As Alaska Airlines Flight 1282 neared 16,000 feet, a boom thundered through the plane as a chunk of the plane’s bodywork was ripped away. The rush of air as the cabin depressurized pulled mobile phones out of hands—an iPhone was found intact on the ground below—and a teenager sitting close to the door was left shirtless. “His shirt got sucked off of his body when the panel blew out because of the pressure, and it was his seatbelt that kept him in his seat and saved his life,” one passenger told the Associated Press.

The consequences of being unbelted at such a moment are unthinkable, but there is one group of passengers who are not required to wear a seatbelt during takeoff in the US: babies under the age of two.

Infant safety is a weird quirk of aviation regulation. In the US, the Federal Aviation Authority (FAA) doesn’t require infants under the age of two to have their own seat, allowing them to be held on a parent’s lap without a belt. Aviation authorities in Canada and Japan follow the FAA’s lead, but the European Union Aviation Safety Agency (EASA) and the UK’s Civil Aviation Authority (CAA) take a different approach, requiring lap-held infants to be secured with an additional seatbelt that loops around their belly and links to the parent during takeoff and landing. That may sound safer, but those same belts are banned in the US and Canada on the grounds that they are riskier in a crash.

Though the FAA doesn’t require seats for infants, it does recommend them. “The safest place for a child under age two is an approved child-restraint system or device, not an adult’s lap,” a spokesperson for the FAA, Mina Kaji, said in an email. Last weekend’s Alaska Airlines incident reignites arguments over whether this policy needs to change.

Why isn’t seating for infants a requirement? Because it’s a bit of a faff and expensive—the costs are judged to outweigh the benefits. “This is a bit of a practical compromise—a separate seat would also need a child seat to be carried, and that should be approved for use on an aircraft,” says Graham Braithwaite, professor of safety and accident investigation at Cranfield University in the UK.

For the FAA, there’s a very specific reason. “For every rule, the FAA is legally required to conduct a cost-benefit analysis to show that the benefits of the proposal exceed the costs,” said Kaji, the FAA spokesperson. The costs of mandating that under-twos be placed in child restraint systems (CRS)—industry lingo for an approved car seat that can be added to a regular seat—aren’t just financials, but also, somewhat unexpectedly, road deaths: The FAA’s own modeling and academic research suggest that, in the US at least, requiring families to buy a ticket for infants two and under would push some to drive instead of flying, leading to an estimated increase in total transportation deaths by 72 over a decade. In comparison, requiring child-appropriate seating for under-twos would have prevented three child deaths between 1979 to 2010, a report from the US National Transport Safety Board (NTSB) notes.

As terrible as that tradeoff is, it’s worth noting that the risk of death on a commercial airliner to anyone, lap-held infants included, remains vanishingly low—so let the pangs of guilt subside if you’ve flown holding your baby on your lap. “Commercial aircraft accidents are still extremely rare, and the logistics of having infants in specific seats, with the installation of a CRS for every flight, may outweigh the safety aspects,” notes Sarah Barry, deputy head of the School of Aviation and Security at Buckinghamshire New University in the UK.

That said, the FAA’s cost-benefit analysis hasn’t convinced everyone. Last year, the Association of Flight Attendants–CWA union called for a change to the rule, and a requirement for a seat for all passengers, as they have for the past three decades.

That campaign was sparked in part by the crash of United Airlines Flight 232 in 1989, in which 112 passengers were killed out of the 296 on board. The aircraft’s systems cut out midair, and with a crash landing imminent, flight attendants told parents of lap-held babies to place them on the ground between their feet, surrounded by blankets, and hold them down as best they could. Of the four lap-held infants on the flight, three suffered injuries, and one—22-month-old Evan Tsao—died of smoke inhalation after slipping into the rear of the craft.

The following year, the NTSB added seats for infants to the FAA’s list of most wanted safety improvements, but the request was removed in 2006 after the FAA’s own modeling showed that buying that extra ticket would motivate 20 percent of families to drive rather than fly—in particular, those with the tightest budgets—and in turn lead to an increase in road deaths.

That belief was backed up by academic research in 2002 that showed the policy change would cause a small net increase in deaths, even if as few as 5 percent of families opted to hit the road instead. Plus, the study found the cost of regulation per death avoided would be about $1.3 billion at the time, deemed by the researchers “a poor use of societal resources.”

The researchers admitted there were limitations to the paper, as they didn’t consider what would happen if seats were free of charge, or if airlines simply gave empty, unsold seats to parents by shifting seating arrangements. (It also ignores the fact that ticket prices often increase for other reasons, and that the cost-benefit balance may shift in other countries where driving is not an option—trains are slightly less safe than planes, but nowhere near as dangerous as highway driving.)

Of course, there’s more to inflight safety than deaths, notably injuries through turbulence. In the US, no one has died from turbulence on a commercial airliner since 2009, though there have been 146 serious injuries—think broken bones, burns, or organ damage—of which the vast majority were suffered by crew members rather than passengers.

However, research from the University of Reading in the UK last year showed that flights were getting bumpier due to climate change, with severe turbulence across the North Atlantic increasing by 55 percent since 1979. Though severe turbulence is experienced for only a small portion of overall flying time, and is spotted in less than 0.1 percent of the atmosphere, that could still lead to more injuries.

There are risks beyond turbulence, though most are more mundane than an entire chunk of aircraft hull falling off or other passengers deciding it’s time to crank open the door midflight, as happened on a flight last year. Indeed, research from 2019 into the number of in-flight medical events impacting children suggested that unrestrained children on laps are at risk not only from turbulence, but also from simply falling out of the seat.

The researchers suggest the most common accidents could be reduced by banning younger passengers from aisle seats in favor of window ones, which would help avoid burns from hot beverages passed over them, bumped limbs from passing carts, and injured heads from luggage falling out of overhead lockers—while also helping children avoid restlessness with the view.

Beyond window seats for all children, what can be done? The lap belts favored in Europe may help infants from slipping out of a dozing parent’s grasp, but research from RMIT University in Australia in 2015 showed downsides. In case of a crash or in-air incident, unrestrained lap-held babies were likely to be projected into the cabin but restrained infants instead suffered serious stomach injuries from the pressure of the belt—and both were at risk of head injury from the seat in front of them or the parent holding them. A vest that clips a baby to a parent’s seatbelt has been approved by the FAA for cruising altitudes, but it doesn’t solve that latter problem.

Changing the design of seats to better accommodate infants or developing better ways to hold on to them in case of a crash are possible solutions, but new technologies need time for development and approval—and that happens slowly in aviation. “Airlines are always looking at evolving designs for comfort and safety,” says Emre Eroktem, the head of the aviation management course at City, University of London. “But regulations take more time to change, as these are developed over long periods of time and after tests, and in reaction to incidents and unfortunate accidents.”

Handily, we have an existing solution: Put infants under two in their own seat on the plane, restrained by an approved car seat—a solution that the FAA endorses. Regulators unwilling to force the issue could encourage this safety-first behavior among passengers by pressuring the airline industry to offer heavy discounts on seats for younger children.

But beyond seats for all, the best solution may be reducing accidents and turbulence, with databases and automatic tracking helping to gather the necessary data to make commercial aviation even safer, rather than requiring safety equipment. “In other words, the risk is managed by preventing the accident, and the compromise is so that there is not excessive cost or discomfort for passengers,” says Braithwaite of Cranfield.

But what, as a parent or carer of an infant under two, are you supposed to do with this 30-year-old debate? If you can afford it, take the advice of aviation authorities like the FAA and buy your child a seat, and bring an approved car seat to insert into it. If your budget can’t stretch to an additional ticket for your child, bring a car seat to the airport with you anyway (or, in the US, invest in an approved harness strap for children over 20 pounds or 10 kilograms), and ask at check-in whether there’s a spare seat you could have for free.

And remember that we hear about midair mishaps like the Alaska Airlines blowout for a reason: They’re rare. Try to have a nice flight.