There’s a problem with crash diets: After shedding a few pounds, you indulge yourself and put them straight back on again.

Qantas Airways Ltd. risks learning that lesson with its ambitious plan to create a world-spanning network of ultra-long-haul routes. The new planes making such journeys possible are certainly more fuel-efficient than their predecessors – but if you use them to fly farther than previously thought possible, you risk squandering the benefit.

Fuel costs rose A$416 million ($296 million) in the first half, driving the carrier’s key measure of underlying profit before tax to its biggest decline in four-and-a-half years, according to results announced Tuesday. The pain was particularly concentrated on longer overseas routes, which accounted for just over half of the increased kerosene spending.

This wasn’t what was meant to happen. In the years since jet fuel prices reset lower in 2014, Qantas has been exchanging its fleet of aging and gas-guzzling Boeing Co. 747-400s for 787-9 Dreamliners, arguably the most efficient passenger jets in the skies. That’s a good idea, because fuel is the biggest single cost for most airlines, accounting for between 20 percent and 30 percent of operating expenses.

And yet Qantas’s fuel costs, even when adjusted for the distance being flown, seem to be on the rise. Per available seat kilometer – a measure of seats, multiplied by distance, otherwise known as ASKs – fuel spending is now up at about 1.85 cents, from 1.48 cents two years ago. That doesn’t appear to be purely a function of the higher cost of kerosene: The number of ASKs per barrel of fuel, after rising quite consistently in 2016 and 2017, fell sharply in the last six months of 2018.

How is an airline that’s buying more efficient jets failing to get more efficient performance?

One factor is that not all ASKs are created equal. Passengers and luggage typically make up only 20 percent to 25 percent of the weight of a fully laden long-haul jet at takeoff, with the plane itself accounting for 50 percent of the total and fuel making up the remainder. As a result, taking out 20 economy berths and putting four business-class seats in their place may result in better profit, but will reduce ASKs and barely make a difference to fuel consumption.

The bigger problem comes when you start thinking of all the things you can do with your new fuel-efficient jets. If you compare a 787-9 to a 747-400 flying from Sydney to Hong Kong, the savings should be as dramatic as billed on Boeing’s marketing materials.

It’s more complicated when you start trying to do new things, like the Perth-to-London direct flights that Qantas has been operating since last year and the Sydney-to-London ones it wants to start in 2022. Such ultra-long-haul trips are attractive to high-margin business-class passengers and leave Qantas less at the mercy of hub airports such as Dubai, Singapore and Hong Kong.

At the same time, by not stopping to refill en route, the planes risk burning an awful lot of kerosene just to carry around the extra fuel they need for what was once the second leg of the journey.

This isn’t a new phenomenon. The economist William Stanley Jevons noticed in the 19th century that British coal consumption increased after James Watt invented a more efficient steam engine, because lower-cost power ended up being used more widely than ever before. Efficiency gains often tend to be wasted by less efficient usage habits.

That’s all the more reason for Qantas to consider where it makes its money. By our estimates, each ASK on its domestic mainline carrier generates about 18.66 Australian cents of revenue and 2.62 cents of underlying Ebit. All those business-class and full-service seats on international flights barely do better than the discount Jetstar Airways Pty. division on that revenue measure, and perform substantially worse when it comes to earnings.

Despite this, it’s international travel where Qantas has been putting on extra capacity, while domestic mainline flights are being throttled back. Perhaps this increase is simply what’s needed to maintain its competitive position, though that logic has caused problems in the past: Losses on overseas routes came close to destroying the carrier in the early part of this decade.

The delivery of more kerosene-sipping jets into a now-vanishing era of low fuel prices has resulted in some banner years for Qantas shareholders, but that still came on the back of swingeing cost cuts. Chief Executive Officer Alan Joyce should be careful not to forget that. What efficiency gives with one hand, it can take away with the other.