Investigation
Australia requires a seatbelt sensor in every new car. From November 2026 it requires new buses to carry a sticker, a lit sign and one recorded announcement. We read the regulatory record to find out why the difference exists. The documents reviewed point to the process used to adopt international vehicle standards. It is that nobody ever asked the question.
1. The question
Two Australian rules govern seatbelt warnings. They were written for the same purpose, by the same department, under the same Act.
ADR 4/05 covers cars. Since 2013 for new models, every new passenger car sold in Australia has had to carry a seatbelt reminder at the driver’s position: a sensor that knows whether that specific buckle is latched, and an alert that responds to the answer.
ADR 68/01 covers buses. From 1 November 2026 for new models and 1 November 2027 for all new vehicles, buses fitted with seatbelts must carry three things: a label on the back of each seat reading “Fasten seatbelt while seated”, a lit sign on the cabin bulkhead reading “Fasten seatbelt”, and a recorded announcement saying “Please fasten seatbelt while seated” when the last door closes.
No buckle sensor. No occupancy sensor. No detection of any kind. The industry body’s own guidance states it plainly:
I have read some articles that say seatbelt monitoring systems are required as part of this ADR change. Is this correct? — No, seatbelt monitoring systems that detect when someone is in the seat are not required.
Bus Industry Confederation, ADR 68/01 & Seatbelts FAQ, April 2025
So the driver of a Hyundai i30 gets a device that knows the state of their belt. Forty-eight people on a coach at 100 km/h get a recorded voice that plays whether every belt is fastened or none of them are.
| Cars — ADR 4/05 | Buses — ADR 68/01 | |
|---|---|---|
| Applies from | 1 July 2013 new models | 1 November 2026 new models |
| Knows if a belt is undone | Yes — buckle sensor | No |
| Knows if a seat is occupied | Not required | No |
| Alert responds to belt state | Yes | No — plays on a timer regardless |
| Still active during the trip | Yes | No — plays once, before 500 m or 25 km/h |
| Seats covered by sensing | 1 (the driver) | 0 |
| Belt wearing rate in that population | About 97% | 7% to 45% |
The vehicle class with the worst compliance problem in Australian road transport received the weaker instrument. This piece is about finding the document that explains that.
A naming problem worth clearing up first. The bus industry calls the ADR 68/01 measures “seatbelt reminder systems”. The car industry calls the ADR 4/05 device a “seatbelt reminder system” too. They are not the same thing. One senses a buckle and reacts. The other is a sign and a recording. Throughout this piece, “warning system” means something that detects the state of a specific belt. “Awareness measure” means a sign, a label or an announcement. The difference between those two categories is the entire subject.
2. How a rule like this gets made
You need two pieces of machinery to follow the rest of this.
Harmonisation. Since the late 1990s Australia has had a standing policy of aligning the Australian Design Rules with United Nations vehicle regulations rather than writing its own. Australia signed the UN 1958 and 1998 Agreements. Most ADRs now work by reference: a manufacturer shows a UN approval and the ADR is satisfied. The stated reasons are cost, trade and getting the latest vehicles into the market sooner. This matters more than it sounds.
Impact analysis. Before a rule that imposes costs is made, the department is expected to prepare a Regulation Impact Statement, now called an Impact Analysis, and have it assessed by what is today the Office of Impact Analysis. It contains a cost-benefit analysis. To value a prevented death, the government uses a published figure: the value of statistical life, currently $5.87 million in 2025 dollars.
That second point answers a question people often ask rhetorically. Is there a price on a life? Yes. It is written down, it is updated annually, and it is used to decide whether safety rules get made. Whether that is right is a separate argument. It is how the system works, and it is the arithmetic this investigation has to engage with.
3. Possibility one: the sensors cost too much
This is the intuitive answer, and it is the one the industry usually offers. Two sensors per seat across 57 seats, wiring, a control module, certification. A bus is not a car.
So we went and found what the government paid to justify the car rule. The Regulation Impact Statement for ADR 4/05 is public. It contains a costing table.
| Coverage | Simple system | Semi-complex system |
|---|---|---|
| Driver only | $18 | $61 |
| Driver and front passenger | $43 | $92 |
| All occupants | $104 | $195 |
Eighteen dollars for the driver’s position. One hundred and four dollars to cover every seat in the car. The marginal cost of each additional seat, on the government’s own numbers, is roughly twenty dollars.
Now the benefits side. The RIS concluded:
the safety benefits of installing a driver’s seat belt reminder, over the life of a vehicle, outweigh the installation costs.
Regulation Impact Statement, Amendments to Vehicle Standard (ADR 4/05 — Seatbelts)
And here is the size of the win it was making that claim about. On a 15-year vehicle life and a 7 per cent discount rate, the numbers the department took to the Office of Best Practice Regulation were:
| Measure | Value |
|---|---|
| Annual net benefit, 15-year fleet life, 7% discount | $276,219 |
| Benefit-cost ratio, same assumptions | 1.25 |
| Assumed effectiveness of the device | 7% |
| Belt wearing rate in the population it targeted | About 89% already had the device fitted voluntarily; wearing about 97% |
Read those two figures again, because they reframe the whole question.
Australia mandated the car driver’s seatbelt reminder on an annual net benefit of $276,219 — about one twentieth of the value the same government places on a single human life — and a benefit-cost ratio of 1.25. A ratio of 1.0 is the break-even point. This rule passed by a quarter.
It gets thinner. The device was credited with 7 per cent effectiveness, a figure the RIS describes as deliberately conservative, chosen over an Australian estimate of 10 per cent. And 89 per cent of cars already had one fitted voluntarily, so the rule was only ever reaching the last eleven per cent of the market.
Verdict on possibility one. Cost cannot explain the gap, because the standard applied to cars was not demanding. Australia was prepared to mandate a device on a benefit-cost ratio of 1.25 and a net benefit smaller than one statistical life, for a device assumed to be 7 per cent effective, in a population already wearing belts at 97 per cent. If that is the bar, the bus case does not obviously fail it. Nobody has published the sum.
4. Possibility two: the numbers genuinely failed for passengers
This is the serious version of the cost argument, and it deserves a proper hearing, because the ADR 4/05 record contains something the industry rarely mentions.
The 2004 analysis did not only cost the driver’s seat. It ran full benefit-cost ratios for three levels of coverage: driver only, front seats, and all occupants. Here is what happened when they priced reminders for passengers.
| Assumed effectiveness | Driver | Front seats | All occupants |
|---|---|---|---|
| 7% — the figure used | 1.25 | 0.69 | 0.33 |
| 10% | 1.79 | 0.99 | 0.47 |
| 15% | 2.69 | 1.49 | 0.70 |
Look down the right-hand column. Covering every seat in a car never gets close to paying for itself. Even at 15 per cent effectiveness — more than double what the department was willing to assume — the ratio is 0.70. Stretch the analysis to a 20-year vehicle life and the best all-occupants figure in the entire document is 0.49. The driver’s column crosses 1.0. The passenger columns never do.
So there is a documented cost-benefit case for not putting sensors in passenger car seats. It is a real finding.
But notice why it fails. A benefit-cost ratio for a seatbelt reminder is driven by one number above all others: how many people in that seat are currently unbelted. In an Australian car, about 97 per cent of occupants already wear the belt. There is almost nothing left to win. You are spending twenty dollars a seat to chase the last three per cent, and the arithmetic says don’t.
Every input in that calculation reverses on a bus.
| Variable | Car passenger seat | Coach passenger seat |
|---|---|---|
| Occupants currently unbelted — the size of the prize | About 3% | 55% to 93% |
| Seats protected by one installation | 4 | Up to 57 |
| Service life the cost is spread over | About 15 years used in the RIS | Approaching 26 years |
| Consequence of being unbelted in a rollover | Severe | Severe — and at Greta, all ten deaths involved ejection |
And the effectiveness assumption — the variable the car analysis was most cautious about — has since been measured on an actual bus. A Korean government institute fitted occupancy and buckle sensors to every seat of an interurban bus in commercial service and observed belt use rise from 9.6 per cent to 59.0 per cent. In the units the RIS uses, that is a reduction in the unbelted population of about 55 per cent. The car rule was justified on an assumption of 7 per cent.
Verdict on possibility two. This is the strongest cost argument available, and it is genuine — passenger-seat sensing failed the test in cars. But it failed because car passengers already buckle up. The sum was done for a population wearing belts at 97 per cent and has never been redone for one wearing them at 7 to 45 per cent, in vehicles carrying fourteen times as many people for ten years longer. The calculation was not rejected for buses. It was never run.
5. Possibility three: there was no international rule to copy
This is where the record gets uncomfortable, and where we think the real answer is.
Australia looked at seatbelt reminders in 2004 and decided not to regulate. The ADR 4/05 RIS explains why, and the reason is remarkable:
The RIS did find a positive case for fitting a simple low cost seatbelt reminder to the driver position, but at the time did not support a regulatory approach. The primary reason for this was that industry surveys and Department of Transport and Regional Services forecasts indicated that 95 per cent of passenger vehicles would be fitted with these devices by 2007. It was also highly significant that at the time there was no international regulation in place that covered seatbelt reminders. This would have made any Australian requirement unique and so costly for industry to implement.
ADR 4/05 Regulation Impact Statement, section 4
A positive safety case, set aside partly because no international regulation existed to align with. Then, seven years later:
A follow-up survey in 2010 indicated that a voluntary fitment rate of 95 per cent had not been achieved and is still not being achieved, reaching approximately 89 per cent. … This, coupled with UNECE R 16 mandating seatbelt reminder systems with a phase-in period starting from 2009, now delivers a compelling case in 2011 to also mandate the same requirements in Australia.
ADR 4/05 Regulation Impact Statement, section 4
Nothing about human physiology changed between 2004 and 2011. Belts did not become more effective. What changed is that the United Nations wrote a regulation, and Australia adopted it.
Which brings us to the sentence that, as far as we can tell, is the actual answer to the question in the title of this piece. Here is what UN Regulation 16 requires:
| Vehicle category | Front seats | Rear and passenger seats |
|---|---|---|
| M1 — passenger cars | Required | Required |
| N1 — light commercial | Required | Required |
| M2 and M3 — buses and coaches | Required | Not required |
The international rulebook that Australia harmonises to requires a reminder for the bus driver and the person sitting next to them, and nothing for the forty-eight people behind them. Our systematic review of the passenger car evidence sets out that rule-by-rule comparison in full, and what transfers to a bus. There is no UN regulation covering bus passenger seats for Australia to adopt.
Put the two documents together and the chain is complete:
- Australian policy is to align the ADRs with UN regulations rather than write unique national requirements.
- In 2004, the absence of an international regulation was given as a primary reason not to regulate seatbelt reminders, despite a positive safety case.
- In 2011, the existence of one was given as the reason the case had become compelling.
- UN Regulation 16 does not cover bus passenger seats.
Verdict on possibility three. This explanation concerns the regulatory process; evaluating safety outcomes requires evidence from bus operations. It is a procedural one about where rules come from. Bus passengers do not have a seatbelt warning requirement because no international committee has written one for Australia to inherit — and Australia’s stated position, in the one document where it addressed the point directly, is that writing a unique national requirement is something to avoid.
6. Possibility four: the driver is a special case
The obvious objection to everything above is that ADR 4/05 covers the driver. A driver controls the vehicle, holds a licence, carries a legal duty and can be fined. Extending that logic to a passenger is a category error. Maybe the rule is about drivers, not about seats.
That objection held until January 2025, when the United States removed it.
The US final rule amending FMVSS 208 requires a seat belt warning system for every rear seating position, phasing in from September 2027, and extends the front passenger requirement as well. The regulator estimated it would save about 50 lives and prevent more than 500 injuries a year.
The population that rule protects is people who are not driving, hold no duty in respect of the vehicle, and cannot be reached by any interlock. In other words, passengers. US rear-seat belt use sits around 80 per cent against about 92 per cent in front — and on that 12-point gap, the regulator concluded per-seat sensing was justified.
Australian bus passengers sit between 7 and 45 per cent.
There is one wrinkle, and it is telling. The US rule applies to vehicles with a gross weight rating up to 4,536 kg — which covers cars, light trucks and small buses, and excludes coaches. So the largest passenger vehicles on the road are once again outside the instrument. Not because they were assessed and excluded. Because the rule was written around a weight class.
Verdict on possibility four. Dead. The “it’s about the driver” argument is no longer available to anyone, because the world’s most active vehicle safety regulator has just mandated per-seat sensing for people who are not driving. Coaches were not argued out of that rule either. They were simply outside its scope.
7. Possibility five: nobody asked
At this point we went looking for the document that weighs per-seat sensing for bus passengers and rejects it. A cost-benefit analysis, a consultation response, a paragraph in an explanatory statement. Anything.
There isn’t one. Here is exactly where we looked.
| Where we looked | What we found |
|---|---|
| Office of Impact Analysis, published impact analyses | The ADR 4/05 seatbelt RIS is there, assessed as adequate in 2012. We located no published Impact Analysis for ADR 68/01 or for occupant protection in buses. |
| Departmental announcement of the new rule | States the requirements and the response to the Hunter Valley crash. No cost-benefit figures, no discussion of sensing. |
| Industry consultation record, Bus Safety Working Group | Describes the three awareness measures and the removal of the seatbelt exemption. No mention of sensing being considered. |
| Bus Industry Confederation guidance to operators | Confirms monitoring systems are not required. The reason offered for the measures chosen is that seatbelts reduce injury — not that signs and announcements raise wearing. |
| The Greta crash investigation recommendations | 41 recommendations. The seatbelt ones are a statewide awareness campaign and stronger enforcement. Detection is not among them. |
That last row matters more than the others. The reform pipeline that produced ADR 68/01 started with a crash investigation, and that investigation’s seatbelt recommendations were campaigns and enforcement. Sensing never entered the pipeline, so it was never at the table when the rule was drafted, so no document exists in which it was rejected.
Read the justification the industry body gives for the measures that were chosen, and you can see the gap directly:
There was a need to improve occupant safety, specifically seatbelt usage in buses where they are provided. Studies demonstrated that increasing seatbelt usage significantly reduces the risk of occupant injury or fatality in the event of an accident.
Bus Industry Confederation, ADR 68/01 FAQ, April 2025
Both sentences are true. Neither is a justification for the instrument chosen. The first says the problem is real. The second says seatbelts work. Nothing there says a sticker, a sign and one announcement will get more passengers to wear one — and no measured estimate for those three measures exists in any jurisdiction. In the wider literature the effect column for seatback signage and for a timed announcement on a bus is empty — we ranked fifteen behavioural mechanisms across thirteen fields and signs and announcements came bottom of every one.
Verdict on possibility five. Confirmed. There is no published analysis weighing per-seat sensing for bus passengers and finding against it. The gap is not the outcome of a decision. It is the absence of one.
8. The sum, with today’s prices
The cost side of this question is no longer hypothetical. Per-seat seatbelt warning systems are on sale in Australia now, with published prices. Here are ours.
| Vehicle size | Seats | Price per vehicle | Cost per seat at the top of the band |
|---|---|---|---|
| Small | Up to 20 | $3,199 | $160 |
| Medium | 21 to 36 | $5,099 | $142 |
| Large | 37 to 57 | $9,499 | $167 |
| X-Large | 58 to 120 | $13,899 | $116 |
So a full-size coach costs $9,499 to fit, and each passenger seat in it costs about $167 to cover.
Now put that against the government’s own price for a life: $5.87 million.
| Question | Working | Answer |
|---|---|---|
| How many coaches does one prevented death pay to fit? | $5,870,000 ÷ $9,499 | 618 coaches |
| How many would the ten lives lost at Greta pay to fit? | 10 × $5,870,000 ÷ $9,499 | 6,180 coaches |
| How many coaches does the annual net benefit that justified the car rule buy? | $276,219 ÷ $9,499 | 29 coaches |
That last row is the one to sit with. The entire annual net benefit Australia used to justify a national seatbelt reminder rule for every new car in the country was worth 29 coach fitments.
And the per-seat comparison runs the same way. The 2010 analysis priced a car reminder at roughly twenty dollars a seat and declined to mandate it for passengers, because only three per cent of car passengers are unbelted. A coach seat costs $167 to cover today, and the passenger in it is between 18 and 31 times more likely to be unbelted than that car passenger.
What a full impact analysis still needs is fleet size, crash frequency, injury distributions, the share of bus casualties attributable to non-wearing, and an Australian effectiveness figure. Those inputs exist or can be built. Nobody has assembled them.
9. So what is the justification?
There isn’t one. No document exists in which someone weighed a seatbelt warning system for bus passengers against its cost and concluded it wasn’t worth it.
What there is instead is a chain of procedure:
- Australia aligns its vehicle rules to United Nations regulations by policy.
- UN Regulation 16 requires seatbelt reminders on all car seats, and only the front seats of a bus.
- So there was nothing for Australia to adopt for bus passengers, and Australia’s stated position is that writing unique national requirements is costly and to be avoided.
- When ADR 68 was reopened after the Hunter Valley crash, the reform proposals on the table came from a crash investigation whose seatbelt recommendations were campaigns and enforcement.
- The measures that resulted are the three cheapest interventions available, and the effect of none of them has ever been measured on a bus passenger.
Each step is defensible on its own. The result is that Australia’s least-belted passengers got its weakest instrument, and no one had to argue for that outcome.
On the question of cost specifically, the record allows a firmer statement. Cost is not the explanation, because the price Australia was willing to pay for the car rule was very low: a benefit-cost ratio of 1.25, an annual net benefit of $276,219, a 7 per cent effectiveness assumption, and a population already at 97 per cent compliance. A rule mandated on those numbers does not sit easily beside a refusal to even cost the same device for a population at 7 to 45 per cent.
10. Cost should not be what decides this
AITSC publishes equipment prices to support cost and deployment planning. A coach fitment is under ten thousand dollars, fits in an afternoon, needs one auto-electrician, and carries no ongoing fees. More than 50 vehicles have already been fitted for a single Australian operator.
We are willing to work with regulators, transport departments and government at any level to get per-seat seatbelt warning into the Australian fleet. That includes fleet and programme pricing, supplying vehicles and equipment for the trial an impact analysis would need, opening our cost structure to the department preparing that analysis, and sitting down with anyone drafting the next revision of ADR 68.
If price is the obstacle, name the number that makes it work and we will start from there. A passenger’s life is not the kind of thing that should come down to the cost of a buckle sensor, and we do not want cost to be the reason this stays unbuilt.
11. What would close the gap
Two things, in order.
An impact analysis. The department should cost per-seat buckle and occupancy sensing for belt-fitted buses and publish the result, using the same method and the same value of statistical life it applied to cars in 2012. If the ratio comes out below 1.0, that is a real answer and the argument is over. Right now there is no answer at all.
An Australian trial. The missing input in any such analysis is a defensible effectiveness estimate for an Australian bus. One overseas pilot, which used a light rather than a sound, is thin ground for a national rule. We have published a protocol for a 96-vehicle randomised controlled trial that would produce that number, and which fixes in advance the results that would show the intervention adds nothing.
Until one of those exists, the position stands as it is: a car with one unbelted driver knows about it. A coach with forty unbelted passengers does not, and no one has written down why.
Related research: Read the empirical findings from the Korean bus seatbelt reminder study showing how automated reminders boosted compliance from 9.6% to 59%.
Further reading from this research programme
- Per-seat buckle sensing with a persistent localised audible alert — the evidence review, including a fuller comparison of ADR 4/05, ADR 68/01 and the US rule.
- What actually changes passenger behaviour — why signs and announcements sit at the bottom of the evidence in every field that has measured them.
- The Korean bus seatbelt reminder study — the only trial of per-seat sensing ever run on a bus.
- Information or consequence? — the protocol for the Australian trial.
Compliance reference: Consult our comprehensive ADR 68/01 compliance guide for bus operators to meet upcoming national deadlines.
Sources
- Bus Industry Confederation. ADR 68/01 & Seatbelts: Frequently Asked Questions, April 2025. bic.asn.au
- Department of Infrastructure, Transport, Regional Development, Communications and the Arts. Australian Design Rules. infrastructure.gov.au
- Office of Impact Analysis. Harmonisation of the Australian Design Rules — Regulation Impact Statement. oia.pmc.gov.au
- Office of Impact Analysis. Value of statistical life, guidance note. oia.pmc.gov.au
- Office of Impact Analysis. Amendments to Vehicle Standard (Australian Design Rule 4/05 — Seatbelts) 2006 — Regulation Impact Statement, assessed adequate, 2012. oia.pmc.gov.au
- Department of Infrastructure and Transport. Regulation Impact Statement for Amendments to Vehicle Standard (Australian Design Rule 4/04 — Seatbelts) 2006, full text, including Table 1 costs, Table 2 net benefits and Appendix 3 benefit-cost ratios. PDF
- Office of Transport Safety Investigations. Bus safety investigation report: rollover of coach 4666MO, Greta, NSW, 11 June 2023, 2025.
- Park J et al. The effectiveness of seat belt reminders by seat belt wearing rate in buses, paper 23-0190, 27th International Technical Conference on the Enhanced Safety of Vehicles, 2023. PDF
- UNECE. UN Regulation No. 16 — Safety belts, restraint systems, child restraint systems and ISOFIX. PDF
- European Union. Regulation (EU) 2019/2144 on type-approval requirements for motor vehicles. eur-lex.europa.eu
- National Highway Traffic Safety Administration. Federal Motor Vehicle Safety Standards; Occupant Crash Protection, Seat Belt Reminder Systems, Controls and Displays, final rule, 3 January 2025. federalregister.gov
- Department of Infrastructure, Transport, Regional Development, Communications and the Arts. New bus design rules set to reduce fatalities. infrastructure.gov.au
- Assistant Minister for Regional Development. New bus design rules introduced to prevent future tragedy, media release. minister.infrastructure.gov.au
- Bus Industry Confederation. Transmission, September 2024: Revisions to ADR 68. bic.asn.au
- Newcastle Herald. OTSI report: widespread change needed for NSW bus safety. newcastleherald.com.au
- NSW Government. NSW Government acts on initial bus safety recommendations. nsw.gov.au
- AITSC Research Division. Per-seat buckle sensing with a persistent localised audible alert: a systematic review of the passenger car evidence and its application to buses, 14 September 2026.
- AITSC Research Division. Information or consequence? Protocol for a four-arm factorial randomised controlled trial of the ADR 68/01 seatbelt requirements and per-seat buckle sensing on buses, 15 September 2026.
- AITSC. ADR 68/01 Seatbelt Warning System — kit prices, specifications and installation detail. aitsc.au
