Friday, October 28, 2016

Reverse Engineering Airline Costs from Fares

I thought it would be good fun to do a scatter plot of flights with how much they cost vs how long they are. I did them in two groups of international and US domestic flights (about 20 each), getting the best fares 3 weeks out with Google Flights, and then got the linear equation that described them.

What it showed:

The Y-intercept for US flights was $92; that is, it costs you $92 to get on the plane, and then $32 for every hour you fly. So flying from one of the Norcal cities to San Diego (1.5 hrs) should be about $140. Checks out.

That $92 would be for maintenance and all the other fixed costs. Now, not including the TTD (through-the-door) cost, at a domestic rate of $32 for every hour you fly, you can figure out how much that balances against fuel costs. At a jet fuel cost of $5.02/gallon at the time of writing, and a 737 mid-air burn rate of 750 gallons per hour, a 737-MAX (200 passengers) that's less than two-thirds full will lose money just considering fuel costs alone. Including the TTD, they have to be only 15% full before it drops below that. Again, that's just fuel costs. Pilots and flight crew eat up money and that $92 might be better thought of as gate-time costs. Of course not everyone will get the good fare that I was using for my comparison, but the average fare certainly won't be double what I got here.

Interestingly, international flights showed a negative Y-intercept of -$12 - which initially I thought could represent a subsidy, but is essentially zero. Plus, then I noticed that the per-hour cost of flying is almost four times as expensive as the domestic numbers I had above. Of course long international flights aren't using 737s, and fuel is probably taxed more than in the U.S. but does this really work out to a factor of four? More than likely this is because most countries' airlines are heavily subsidized (and/or socialized), and even beyond that international route scheduling changes only very slowly and so often doesn't represent what the market would actually bear. For this reason, when I was traveling to Vancouver a lot I found it was worth both my time and money to fly to Seattle and drive, regardless of whether I was looking at American or Canadian airline prices.

No comments: