PULSCAR ANALYSIS OF NHTSA PUBLIC DATA — POWER STEERING FLUID LEAKS Published September 2026. License: CC BY 4.0 (credit "PulsCar analysis of NHTSA public data, September 2026"). ------------------------------------------------------------------------------- 1. SOURCE FILES ------------------------------------------------------------------------------- NHTSA public ODI flat files, downloaded 17 September 2026 from https://static.nhtsa.gov/odi/ffdd/ (field definitions: cmpl/CMPL.txt, rcl/RCL.txt, ftsb/FLAT_TSBS.txt), loaded into a local SQLite database. complaints 777,040 rows, received 1 January 2015 - 8 September 2026 recalls one row per make/model/year per campaign (counted DISTINCT) tsbs 187,827 distinct manufacturer communications investigations ODI defect investigations Owner complaints are self-reported, unverified and have no denominator. Nothing here is a failure rate, and no reliability ranking is made from these counts. ------------------------------------------------------------------------------- 2. THE COMPLAINT SET (998) ------------------------------------------------------------------------------- Query: a) narrative contains POWER STEERING and one of LEAK / LEAKING / LEAKED -> 1,328 b) AND at least one occurrence of LEAK sits within 110 characters of POWER STEERING, or of P/S, or of STEERING FLUID / PUMP / HOSE / RACK / GEAR / RESERVOIR — so that the leak and the steering system are in the same thought rather than in the same complaint c) AND the model is not a motorcycle (same exclusion pattern used throughout our work) Result: 998 complaints, received 1 January 2015 - 8 September 2026. Headline figures on those 998: name a hose, line, fitting or clamp 324 32.5% name the rack, steering gear or pinion 315 31.6% name the pump 258 25.9% name a seal 59 5.9% name a reservoir or a cap 67 6.7% name a rack boot 50 5.0% mention fire, smoke or burning 112 11.2% carry NHTSA's fire flag 30 3.0% (baseline 1.8%) name the exhaust, a manifold or a catalytic converter 48 4.8% describe lost or heavy steering assist 153 15.3% mention a recall 334 33.5% mention a serpentine belt 44 4.4% towed 26 2.6% (baseline 5.1%) crash reported 11 1.1% (baseline 4.4%) Failure mileage: 712 report a usable figure (0 and >= 400,000 dropped). Median 85,000 miles; 25th percentile 56,325; 75th percentile 126,127. The fire numbers are the reason this set was worth building. A power steering leak is the only one of the common under-hood leaks whose fluid routinely lands on an exhaust component, and 3.0% of these complaints carry the federal fire flag against an all-complaint baseline of 1.8%. Model-year distribution is old-skewed: 140 of the 998 are model year 2015 or newer and only 30 are 2020 or newer, and the newest entries are dominated by trucks and vans — Tacoma, Ram 1500 and ProMaster, Sierra 2500, E-350, Savana, Freightliner and Ford chassis cabs. That is consistent with hydraulic assist surviving longest on heavy vehicles, but it is a complaint distribution, not a census of what is fitted to what. NHTSA's own recall component list carries both "STEERING:ELECTRIC POWER ASSIST SYSTEM" (93 campaigns) and "STEERING:HYDRAULIC POWER ASSIST" in its several forms (95 campaigns), so both systems are demonstrably alive in the fleet and no single model year separates them. By make: ps-leak-complaints-by-make.csv (makes with >= 8 complaints). By model: ps-leak-top-models.csv (models with >= 8). Each carries the share naming a hose, the rack and the pump, the share mentioning fire, smoke or burning, and the median reported failure mileage. ------------------------------------------------------------------------------- 3. 110 COMPLAINTS READ IN FULL — WHICH ADDRESS ------------------------------------------------------------------------------- 110 drawn at random (random.seed(2026)) and read in full, one primary outcome each: a hose, line, fitting or clamp is the named address 27 24.5% the rack or steering gear is the named address 23 20.9% a leak is confirmed but the narrative never names the address 16 14.5% off target, or a multi-fault list in which a steering leak is one line 16 14.5% the pump or its shaft seal is the named address 10 9.1% no leak: electric power steering, or hydraulics ruled out 10 9.1% fire, smoke or burning is the substance of the report 6 5.5% a rack boot is the named address 2 1.8% FALSE POSITIVES. The off-target category — 16 of 110, 14.5% — is the measured false-positive rate of the published query. Reading the first 30 of the sample on their own gives 4 of 30, which is consistent. The "no leak" category is not a false positive: those complaints are on topic and are some of the most useful in the set, because they are people who went looking for a leak in a car that has no hydraulic fluid to lose. WHAT THAT CHANGED IN THE ARTICLE. The page previously ranked the five leak addresses as hoses, then the pump, then the rack. Both the keyword counts on the 998 (hoses 32.5%, rack 31.6%, pump 25.9%) and the 110 read in full (27, 23, 10) put the rack ahead of the pump, and the article now says so. The page also previously claimed that "half the time the answer is embarrassingly cheap." We found no support for that: reservoirs and caps are named in 6.7% of the 998 and are the primary address in none of the 110 we read. That claim has been removed rather than re-worded. Published as ps-leak-addresses-sample110.csv: complaint id, make, model, model year, reported failure mileage, assigned address. No free text, no VIN, no dealer name, no city. ------------------------------------------------------------------------------- 4. RECALLS (70 campaigns, 1,818,358 vehicles) ------------------------------------------------------------------------------- Campaigns whose defect description names power steering together with a leak, fluid or a hose: 70 distinct campaigns covering 1,818,358 vehicles or units. 30 of the 70 name FIRE in the consequence field. The largest and most instructive: 12V-222 (Honda, May 2012, 625,762 vehicles; 2007-2008 Acura TL and 2003-2007 Accord V6). "Prolonged under-hood and power steering fluid temperatures may cause the power steering hose to deteriorate prematurely, resulting in cracks and power steering fluid leakage." Consequence: fluid "can leak onto a hot catalytic converter, leading to smoke and possibly an under-hood fire." 10V-174 (Honda, April 2010, 167,255 vehicles; 2004-2008 Acura TSX). "Prolonged high under hood temperatures may cause the power steering hose to deteriorate prematurely and may cause the hose to crack and leak power steering oil." 14V-339 (GM, June 2014, 124,527 vehicles; 2015 Silverado and Sierra 2500/3500 HD). "The power steering hose clamp may disconnect from the power steering pump or gear, causing a loss of power steering fluid" — and a loss of power brake assist with it. 15V-878 (Chrysler, December 2015, 60,140 vehicles; 2015 Jeep Compass and Patriot). A retention clamp installed in the wrong place allows the low pressure return hose to detach; consequence names fire. 17V-184 (Nissan, March 2017, 56,766 vehicles; 2013-2014 Murano). Hose clamp does not adequately secure the hose; consequence names fire. 20V-663 (Daimler Vans, October 2020, 47,667 vehicles; 2015-2020 Sprinter). Return line hoses may detach at their connection points. 22V-089 (Ford, February 2022, 44,219 vehicles; 2021-2022 E-350 and E-450). Inadequate connection between the power steering pressure line and the brake Hydroboost unit; loss of power steering AND power brake assist. Caveat: "potentially affected" is a population, not a count of failures, and the same vehicle can appear in more than one campaign. ------------------------------------------------------------------------------- 5. INVESTIGATION EA05-019, "POWER STEERING FLUID FIRES" ------------------------------------------------------------------------------- Opened 7 November 2005, closed 12 January 2006. ODI received 29 reports of engine compartment fires related to power steering hose leakage on model year 1999 and 2000 Chrysler Cirrus, Dodge Stratus and Plymouth Breeze vehicles with 2.4L engines. The resume states that the fires occurred either from a hose leak or "after an extended drive and shortly after being parked - conditions consistent with ignition of power steering fluid on a hot engine or exhaust surface during the vehicle hot soak period." DaimlerChrysler filed recall 06V-001 on 6 January 2006, stating that "the high pressure power steering hose may prematurely heat, age and develop a leak," that it was aware of 94 complaints and field reports, and that the remedy was a replacement hose "made from a hose material with increased heat resistance capability." The investigation was closed on that basis. The hot-soak detail is the most practically useful sentence we found anywhere in this work, and it is quoted in the article: the dangerous moment is not necessarily while driving. ------------------------------------------------------------------------------- 6. MANUFACTURER COMMUNICATIONS (94) ------------------------------------------------------------------------------- 94 distinct bulletins name power steering together with a leak or fluid. By make the largest groups are Chevrolet, GMC, Cadillac and Buick — the GM full-size platforms. The one read for this article: GM PIP5516 and PIP5516A (February 2018 and April 2019), for 2016-2017 Chevrolet Traverse, GMC Acadia and Buick Enclave: "a possible power steering fluid leak above the engine harness to the rear (Bank 1) heated Oxygen sensor and direction to replace all involved parts due to this contamination and leak." This is the collateral-damage pattern the article now covers: a steering leak arriving at the service desk as an emissions fault. ------------------------------------------------------------------------------- 7. WHAT WE DID NOT DO ------------------------------------------------------------------------------- - No normalization by vehicles on the road, and no reliability ranking. The Honda Accord's 72 complaints and the Chevrolet Traverse's 68 reflect how many were built and how old they are as much as anything else. - No claim about which model years are hydraulic and which are electric. We report the complaint distribution and NHTSA's own component categories, and the article tells readers to look for a reservoir instead. - No fluid-specification claims by brand. The complaint file does not contain them, we did not open every owner's manual, and the article now says the cap and the manual are the only authority. - No price re-verification. The repair price ranges in the article are prior published estimates and are labelled as such; consumer estimator sites block automated access and we do not work around that. ------------------------------------------------------------------------------ CASE SENSITIVITY — SECTION ADDED 19 SEPTEMBER 2026 ------------------------------------------------------------------------------ NHTSA stores complaint narratives in ALL CAPS for records filed through roughly 2020 and in mixed case for records filed from 2021 on. A case-SENSITIVE search of the narrative field loses about 92% of the complaints filed after 2020, so this is the single largest way an analysis of this database can go wrong while still looking finished. Every query behind this file is case-insensitive. SQLite's LIKE and its FTS5 full-text index both ignore case by default, and no query in this work uses GLOB or any case-sensitive comparison. This was not taken on trust. On 19 September 2026 every published folder was re-checked against the same database, comparing the figures here with what a case-sensitive query would have returned. The result: step (a) was re-run case-insensitively on 19 September 2026 and returned 1,328 complaints — exactly the figure published above. The same step run case-sensitively returns 982. The published work is the case-insensitive one. (The terms are written in capitals in section 2 as a notation for readability; that is the write-up's convention, not the query's.)