Per mile driven, pedestrians in Great Britain were 2.15 times as likely to be hit by an electric or hybrid-electric car as by a petrol or diesel car (RR 2.15, 95% CI 2.05 to 2.26), and in urban areas the rate ratio reached 2.97 (Edwards et al., Journal of Epidemiology and Community Health 2024). That risk sits on top of a fleet that is scaling fast: global electric car sales passed 20 million in 2025 (IEA Global EV Outlook 2026). Writing the rule has proved easier than meeting it: between December 2024 and June 2026, the eight US recall campaigns we identified for alert sounds that failed FMVSS 141 covered 222,257 vehicles (NHTSA recall records, our tally). We aggregated data from NHTSA, the UNECE, the European Union, the Virginia Tech Transportation Institute, ACEA, the IEA, and six peer-reviewed studies. For the wider picture of how blind pedestrians use sound at crossings, see our tactile paving and accessible pedestrian signal statistics.
TL;DR
- Pedestrian casualty rate per 100 million miles: 5.16 for electric and hybrid cars vs 2.40 for petrol and diesel cars (Edwards et al., JECH 2024).
- Hybrids had 1.52 times the odds of a pedestrian crash in low-speed maneuvers (NHTSA DOT HS 812 371, 2017).
- FMVSS 141 requires alert sound from stationary through 30 km/h (49 CFR 571.141).
- UN R138 floor: 50 dB(A) at 10 km/h and 56 dB(A) at 20 km/h; ceiling: 75 dB(A) driving forward (UN Regulation No. 138).
- All new hybrid and pure electric vehicles in the EU need an AVAS since 1 July 2021 (EU Regulation 540/2014).
- Tesla’s Boombox recall covered 594,894 vehicles for obscuring the pedestrian warning sound (NHTSA Recall 22V235, 2022).
- Three Chevrolet Equinox EV alert-sound recalls covered 112,483 vehicles in total (NHTSA Recalls 24V925, 25V639, 25V878).
- Added warning sounds raised mean detection distance at 10 km/h to 39.8 m vs 22.1 m with no added sound (VTTI 2023).
- In a 90-condition lab test, pedestrians detected EV warning sounds less than 20% of the time in every condition (Hsieh et al., IJERPH 2021).
- NHTSA valued the benefits of FMVSS 141 at $320.0 million at a 3% discount rate (NHTSA FMVSS 141 Final Rule 2016).
- Global electric car sales are forecast at about 23 million in 2026 (IEA Global EV Outlook 2026).
1. Pedestrian Crash Risk: The Evidence Behind the Mandate
Electric and hybrid cars injured 5.16 pedestrians per 100 million miles in Great Britain, more than double the 2.40 rate for petrol and diesel cars (Edwards et al., Journal of Epidemiology and Community Health 2024). The finding that matters most is where the gap appears: the urban rate ratio was 2.97, while the rural ratio of 0.91 showed no clear excess risk. That pattern is consistent with a low-speed hearing problem rather than a vehicle-dynamics problem, because cities are where cars move slowly and engine sound is the main cue.
US data points the same way from a different angle. NHTSA’s 2017 research note, built on crash records from 16 States, found the hybrid disadvantage was sharpest in low-speed maneuvers such as turning, backing and parking (NHTSA, DOT HS 812 371, 2017). Its earlier 2011 update found that 1.24% of hybrid crashes during low-speed maneuvers involved a pedestrian (90 of 7,258) versus 0.75% for combustion cars (2,015 of 268,440), with an all-speed odds ratio of 1.35 (95% CI 1.16 to 1.56) (NHTSA, DOT HS 811 526, 2011).
| Metric | Value | Source |
|---|---|---|
| Pedestrian casualties per 100 million miles, electric and hybrid cars (Great Britain, 2013-2017) | 5.16 (95% CI 4.92 to 5.42) | Edwards et al., JECH 2024 |
| Pedestrian casualties per 100 million miles, petrol and diesel cars (same period) | 2.40 (95% CI 2.38 to 2.41) | Edwards et al., JECH 2024 |
| Rate ratio, electric and hybrid vs petrol and diesel, all areas | 2.15 (95% CI 2.05 to 2.26) | Edwards et al., JECH 2024 |
| Rate ratio, electric and hybrid vs petrol and diesel, urban areas | 2.97 (95% CI 2.41 to 3.7) | Edwards et al., JECH 2024 |
| Odds ratio of a pedestrian crash, hybrid vs combustion vehicle, all speeds (16 US States) | 1.20 (95% CI 1.09 to 1.33) | NHTSA DOT HS 812 371, 2017 |
| Odds ratio of a pedestrian crash, hybrid vs combustion vehicle, low-speed maneuvers | 1.52 (95% CI 1.33 to 1.74) | NHTSA DOT HS 812 371, 2017 |
| Odds ratio of a pedalcyclist (bicycle rider) crash, hybrid vs combustion vehicle, all speeds | 1.51 (95% CI 1.34 to 1.70) | NHTSA DOT HS 812 371, 2017 |
| Odds ratio of a pedalcyclist crash, hybrid vs combustion vehicle, low-speed maneuvers | 1.46 (95% CI 1.22 to 1.74) | NHTSA DOT HS 812 371, 2017 |
Context note: both datasets predate alert-sound mandates. Most recent available data: Edwards et al. (2013-2017 collisions, published 2024) and NHTSA DOT HS 812 371 (2000-2011 crash years, published 2017). The 2017 note counted 420 hybrid pedestrian crashes among 68,950 hybrids, against 77,283 among 15,259,443 combustion vehicles, so the hybrid sample is small. In Great Britain, the vehicle type was missing for 24% of pedestrians hit by a car or taxi.
2. What the Rules Require: FMVSS 141, UN R138 and EU 540/2014
The US and UN rules disagree on where “quiet” ends: FMVSS 141 requires alert sound up to 30 km/h, while UN Regulation No. 138 stops at 20 km/h. NHTSA chose a 30 km/h crossover speed (the speed above which tire and wind noise make a vehicle audible on its own) and set target detection distances of 5 m at 10 km/h, 11 m at 20 km/h and 19 m at 30 km/h, based on a 1.5-second mean reaction time for surprise events (NHTSA FMVSS 141 Final Rule 2016).
The two systems also measure loudness differently. FMVSS 141 works in one-third octave bands (narrow slices of the frequency spectrum) and adds a rule that volume must rise by 3 dB between each speed step, so a pedestrian can hear acceleration (49 CFR 571.141). UN R138 instead pairs overall minimums with a pitch rule: at least one tone must shift by an average of 0.8% per 1 km/h (UN Regulation No. 138, EU Official Journal text).
| Metric | Value | Source |
|---|---|---|
| FMVSS 141 speed range requiring alert sound | Stationary through 30 km/h | 49 CFR 571.141 |
| FMVSS 141 minimum volume increase between speed steps (stationary to 10, 10 to 20, 20 to 30 km/h) | 3 dB | 49 CFR 571.141 |
| FMVSS 141 two-band option, required band sum | 44 dB(A) stationary, 51 dB(A) at 10 to under 20 km/h, 57 dB(A) at 20 to under 30 km/h, 62 dB(A) at 30 km/h | 49 CFR 571.141 |
| FMVSS 141 compliance schedule (current text) | At least 50% of production from March 1, 2020; all covered vehicles from March 1, 2021 | 49 CFR 571.141 |
| UN R138 minimum overall sound level | 50 dB(A) at 10 km/h, 56 dB(A) at 20 km/h, 47 dB(A) reversing | UN Regulation No. 138 |
| UN R138 maximum overall AVAS sound level, driving forward | 75 dB(A) | UN Regulation No. 138 |
| UN R138 minimum pitch change of at least one tone, 5 to 20 km/h | Average of 0.8% per 1 km/h | UN Regulation No. 138 |
| EU AVAS mandate dates | New vehicle types from 1 July 2019; all new hybrid and pure electric vehicles from 1 July 2021 | EU Regulation 540/2014 |
Outlier note: Europe now runs a floor and a ceiling at the same time. The same EU regulation that mandates AVAS also tightens pass-by noise limits for ordinary cars: for M1 passenger cars at or below 120 kW per 1,000 kg, the limit falls from 72 to 70 to 68 dB(A), with the Phase 3 value applying to first registrations from 1 July 2026 (EU Regulation 540/2014). The 2014 text also allowed a driver switch that resets to on at every restart, while the UN R138 text states that any pause function shall be prohibited. Community noise policy pulls the other way, as our aircraft noise exposure statistics show for airports.
3. Compliance Failures: The FMVSS 141 Recall Record
The largest pedestrian-sound recall in this roundup was caused not by a broken speaker but by a feature: Tesla’s Boombox, which could play sounds through the external speaker and obscure the Pedestrian Warning System, led to a recall of 594,894 vehicles (NHTSA Recall 22V235, 2022). That campaign superseded an earlier one covering 578,607 vehicles that disabled Boombox in Drive, Neutral and Reverse; the second extended the block to Summon and Smart Summon.
The 2024-2026 failures are subtler and almost all software. GM’s 2024 Equinox EV recall was triggered when NHTSA’s own compliance test found the alert did not produce the required relative volume change from stationary to 10 km/h, with an estimated 100% of the 23,700 vehicles affected (NHTSA Recall 25V639 Part 573 report). Across three Equinox EV campaigns the total reached 112,483 vehicles (7,606 + 23,700 + 81,177). The lesson for compliance teams: FMVSS 141 is not only a loudness test, it is a test of how the sound changes.
| Metric | Value | Source |
|---|---|---|
| Tesla Model S, X, Y and 3 recalled for Boombox obscuring the pedestrian warning sound (first campaign, February 2022) | 578,607 vehicles | NHTSA Recall 22V063, 2022 |
| Tesla vehicles in the superseding Boombox recall (April 2022) | 594,894 vehicles | NHTSA Recall 22V235, 2022 |
| 2025 Chevrolet Equinox EV, alert sound volume insufficient (December 2024) | 7,606 vehicles | NHTSA Recall 24V925, 2024 |
| 2024 Chevrolet Equinox EV, volume change from stationary to 10 km/h too small, found by NHTSA testing | 23,700 vehicles (100% estimated defective) | NHTSA Recall 25V639, 2025 |
| 2025-2026 Chevrolet Equinox EV, insufficient volume change between stopped and low-speed movement | 81,177 vehicles | NHTSA Recall 25V878, 2025 |
| 2023-2025 Toyota Corolla Cross Hybrid, not enough warning sound in reverse | 73,528 vehicles | NHTSA Recall 26V203, 2026 |
| 2024-2026 Denago EV Nomad, Scout and Rover models, no warning sound while moving | 34,136 vehicles | NHTSA Recall 26V381, 2026 |
| Vehicles in the eight FMVSS 141 alert-sound recalls we identified, December 2024 to June 2026 | 222,257 vehicles | NHTSA recall records 24V925 to 26V381, our tally |
Math note: 222,257 = 7,606 (24V925) + 94 (25V366) + 23,700 (25V639) + 1,899 (25V830) + 81,177 (25V878) + 117 (26V099) + 73,528 (26V203) + 34,136 (26V381). The three smaller campaigns are Mercedes-Benz models with an incorrect warning sound while reversing (94 vehicles, NHTSA Recall 25V366), the Mercedes-AMG EQE SUV 53 whose sound did not keep rising with speed (1,899 vehicles, NHTSA Recall 25V830), and the 2025 Mercedes-Benz eSprinter, too quiet in reverse (117 vehicles, NHTSA Recall 26V099). The tally is not exhaustive. Every remedy listed was a software update or, for Denago, an AVAS remedy kit installed by dealers.
4. Detectability: Do Added Warning Sounds Actually Work?
On a real road at 10 km/h, an added warning sound played through the factory speaker was detected at 70.4 m, more than three times the 22.1 m measured with no added sound (Virginia Tech Transportation Institute, VTTI-05-086, 2023). The design of the sound matters as much as its presence: low-frequency-emphasis sounds performed best, while high-frequency sounds performed about the same as no sound at all. In VTTI’s earlier phase, about 30% of no-sound trials fell inside a possible strike window, while added-sound conditions rarely did.
The gap closes as speed rises. At 20 km/h, NHTSA’s own pass-by tests found combustion vehicles only 3.5 dB(A) louder on average than unequipped hybrids and EVs, against 6.2 to 8.5 dB(A) at 10 km/h (NHTSA FMVSS 141 Final Rule 2016). In a JASA study, “tire-road interaction noise becomes dominant” at 20 km/h, which extends detection range, while detection probability was not always 100% at 10 km/h even where the sounds met NHTSA minimums (Roan et al., JASA 2021). That is also why road surface choices interact with vehicle sound, a theme covered in our porous asphalt and quiet pavement statistics.
| Metric | Value | Source |
|---|---|---|
| Mean on-road detection distance at 10 km/h, average across added-sound configurations | 39.8 m (SD 17.6 m) | VTTI 2023 |
| Mean on-road detection distance at 10 km/h, no added sound | 22.1 m | VTTI 2023 |
| Mean on-road detection distance at 20 km/h, average across all configurations | 58.0 m (SD 7.7 m) | VTTI 2023 |
| Drop in mean detection distance when background noise increased (10 km/h / 20 km/h) | About 33% / about 28% | VTTI 2023 |
| Sound and speaker combinations that met both FMVSS 141 stationary requirements (band sum and one-third octave bands) | 3 | VTTI 2023 |
| Background noise levels tested in the JASA probability-of-detection study (16 subjects, 10 and 20 km/h) | 55 dBA and 60 dBA | Roan et al., JASA 2021 |
| Detection rate with warning sound at 51 dBA vs 46 dBA (lab, 30 participants) | 11.26% vs 6.96% | Hsieh et al., IJERPH 2021 |
| Detection rate with 50 dBA vs 70 dBA background noise (same lab study) | 13% vs 6.78% | Hsieh et al., IJERPH 2021 |
Context note: VTTI ran its on-road tests with background noise held at about 55 dBA, and its lab replications underestimated road results by 30.8% at 10 km/h and 46.1% at 20 km/h. The Hsieh lab study (Hsieh et al., IJERPH 2021) played warning sounds of 46 and 51 dBA against 50 to 70 dBA of background noise at distances of 2 to 10 m, and found distance had no significant effect on detection.
5. Cost and Benefit: What the Mandate Was Expected to Buy
NHTSA estimated that FMVSS 141 would prevent 2,390 injuries to pedestrians and pedalcyclists over the lifetime of a single model-year fleet (MY2020), plus 11 more from low-speed vehicles (NHTSA FMVSS 141 Final Rule 2016). Against roughly $41 million a year in installation and fuel costs, the agency valued the benefits at $320.0 million at a 3% discount rate, so on paper the rule pays for itself several times over.
The weight and fuel penalty is close to a rounding error: about 1.5 pounds per light vehicle and 2.3 extra gallons of fuel over a passenger car’s lifetime. The real compliance cost, as section 3 shows, has turned out to be software validation rather than hardware.
| Metric | Value | Source |
|---|---|---|
| Injuries to pedestrians and pedalcyclists prevented over the lifetime of the MY2020 hybrid and electric fleet | 2,390 | NHTSA FMVSS 141 Final Rule 2016 |
| Monetized benefits (3% / 7% discount rate) | $320.0 million / $247.5 million | NHTSA FMVSS 141 Final Rule 2016 |
| Annual fuel and installation costs for light EV, hybrid and low-speed vehicle fleet (3% / 7%) | $41.8 million / $41.3 million | NHTSA FMVSS 141 Final Rule 2016 |
| Cost range per alert-sound system | $50.49 to $125.34 | NHTSA FMVSS 141 Final Rule 2016 |
| Average added weight per light vehicle (with 2.3 extra gallons of fuel over a passenger car’s lifetime) | 1.5 pounds | NHTSA FMVSS 141 Final Rule 2016 |
Context note: these are 2016 projections, not measured outcomes. No public study we found has yet compared pedestrian crash rates of AVAS-equipped vehicles with pre-mandate vehicles, so the benefit figure remains an estimate. The compliance dates in the 2016 rule (September 2018 and 2019) were later pushed back to the March 2020 and March 2021 schedule shown in section 2.
6. Scale: How Many Quiet Vehicles Need an Alert Sound
More than 20 million electric cars were sold worldwide in 2025, up about 20% on 2024, and the IEA expects about 23 million in 2026 (IEA Global EV Outlook 2026). Every one of those battery electric and plug-in vehicles sold in a market that follows FMVSS 141, UN R138 or an equivalent rule needs a compliant alert sound, which turns AVAS from a niche speaker part into a high-volume safety component.
Europe shows how wide the scope can be. In 2025, battery-electric cars took 17.4% of new EU registrations, hybrid-electric cars 34.5% and plug-in hybrids 9.4% (ACEA 2025 registrations). Added together (17.4 + 34.5 + 9.4), that is 61.3% of new cars with some form of electric drive, although ACEA’s hybrid-electric category includes mild hybrids that cannot move on electric power alone and therefore fall outside UN R138’s scope.
| Metric | Value | Source |
|---|---|---|
| Global electric car sales, 2025 (battery electric, plug-in hybrid and extended-range) | Over 20 million | IEA Global EV Outlook 2026 |
| Growth in global electric car sales, 2025 vs 2024 | About 20% | IEA Global EV Outlook 2026 |
| Electric cars as a share of new car sales in China, 2025 | Nearly 55% | IEA Global EV Outlook 2026 |
| US electric car sales, 2025 | About 1.5 million (just under 10% of car sales) | IEA Global EV Outlook 2026 |
| Forecast global electric car sales, 2026 | About 23 million | IEA Global EV Outlook 2026 |
| New battery-electric car registrations in the EU, 2025 | 1,880,370 (17.4% share) | ACEA 2025 |
| New hybrid-electric car registrations in the EU, 2025 | 3,733,325 (34.5% share) | ACEA 2025 |
| New plug-in hybrid car registrations in the EU, 2025 | 1,015,887 (9.4% share) | ACEA 2025 |
Outlier note: FMVSS 141 also reaches beyond passenger cars. It covers electric and hybrid vehicles up to 4,536 kg (10,000 lb) gross vehicle weight rating plus low-speed vehicles, which is why a maker of small electric low-speed vehicles such as Denago appears in the recall record (see section 3).
7. Pedestrian Perception and Sound Design
When 226 online participants decided when it felt safe to cross in front of an unseen vehicle at 30 km/h, sounds with higher perceived loudness, such as continuous high-frequency tones, held them back most effectively (Bazilinskyy et al., Applied Acoustics 2023). Intermittent beeps worked less well, likely because the gaps carried no information, and plain tire noise was effective while being the least annoying sound tested. The design tension is clear: the loudness that deters risky crossings also adds noise to every street the vehicle drives down, the same trade-off emergency services face, as our emergency siren noise statistics show.
Field and survey work adds behavioral texture. In Coimbra, Portugal, researchers observed 379 pedestrians at six unsignalised crossings and found 41 of them (10.8%) interacted with an electric vehicle; pedestrians relying on hearing alone were often surprised by its approach (de Oliveira et al., European Transport Research Review 2024). In Spain, nearly half of 417 surveyed workers saw electric vehicles as a risk to other road users, yet they did not change how they walked in company parking areas (Pardo-Ferreira et al., IJERPH 2020).
| Metric | Value | Source |
|---|---|---|
| Participants in online crossing-decision experiment on exterior vehicle sounds | 226 | Bazilinskyy et al., Applied Acoustics 2023 |
| Vehicle speed simulated in that experiment | 30 km/h (constant) | Bazilinskyy et al., Applied Acoustics 2023 |
| Pedestrians observed at six unsignalised crossings in Coimbra | 379 | de Oliveira et al., ETRR 2024 |
| Share of observed pedestrians who interacted with an electric vehicle | 10.8% (41 pedestrians) | de Oliveira et al., ETRR 2024 |
| Blind participants in the Coimbra observational study | 3 | de Oliveira et al., ETRR 2024 |
| Worker questionnaires on electric vehicle noise risk (Malaga, Spain) | 417 | Pardo-Ferreira et al., IJERPH 2020 |
| Warning-sound level suggested by lab researchers for a roughly 50% chance of detection | At least 57 dB | Hsieh et al., IJERPH 2021 |
Context note: these are small or non-representative samples (an online panel, three blind volunteers, one employer survey), so they describe behavior and perception rather than population risk. Detection research on blind pedestrians specifically remains thin: the VTTI on-road tests in section 4 recruited 20 university students rather than blind pedestrians.
Summary: Electric Vehicle AVAS Acoustic Warning Compliance by the Numbers
| Metric | Value | Source |
|---|---|---|
| Pedestrian casualties per 100 million miles, electric and hybrid vs petrol and diesel cars (Great Britain) | 5.16 vs 2.40 | Edwards et al., JECH 2024 |
| Urban rate ratio, electric and hybrid vs petrol and diesel cars | 2.97 | Edwards et al., JECH 2024 |
| Odds ratio of a pedestrian crash, hybrid vs combustion vehicle, low-speed maneuvers (US) | 1.52 | NHTSA DOT HS 812 371, 2017 |
| Odds ratio of a pedalcyclist crash, hybrid vs combustion vehicle, all speeds (US) | 1.51 | NHTSA DOT HS 812 371, 2017 |
| Share of crashes during low-speed maneuvers that involved a pedestrian, hybrid vs combustion vehicle (2011 update) | 1.24% vs 0.75% | NHTSA DOT HS 811 526, 2011 |
| FMVSS 141 speed range | Stationary through 30 km/h | 49 CFR 571.141 |
| FMVSS 141 minimum volume increase between speed steps | 3 dB | 49 CFR 571.141 |
| UN R138 minimum overall level at 10 km/h / 20 km/h, and maximum AVAS level driving forward | 50 dB(A) / 56 dB(A), maximum 75 dB(A) | UN Regulation No. 138 |
| EU date from which all new hybrid and pure electric vehicles need an AVAS | 1 July 2021 | EU Regulation 540/2014 |
| Largest pedestrian-warning-sound recall (Tesla Boombox) | 594,894 vehicles | NHTSA Recall 22V235, 2022 |
| Largest single Equinox EV alert-sound recall | 81,177 vehicles | NHTSA Recall 25V878, 2025 |
| Vehicles in eight FMVSS 141 alert-sound recalls, December 2024 to June 2026 | 222,257 | NHTSA recall records, our tally |
| Mean detection distance at 10 km/h with added sound vs no added sound | 39.8 m vs 22.1 m | VTTI 2023 |
| Highest detection rate in any condition of a 90-condition lab test | Below 20% | Hsieh et al., IJERPH 2021 |
| Lifetime injuries prevented, MY2020 fleet (NHTSA projection) | 2,390 | NHTSA FMVSS 141 Final Rule 2016 |
| Global electric car sales, 2025 | Over 20 million | IEA Global EV Outlook 2026 |
| Battery-electric share of new EU car registrations, 2025 | 17.4% | ACEA 2025 |
Methodology and Sources
- NHTSA, FMVSS 141 Minimum Sound Requirements for Hybrid and Electric Vehicles, Final Rule (Federal Register, December 14, 2016): crash-risk basis, detection targets, cost-benefit analysis.
- 49 CFR 571.141, Standard No. 141 (eCFR, current text): speed range, band minimums, relative volume rule, phase-in, anti-tampering clause.
- NHTSA, DOT HS 812 371, Updated Analysis of Pedestrian and Pedalcyclist Crashes With Hybrid Vehicles (2017).
- NHTSA, DOT HS 811 526, Incidence Rates of Pedestrian and Bicyclist Crashes by Hybrid Electric Passenger Vehicles: An Update (2011).
- NHTSA recall records 22V063, 22V235, 24V925, 25V366, 25V639, 25V830, 25V878, 26V099, 26V203 and 26V381, read through the NHTSA recalls API and the 25V639 Part 573 report.
- UN Regulation No. 138, Quiet Road Transport Vehicles (EU Official Journal text).
- EU Regulation 540/2014 on the sound level of motor vehicles (EUR-Lex).
- Edwards, Moore and Higgins, Pedestrian safety on the road to net zero, Journal of Epidemiology and Community Health (2024), with figures cross-checked against the BMJ press release.
- Virginia Tech Transportation Institute, VTTI-05-086, A Data Driven Approach to the Development and Evaluation of Acoustic Electric Vehicle Alerting Systems for Vision Impaired Pedestrians (2023).
- Roan et al., Probability of detection of electric vehicles with and without added warning sounds, JASA (2021).
- Hsieh et al., Effect of Environmental Noise, Distance and Warning Sound on Pedestrians’ Auditory Detectability of Electric Vehicles, IJERPH (2021).
- Bazilinskyy et al., Exterior sounds for electric and automated vehicles: Loud is effective, Applied Acoustics (2023).
- de Oliveira et al., Insights into critical urban pedestrian dynamics, European Transport Research Review (2024).
- Pardo-Ferreira et al., New Risk Situations Related to Low Noise from Electric Vehicles, IJERPH (2020).
- IEA, Global EV Outlook 2026: the IEA site blocked automated access, so its 2025 sales figures were read through Industrial Info Resources’ report on the outlook.
- ACEA, New car registrations 2025.
- Data watch: the crash-risk evidence is old (NHTSA crash years 2000-2011; Great Britain collisions 2013-2017) and predates AVAS mandates, so no published study yet measures the real-world effect of alert sounds. NHTSA’s hybrid pedestrian odds ratio varies by version: 1.35 in 2011, 1.20 in the 2017 note (1.21 when Honda hybrids are recategorized), and the 2016 final rule also cites a 1.18 regression result. For Great Britain, the press release describes collisions as three times as likely in urban as in rural areas, a different comparison from the paper’s urban rate ratio of 2.97 against petrol and diesel cars. The 2016 final rule quotes per-vehicle costs in more than one way ($50.49 to $125.34 per system elsewhere vs $129.84 for unequipped hybrids and $54.99 for electric vehicles), and some of its benefit tables are garbled in the published text. UN R138 values come from the consolidated text published in the EU Official Journal (version of 10 October 2017); later UN series amendments may differ, and only the UNECE original has legal effect. The recall tally of 222,257 covers only the campaigns we identified and may be incomplete; Toyota’s 26V203 owner letters went out September 1, 2026, so remedy completion data is not yet available. IEA figures are shown as published; newer IEA monthly data may revise 2026 forecasts.
Last updated: October 18, 2026. We update this roundup quarterly, and the next refresh is expected when NHTSA publishes its next FMVSS 141 recall filings and the IEA releases Global EV Outlook 2027.