Hardly any piece of mountain equipment is judged as differently as the avalanche airbag. Some regard it as the greatest advance since the transceiver, others as an expensive placebo whose main effect is to send people into steeper terrain.
There is a solid study on the question, and its answer sits between the two camps. It comes from Pascal Haegeli and colleagues, was published in 2014, and assessed 424 serious avalanche accidents between 1994 and 2012 across seven countries, Switzerland among them.
The headline figure
It looks at people caught by an avalanche of size 2 or larger, that is, one big enough to bury someone seriously.
The mechanism is not a protective shell but physics. Moving snow sorts itself by particle size, and large items travel upwards. The airbag increases the body’s volume without meaningfully increasing its weight, and so moves it up in that sorting. This shows in the study’s second figure: the risk of critical burial falls from 47 to 20 per cent.
The airbag therefore acts before burial, not inside it. Someone who is not completely buried never enters the survival curve at all.
The catch: one in five does not inflate
Here it gets uncomfortable. The study found a failure rate of 20 per cent. And the reason is usually not the technology:
Show the figures as a table
| Entry | Value |
|---|---|
| Inflates | 80 % |
| User (not deployed) | 12 % |
| Other (device, circumstances) | 8 % |
Sixty per cent of failures come down to nobody pulling the handle. The remainder splits across device faults, torn-off handles, destroyed balloons and cases where the pack was lost during the ride.
This is the real finding of the study, and it has little to do with technology. An airbag is the only safety device in the pack that has to be operated within one or two seconds and under maximum stress. A transceiver is used by somebody who is still standing. The airbag has to be triggered by the person being swept away.
What remains once you price that in
The 22 against 11 per cent applies if the airbag inflates. Factor in the failure rate and the picture for a randomly chosen airbag wearer looks different:
| Share | Mortality | |
|---|---|---|
| Airbag inflates | 80 % | 11 % |
| Airbag does not inflate | 20 % | around 22 % |
| Combined | 100 % | around 13 % |
Instead of a halving from 22 to 11, what remains is a reduction from 22 to around 13 per cent. Still a substantial effect, but markedly less than the headline promises.
And that is the usable conclusion: the biggest lever is not the purchase but practising the pull. Twelve airbags in a hundred fail because nobody deployed them. Those twelve per cent are the only part of the calculation you can influence yourself.
What the airbag cannot do
Three limits that appear in no product description.
It does not help against impact. Between a fifth and a third of avalanche victims die of injuries sustained during the ride, not of asphyxia. The balloon does nothing against a tree or a rock edge, and in terrain with cliffs it does nothing against the fall.
It does not help in terrain traps. If the avalanche runs into a hollow, a stream bed or against the edge of a forest, the snow piles up. Sorting to the top helps little when the top is still several metres of snow.
It replaces nothing. An airbag supplements transceiver, probe and shovel; it does not substitute for them. The 11 per cent who die despite an inflated airbag are partly the very cases where companion rescue then matters after all.
And the question of risk compensation
The commonest objection to airbags is that they tempt people into greater risk. The technical term is risk compensation, and the argument is not far-fetched: it has been studied for seat belts, helmets and anti-lock braking, with mixed results.
Honestly: we found no robust figure on this. The Haegeli study measures effectiveness within the accident, not behaviour beforehand. Whether airbag wearers ski steeper terrain on average cannot be read from it, and we are not presenting a hunch as a finding.
What can be said: if risk compensation exists, it eats into part of the measured benefit, and the measured benefit after deducting the failure rate is already smaller than the striking halving suggests.
What it costs
Swiss specialist retail prices for the 2025/26 winter, as an order of magnitude:
- Cartridge systems from around CHF 560, plus the cartridge and refilling it after every deployment and every practice pull.
- Electronic systems around CHF 1,000. They need no cartridge, can be deployed repeatedly and therefore practised with, and cause no trouble on a flight.
The difference matters little for effectiveness and a great deal for practice. With a cartridge system every practice pull costs money; with an electronic one it does not. Given that twelve airbags in a hundred fail because nobody pulled, that is not a side issue.
What the full emergency kit costs is set out in our breakdown of avalanche equipment.
Sources and caveats
- Haegeli, P. et al. (2014), “The effectiveness of avalanche airbags”, Resuscitation. The analysis covered 424 serious avalanche accidents between 1994 and 2012 in Austria, Canada, France, Norway, Slovakia, Switzerland and the USA in which at least one person caught was wearing an airbag. Adjusted absolute mortality reduction 11 percentage points (from 22 to 11 per cent), confidence interval 4 to 18 points, adjusted risk ratio 0.51. Risk of critical burial 47 per cent without, 20 per cent with an inflated airbag. Non-inflation rate 20 per cent, of which 60 per cent was non-deployment.
- The “around 13 per cent” calculation is our own and assumes an uninflated airbag behaves like no airbag. That is a simplification: an undeployed pack still has some volume.
- Prices: Swiss specialist retail, as of winter 2025/26, entry-level guide figures. They vary considerably by model and discount.
- The study dates from 2014 and is therefore over ten years old. Devices have since become lighter and partly electronic. Whether the failure rate is lower today we do not know; we are not aware of a more recent study on this scale.
This piece assesses a study and is not training. How to operate an airbag in an emergency is learned by practising and on a course, with the SLF and the SAC.
