Compulsory bicycle helmets: the measure that costs the state nothing

Two bills, and one confusion

On 9 June 2026, thirty-seven MPs tabled bill no. 2900 in France’s National Assembly. It makes an approved helmet and a retroreflective vest compulsory, by day as by night, in built-up areas as outside them. It requires the road safety certificate (brevet de sécurité routière). It raises the fine to €270, provides for confiscation of the machine and up to €15,000 in cases of derestriction.

None of its provisions targets bicycles. Its official title is “bill to strengthen the regulation of electric scooter use”, and all its obligations bear on motorised personal mobility devices — a legal category covering electric scooters, self-balancing boards and electric unicycles, and one that its article 5 explicitly excludes from active mobility. An electric bike limited to 25 km/h is not part of it: France’s highway code classes it among cycles.

The bill that really targets cyclists is a different one, older and far shorter. Tabled on 16 September 2025 by MP Yannick Favennec-Bécot and twenty-seven co-signatories, bill no. 1810 runs to two articles. The first creates an article L. 311-1-1 of the highway code making the helmet compulsory “for riders of cycles, of motorised personal mobility devices, and of any two-wheeled land vehicle that is unpowered or lightly powered”. The second provides for entry into force six months after promulgation.

That is all. The rest — “the technical characteristics of helmets, the categories of users concerned and the penalties incurred” — is referred by the text itself to a decree issued after review by the Conseil d’État, France’s highest administrative court. In practice, Parliament would vote a principle whose scope and penalty would be set afterwards, without it.

What that principle rests on is another matter. The explanatory memorandum of bill no. 1810 puts forward a single figure, which it attributes to France’s interministerial road safety observatory: “more than 600 users of motorised personal mobility devices (EDPM) were seriously injured or killed in 2023”. The figure is real, but it covers the category of motorised devices — the very one from which the highway code separates cycles.

One bill, one report, one withdrawal

This is not the first time. On 8 July 2021, senator François Bonneau and several colleagues tabled in the Senate a bill on the wearing of bicycle helmets. The law committee took it up seriously: rapporteur Jérôme Durain heard France’s road safety directorate, biomechanics researchers, user associations and doctors. His report, filed on 5 January 2022, works through the whole scientific file.

The bill reached the floor of the chamber on 13 January 2022. It was withdrawn by its authors the same day.

Four years later, the two bills tabled in the National Assembly have got no further. As of 27 August 2026, the legislative file for bill no. 1810 carries just two entries, its tabling and its referral to the law committee, with no rapporteur appointed and no date of examination. That of bill no. 2900, tabled in June, is at the same stage.

The parliamentary route is going nowhere. The citizens’ route, for its part, has collapsed. Filed on 16 July 2025 on the National Assembly’s petitions platform, petition no. 3027 asked for exactly what the bills ask for: a compulsory helmet for every bicycle and scooter user, high-visibility clothing, and a fine — “symbolic”, its author noted — of €11. It gathered 256 signatures in total. The law committee closed it as a matter of course on 22 January 2026, under the rule that sets aside any petition that has not gathered five thousand signatures within the six months following its filing.

Two hundred and fifty-six signatures. Three months later, in April 2026, pollster Ifop surveyed one thousand people aged eighteen and over on behalf of a commercial comparison site for electric bikes. The report, published on 22 May, concludes that 84 % of French people are in favour of compulsory helmet use “when using a standard bicycle”, and 88 % “when using a bicycle with electric assistance”.

The two figures do not contradict each other: approving a measure when it is put to you and going out to demand it are two different acts, and no pollster claims otherwise. But the gap is worth noting, because it is regularly invoked in one direction only. A majority opinion in a questionnaire is not a social demand.

What the helmet does, and where it stops

On the effectiveness of the helmet itself there is no serious controversy, and this article will not pretend otherwise.

Heard by the Senate rapporteur in 2022, France’s road safety directorate put it this way: “there is a scientific consensus that the helmet is effective if it is well designed, well fitted and correctly worn. On that basis it protects against skull fracture and head trauma. However, cyclists must remain careful, because at higher speed (above 30 km/h) the helmet will no longer protect as effectively”.

The reference meta-analysis points the same way. Published in 2017 in the International Journal of Epidemiology by Jake Olivier and Prudence Creighton, of the University of New South Wales, it aggregates 40 international studies covering 64,708 cyclists and concludes that helmet use is associated with a reduction of 51% in head injuries, 69% in serious head injuries, 65% in fatal head injuries and 33% in facial injuries.

Biomechanics pins down where the limit lies. The TEVU study, carried out in 2017 at the ICube laboratory of the University of Strasbourg by Caroline Deck and Rémy Willinger for France’s road safety authority, modelled several crash configurations. When a car door is opened without warning, the helmet halves the risk of skull fracture and divides by twenty the risk of mild head injury. In a collision with a vehicle travelling at 45 km/h, it divides the risk of skull fracture by three.

Above 45 km/h, still according to that study, this risk reaches 100% — whether or not the cyclist is wearing a helmet.

This threshold is rarely mentioned, and it says the essential of what a bicycle helmet can do. The same work shows a considerable benefit in fall and car-door configurations, and a benefit that fades away in impacts at road speed. A bicycle helmet works very well against the ground, the kerb and the car door; against a mass travelling fast, the protection measured by TEVU comes to nothing.

And when the Senate rapporteur asked Rémy Willinger how many deaths the helmet made it possible to avoid, the answer recorded in his report is unambiguous: no scientific study was then able to put a precise figure on the number of deaths or injuries avoided through helmet use.

Thirty-five years, and Australia still has not settled it

The argument most often set against the obligation does not bear on the effectiveness of the helmet. It bears on the effectiveness of the law: an obligation would push cycling back, and that decline would cost more lives — through inactivity, and through the loss of the protective effect of numbers — than it would save.

Australia is the laboratory for this question. Victoria made the helmet compulsory in 1990 and, according to the University of New South Wales, the other states and territories had followed by 1992. Thirty-five years on, the published work still contradicts itself.

On one side, the work of Jake Olivier, at the University of New South Wales, for whom “there was an immediate 46% reduction in the rate of cyclist deaths per 100,000 population after the introduction of bicycle helmet legislation in Australia” — a decline sustained since 1990, which he puts at 1,332 deaths avoided (translated from the French). His colleague Raphael Grzebieta goes further, saying of opponents of the law that “they are no different from climate change deniers and anti-vaccination groups” (translated from the French) — a comparison this article reports without endorsing.

On the other, the counts by Dorothy Robinson, whose 1996 study notes that in Melbourne, between 1990 and 1992, the rise in the number of helmeted cyclists was far smaller than the fall in the number of cyclists counted. Her reading: the main effect of the law was not to put helmets on cyclists, but to make them disappear.

These two camps publish in peer-reviewed journals, read each other, refute each other, and do not converge. It would be dishonest to present this disagreement as settled one way or the other.

One source, however, makes it possible to measure what is really known: the Cochrane systematic review devoted not to the helmet, but to the laws that impose it. Signed by Alison Macpherson and Anneliese Spinks, it retains six controlled before-and-after studies, and specifies that in each of them the legislation assessed applied only to children and that none of them measured cycling levels. Its authors conclude that legislation appears effective in increasing helmet wearing and reducing head injuries in the populations concerned, but that the evidence is insufficient to support or refute the hypothesis of a decline in cycling.

In other words, according to that reference review: no assessment of an obligation imposed on adults, and no measurement of the effect it is reproached with. Quebec’s national public health institute, which in 2018 synthesised twelve studies conducted in four countries, arrives at the same caution — such a law “may have a negative effect on cycling, particularly among young people, but it is difficult to conclude definitively that such an effect exists, because of significant methodological limitations”.

The same Quebec opinion, on the other hand, sets aside three arguments often advanced against the helmet: neither risk compensation by the helmeted cyclist, nor closer overtaking by drivers, nor neck injuries are, in its view, scientifically established. The case against the obligation is thinner than is claimed; the case for it is just as thin.

Who dies on a bicycle in France

While the debate goes on, the observatory counts. And what it counts does not look like what is being discussed.

In 2025, according to the road casualty report published by the ONISR on 29 May 2026, 235 cyclists were killed on French roads, eleven more than in 2024, and 2,800 were seriously injured, up 9.2%. Set against 2019, cyclist deaths are up 26%.

Three findings in that report deserve to be read together.

First, according to the ONISR, half of the cyclists killed in 2025 were 65 or older. Public debate talks about scooters, urban youth and delivery riders; one cyclist death in two is someone over sixty-five.

Next, still according to that report, the number of cyclists killed in built-up areas fell by 12% in 2025. The rise therefore comes from roads outside built-up areas — those where speeds are high, and where crash configurations are furthest from the ones in which the Strasbourg biomechanics researchers measure a benefit from the helmet.

Finally, the same report notes that 41% of the cyclists killed die in a crash with no other vehicle involved, and that 62% of the cyclists presumed at fault were alone in their crash. These are falls — and that is precisely the scenario in which the helmet works best. A supporter of the obligation finds their best argument here, and it is a legitimate one.

France, though, already has a full-scale test. Decree no. 2016-1800 of 22 December 2016 made the helmet compulsory for children under twelve, riders and passengers alike, from 22 March 2017. Nine years on, no public assessment of that obligation could be found: no implementation review, no measurement of its effect on cycling among children, no estimate of the head injuries avoided. Before extending the measure to everyone, it would remain to publish what it has produced on the only group to which France has already applied it.

A measure “with no direct budgetary impact”

Back to bill no. 1810, and to two sentences of its explanatory memorandum:

It is written as a virtue. It is also, exactly, a description of the problem.

For there is one point on which researchers in both Australian camps agree, including those most firmly in favour of the obligation. Jake Olivier, the very man who measured the 46% fall, writes that “this senseless focus on helmet legislation distracts attention from more important concerns: building separated cycling infrastructure, educating all road users, and cyclist-protective legislation such as minimum passing distances” (translated from the French). His colleague Raphael Grzebieta adds that it is well established that the leading reasons for not cycling in cities, in Australia, are the lack of infrastructure and the danger posed by motor vehicles.

The 2022 Senate report says the same thing in another language: helmet wearing is described there as one strand among others of a broader policy, alongside protected cycle lanes, 30 km/h zones and an increase in the number of cyclists — that “safety in numbers” effect identified by Peter Jacobsen in 2003.

Separated cycle lanes, for their part, create a new burden for the state and for local authorities. They have a direct budgetary impact. They take years, and they arbitrate the use of road space. The helmet does not weigh on the public budget: it weighs on whoever is pedalling.

That is where the imbalance sits, and it is not in the helmet. A helmet protects, the measurement is established, and nothing in this article says otherwise. It is in the choice of a measure which, among all those the literature associates with cyclist safety, is the only one that calls for neither a budget line nor an arbitration over road space.

Since 2021, three bills on the compulsory helmet have been tabled in Parliament, one in the Senate and two in the National Assembly. None has been put to a vote. The first was withdrawn by its authors, the other two are asleep in committee, and the petition carrying the same demand was closed with 256 signatures.

Five years of bills on the helmet, and not a single vote.