already been replaced and the code returned 7 no diagnosis stated 5 an oxygen sensor was the named suspect or was replaced. In 4 of those 5 the owner states the sensor did NOT clear the code 3 a physical exhaust fault — a cracked support-bracket weld, a hole in the pipe, an aftermarket converter that set the light after installation --- 100 Per-complaint categories are published in p0420-what-it-turned-out-to-be.csv so the classification can be checked. Categorisation is one reader's judgement and a second reader would not assign every borderline case identically; the categories are mutually exclusive by construction. READ THIS BEFORE QUOTING THE 36: it is 36 of 100 COMPLAINTS, not 36% of P0420 codes on the road. See section 0. ------------------------------------------------------------------------------ 4. MANUFACTURER BULLETINS THAT MENTION P0420 OR P0430 ------------------------------------------------------------------------------ SELECT t.* FROM tsbs t JOIN tsbs_fts f ON f.rowid = t.rowid WHERE tsbs_fts MATCH 'P0420 OR P0430'; -> 342 distinct documents, dated 7 February 2015 to 10 August 2026. Reading their summaries for each phrase family: software, reprogramming, recalibration or a flash 175 (51.2%) an engine cause named — oil consumption, misfire, injectors, pistons, valve seals, coolant intrusion 88 (25.7%) an oxygen sensor named at all 26 (7.6%) instruct replacing the catalytic converter 9 (2.6%) a warranty extension or customer-satisfaction campaign 7 (2.0%) an exhaust leak named 5 (1.5%) instruct replacing an oxygen sensor 0 (0.0%) By make the largest counts are Ram 70, Honda 66, Ford 38, Jeep 33, Volkswagen 24, Acura 19, Land Rover 12, Chevrolet 11, Dodge 11. These count documents, not failures; a manufacturer that publishes more bulletins is not thereby worse. The zero is the finding worth stating carefully: across 342 documents in which a manufacturer discusses this code, none of the summaries instructs replacing an oxygen sensor as the remedy, and only 26 mention one at all. That is evidence about what manufacturers write, not proof that a sensor is never the answer. Published as p0420-manufacturer-bulletins.csv. ------------------------------------------------------------------------------ 5. OWNER-REPORTED DOLLAR FIGURES ------------------------------------------------------------------------------ Within the 1,380 complaints, every "$" amount between $80 and $8,000 was kept only when, within 100 characters on each side, the text contained a cost verb AND an oxygen-sensor or catalyst phrase AND none of: down payment, car payment, per month, monthly, purchase price, worth, value of, loan, deductible, insurance, lease, rental, trade, transmission, engine replacement, water pump, air conditioning, brake, tire, suspension, theft, stolen, thermostat, crankshaft, camshaft, canister, purge valve, ignition coil, spark plug, fuel pump, lead frame, exhaust manifold, diagnostic, tune-up, charcoal. Identical amounts repeated inside one complaint were counted once. -> 60 figures from 56 complaints. All figures: median $1,000; quartiles $500 and $2,165; range $99 to $6,000. under $300 7 (11.7%) $300-$599 10 (16.7%) $600-$1,199 17 (28.3%) $1,200-$2,499 13 (21.7%) $2,500 and up 13 (21.7%) Split by what the money was attached to in the sentence: attached to an oxygen sensor only: 38 figures, median $800 (quartiles $387 and $1,194) attached to a catalytic converter only: 17 figures, median $2,165 (quartiles $1,704 and $3,500) both or unclear: 5 figures Sample reading: 30 of the extracted figures were drawn at random and read in context. 25 were a genuine oxygen-sensor or converter figure. Two were not — one a multi-repair running total, one a price *difference* between two dealer quotes. Three bundled the sensor with other parts in a single number (an exhaust manifold, a harness and module, a five-part repair). Observed impurity rate near 17%. These are 60 figures, which is a small sample, and they are owner-typed rather than invoices we have seen. The sensor median of $800 sits well above the $150-$500 the article quotes for a single sensor installed; the honest reading is that these are the sensor bills people were angry enough to report to a federal agency, several of them dealer prices or multi-sensor jobs. Use the sensor/ converter GAP — roughly $800 against roughly $2,165 — rather than either number on its own. ------------------------------------------------------------------------------ 6. RECALLS ------------------------------------------------------------------------------ 13 distinct campaigns name a catalytic converter or an oxygen sensor in the defect, component or remedy text. The largest by far, and the one that shapes the complaint file: 19V-120 Kia, February 2019, 378,967 vehicles — 2012-2016 Kia Soul with the 1.6L GDI engine. High exhaust gas temperatures may damage the catalytic converter, possibly resulting in abnormal engine combustion and damage to one or more pistons and possible connecting rod failure. Consequences recorded: engine stall, and a broken rod puncturing the block so leaking oil can contact the exhaust and cause a fire. Remedy: upgrade the Catalytic Overheating Protection ECU logic; replace the converter if damaged; replace the engine if damaged. Kia's number: SC176. That single campaign explains the shape of our complaint set: Kia Soul is the largest single model in both the full set (118 of 1,380) and the P0420 subset (83 of 376). It is a reporting effect and a real defect at the same time, and it is the clearest published example of a catalyst code being a symptom of something upstream rather than a converter that simply wore out. ------------------------------------------------------------------------------ 7. FEDERAL EMISSIONS WARRANTY (not our analysis - primary regulation) ------------------------------------------------------------------------------ 40 CFR 85.2103, read on eCFR (text in force 1 January 2026): (c)(1) For light-duty vehicles, light-duty trucks and medium-duty passenger vehicles the emission PERFORMANCE warranty period is 24 months or 24,000 miles, except 8 years or 80,000 miles for any nonconformity resulting from a failed specified major emission control component identified in paragraph (d). (d)(1) The emission DEFECT warranty period is two years or 24,000 miles, except that the following specified major emission control components have a warranty period of eight years or 80,000 miles: catalytic converters and SCR catalysts, AND RELATED COMPONENTS; particulate filters and traps; EGR components on compression-ignition engines; the emission control module; and certain batteries. 40 CFR 85.2104 sets out when a claim may be denied for non-compliance with the maintenance instructions, and what evidence of compliance may consist of: a validated maintenance log book, a showing that the vehicle was submitted for scheduled servicing, or an owner's statement plus a showing that proper parts were purchased and used. 85.2104(e) states the owner may perform maintenance more frequently than required. 40 CFR 85.2105(a): no valid emission performance warranty claim shall be denied on the basis of the use of a properly installed certified aftermarket part. 40 CFR 85.2109(a)(2) and (a)(4): the manufacturer must furnish a list of all items covered by the emission performance warranty for the full useful life, described as they would appear on a service receipt, and must state in boldface that maintenance, replacement or repair of emission control devices and systems may be performed by any automotive repair establishment using any certified part. WHAT WE DID NOT CLAIM: we do not claim that every oxygen sensor carries 8 years/80,000 miles. The regulation gives that period to catalytic converters "and related components", and it requires each manufacturer to publish the list of covered items in the vehicle's warranty booklet. Whether a particular vehicle's oxygen sensor is on that list is a question for that booklet, and that is exactly what the article tells readers to check. California and other states that follow CARB rules have their own, sometimes longer, coverage; we did not analyse those. ------------------------------------------------------------------------------ 8. LIMITATIONS, STATED PLAINLY ------------------------------------------------------------------------------ - Selection bias is the dominant limitation here and it runs one way: cheap, successful repairs are invisible in this data. - Only 726 of the 1,380 complaints carry a usable odometer; only 121 of the 376 P0420 complaints do. - The 100-complaint read is a sample of complaints, not of repairs. - The bulletin counts are keyword counts over summary text, not over the full PDFs; a bulletin whose summary omits a word is not counted for it. - Nothing here measures how often a shop sold a converter that was not needed. ------------------------------------------------------------------------------ 9. FILES ------------------------------------------------------------------------------ p0420-what-it-turned-out-to-be.csv the 100 read complaints with vehicle, odometer, assigned primary outcome and the first 180 characters of the narrative o2-and-catalyst-complaints-by-make.csv counts by make, P0420 share, median failure miles, oil-consumption/misfire mentions, recall mentions o2-and-catalyst-owner-reported-costs.csv 60 figures, each tagged sensor, converter or unclear p0420-manufacturer-bulletins.csv 342 bulletins with software / converter / oxygen-sensor / engine-cause flags method.txt this file No personal data is published: no VINs, no names, no cities, no dealer addresses. Narrative extracts are short quotations from complaint texts NHTSA already publishes in full.