The answer is that the average cost, and the marginal cost, of an airline ticket are very far apart.

The same applies to hotel rooms, by the way, but hotels do seem to have different market dynamics, probably because a hotel is cheaper in the first place.

The cheapest NYC - LAX round trip in May, with a five-night stay, via ITA is $329. (It is EWR - CLT - LAX on AA).

Virgin America and Jet Blue each charge $1,600 or so fully flexible; AA is $3k.

I don't have the figures but I suspect that the mean cost/passenger of operating that flight is going to be around $1k, so a return trip on an A321 will cost AA in the region of $200k. Clearly AA can't sell 600 seats at $300; nor can they sell all 200 seats at $3k.

However they might be able to sell 30 seats at $3k and 190 seats at prices ranging between $300 and $1,000 - perhaps an average of $650. With 10% oversold and no-shows they might have to refund an average of 3 x $3k and 19 x $200 (assuming that some of the less flexible fares aren't completely inflexible).

This gives them total revenue for the flight of $90k + $123,500 less $12,800 for refunds = $200,700 and therefore wafer-thin profit margins.

If everyone were purchasing the same fare class, 30 people who need complete flexibility would be much, much better off, but 190 people who are currently paying an average of $650 per ticket would need to pay 50% more even if airlines continued to set prices in response to demand, as hotels do.

So if you only ever travel on the cheapest possible fare, based on the example above, your £470 BHX-BKK would be £600+ every single time. To put it another way, you'd be paying for that change fee once every three flights even though, in my example, only 10% of passengers no-show.