The High Cost of Zero Marginal Cost Pricing: Evidence from American Airlines

The LA Times reports a very funny story. In the late 1980s, American Airlines sold lifetime first class unlimited tickets for $350,000 each and often sold a "companion pass" for an extra $150,000. Once you own one of these "golden tickets", you can book two first class tickets for free anywhere you want to go.   You don't have to be Jim Heckman to anticipate that this incentive regime will induce a selection effect (a non-random subset of people signing up) and then induce some extreme treatment effects (as these high demanders now face a zero marginal cost of flying).

In a diverse world, what types of people would give American Airlines $500,000 in upfront cash in return for unlimited future plane rides?   From a selection point of view, these will be wealthy people who believe that they have a very high demand for fancy flying.

Once you purchase one of these gold tickets, and now face a zero marginal cost per ride,  econ 101 would predict that you will demand lots of rides.   This is the "treatment effect".  The article offers one example of a Mr. Rothstein who made 3,009 reservations in less than 4 years and canceled 2,523 of them.  These guys would book several first class tickets and hold them in case of weather delays or other unexpected shocks to their schedule.  They would also sell of their "companion ticket" and pocket the money while some of the gold ticket owners would make new friends and allow them to fly for free in first class.  American Airlines was surprised by the amount of fraud and "over use" that their incentive system introduced.

American Airlines should have called a micro economist and asked for a prediction about the unintended consequences of zero marginal cost pricing before it implemented this policy!