United airlines part 1: corporate socialists

There is much to be said about United Airlines recent debaucle wherein they publicly dragged a paying customer off an aircraft, injuring him in the process. Most of what needs to be said and written has been/will be better expressed by craftier social pundits than me. There is however one element to this story that I feel I can comment on with some agility…and that I find particularly deplorable:

United Airlines created a market-based system (the ‘overbook and compensate’ protocol) which allowed them to profit from the uncertainties associated with air travel. They were perfectly happy to use this system to profit at the expense of customers when the system favored them…but when the market that THEY CREATED suddenly spawned conditions that should have allowed travels to profit at United’s expense, they quit. They turned against the market and sent in the jack-booted thugs. And that is bullshit!

…but first, an apology

I have what I think is a pretty handy analytical/computational example of the optimal overbooking problem. I think it nicely illustrates some of the things I want to talk about here like pricing risk…but I’m still tweaking it. It would probably make more sense to post that one first…but I’m done with this one now and I wanna post it. sorry.

the analytical/computational model should follow shortly.

Executive Summary

I’ll freely admit that this is a bit of meandering blog post. It takes a few detours. The nub of what I want to say can be condensed down to this:

Yes United Airlines was monumentally stupid in their handling of the situation on flight 3411. But the system we have in place for handling these types of incidents is equally stupid. Specifically, in the event of an overbooking event and the absence of any passenger volunteering to get bumped in exchange for some penny-ante compensation, airlines are actually allowed to select passengers for removal. And in the event of such an occurrence, the maximum compensation the a-grieved passenger can demand is 3 times the face value of his/her ticket. That is fucking nuts!

One of the reasons this bothers me is the following:

Basically, United Airlines operates inside a system which allows them to reap extraordinary profits from risky behavior when the uncertainty goes their way (some passengers on overbooked flights don’t show up) while simultaneously protecting them from the consequences of their risky practices when the uncertainty moves against them.

Let’s suppose I wanted the benefits of such a system. A system that allowed me to reap the benefits of a market when market conditions favored me and get protection from market forces when condition don’t. Suppose I wanted health insurance that provided me all the benefits of market competition (low premiums, expanded choice sets) but I also wanted to be insulated from adverse conditions arising from uncertainties (price caps on important prescription drugs for example). I’d be labeled a socialist, communist, and denegrated as downright un-American by the same people who think corporate entities like United Airlines should be able to privatize market gains in good times and socialize loses in bad times.

The problem with this UA fiasco from an economic - or even environmental economic - point of view is that what the company really wanted in this case was to use risk and uncertainty pricing in order to profit but at the same time they wanted to be insulted from the consequences of pricing risk incorrectly.


I don’t know if I can add any value to this saturated conversation but I’m going to try for a few reasons:

  1. I spend a lot (almost all) of my time studying public policy and regulated industries. This nonsense where firms/industries get to exploit extreme gains from poorly regulated markets and at the same time use market interventions to limit losses is everpresent. Industries, particularly those subject to environmental regulation, are pretty predictable in their distribution of ‘pro-market’ propoganda. But when it comes down to it, in my experience, a decent chunk of these firms aren’t really pro-free market. They’re just pro-market when the resulting allocation of resources allows them to profit at others’ expense…and, like everybody else, they want protection from market forces when those forces don’t favor them.

  2. I think the general practice of overbooking as a strategy for risk management is pretty interesting and the process of writing this rant motivated me to go back and read some oldie-but-goodie scholarly work on choice under uncertainty, bounded rationality, and pricing risk. I’ve tried to incorporate and annotate these citations…though I freely admit their inclusion in this post is a little forced.

Establishing some important facts

  1. Fact #1: Facing a full flight and the need to free up a couple seats for a flight crew that was needed in Louisville that night, United Airlines began offering monetary incentives for passengers to give up their seat and take a different flight. After finding no takers on their offer at 400, they upped their bid to 800 then, inexplicably, stopped.

  2. Fact #2: I’m not a lawyer but, as I understand it, they are actually legally allowed to ‘select’ people for removal and compensate them up to (but no more than mind you) 3 times the face value of the ticket.

  3. Nerds can probably skip this one…For those that don’t speak optimization as a second language: the ‘overbook and compensate’ strategy is a pretty uncontroversial (uncontroversial in terms of being a perfectly sensable reaction the realities of selling air travel for money) profit maximizing strategy. There is some positive probability that not everyone you sell a ticket to will show up for the flight they purchased a seat on. Each empty seat on an airplane costs the airline money (to oversimplify thing just think about the fuel required to get a jet from point A to point B. If there are 30 seats on the plane and all 30 are filled the airline’s fuel cost is distributed over 30 people. If the airline only has 25 passengers it has to use the same amount of fuel but the cost is distrubted over fewer people). Since there is a postive probability that not everyone will show up for the flight, and there a cost to the airline of empty seat, the airline has an incentive to sell more seats than physically exist on the plane. However, if the airline sells 31 seats on a plane that only has 30 seats, AND everyone that bought a ticket does show up, the airline some cost of bribing a ticket-holder to take another flight. The airline chooses how many seats to overbook on a flight by maximizing expected profit which dependeds on i) the probability of individuals showing up for thier flight, ii) the cost to the airline of empty seats, and iii) the cost to the airline of selling too many seats.

Point 1: Capping ‘Bumped Passenger Compensation’ is Total Bullshit

Let’s assume we can all agree that publicly dragging a guy of your airplane was a monumentally stupid thing to do. As many others have pointed out, this action was dumb and deplorable for a host of reasons…the main financial one being that UA is almost sure to pay a huge court settlement (I’m just guessing but it’s probably reasonable to think it will be in the high six figures to millions) when they could have just

  • chartered a private flight for their crew in Chicago that needed to be in Louisville the next day.
  • chartered a private flight for some of the passengers on the plane
  • get an Uber for the guy or the flight crew

I’m putting aside the main issue of how much easier it would have been to do one of the three things above (or almost literally ANYTHING other than what they actually did)…because, as I mentioned above, this meta-issue is better left to someone with more expertise in law, marketing, PR, etc. What I want to focus on is why United Airlines didn’t do the one thing that their practice of overbooking and (occasionally) bribing passengers to take another flight was actually set up to do:

offer more money for somebody to get off the damn plane. (HOLY SHIT!!!)

I don’t want to denigrate the human instinct that say even if they absolutely HAD to remove someone from that plane, they didn’t have a drag a man through the plane like an animal. But that kind of commentary is not my comparative advantage. I do want to comment on a nuanced element of this debacle that unsettles my inner economist and that I don’t think is getting the right amount of outrage:

At the nub of what happened is this: United Airlines was perfectly happy to participate in a market-based system when market conditions favored them but, when market conditions changed ever-so-slightly, they quite the market, threw up their hands, and sent in the jack-booted thugs.

What if I want to buy something I can’t afford?

Individuals get a lot of shit from the ‘free-marketeers’ for wanting special treatment or wanting the market to make allowances for their special circumstances. But what would happen if I tried to do what United Airlines did?

I’m sure there are lots of example that would be relevant here but allow me to start with the one that popped into my head when I first thought about this: Nobody wanted to get off that plane for 800 bucks. When UA oversold that flight, in my mind, they endowed people with a right. They sold a service. They’re own practices and precedent dictated that, when such things occur, UA is obligated to buy back that right. The friction here came from the fact that when UA went to buyback some of the rights they sold, they found that the market had changed dramatically from the market under which they sold the original rights. Shit happens! If I sell my house for 800,000 and then find out that the Real Estate market is so hot that I can’t buy a comparable house for 800,000 I don’t have the right to go into the house I just sold to some other people and kick them out because they refuse to accept my offer to buy my house back for 800k/850k/etc.

At that moment the market isn’t defined by my willingness to pay, it’s defined by the owner’s willingness to accept. What United Airlines thought they should have to pay for a seat on that oversold flight should have been irrelevant. They should have had to pay what the market determined was fair value for a seat on that flight…which apparently was more than 800 bucks.

The play-by-play

Let’s review in somewhat pedantic terms what happened on UA flight 3411:

  • United Airlines observers that a certain percentage of seats sold on flights go unfilled for various reasons
  • United formulates a profit maximizing response to this fact that involves selling more seats for a flight than physically exist on that airplane.
  • When United accepts your payment of $X$ in exchange for an airline ticket what they are doing is tranferring ownership of a seat on an airplane to you. It is important to understand that, at this point in the transaction, United Airlines ultimately determines the market. They may be reacting to conditions of supply and demand present at the moment of purchase…but, in the end, they own the seat and they get to decide if the $X$ you are willing to pay meets their minimum willingness to accept.
  • Implicit in UA’s profit maximizing strategy is the possibility that they will ‘overbook’ the flight.
  • In the case that they overbook, they agree to ‘buyback’ the seat that they ceased to be owners of when they took your money. Here it is important to note that, at this point, you get to (should get to) determine the market.
  • On flight 3411 United found themselves in the position of having to buyback some seats on their aircraft. They reportedly started at 400 clams and found no takers.
  • I don’t know what people paid for their seats on UA flight 3411 but comparable tickets can be found on Kayak for around 300. So, what United Airlines was faced with on Sunday, April 9, was a market that had changed. They probably sold a bunch of seats on that flight for $300 bucks then found out what the rest of us deal with every time we try to buy a Big Mac in an airport or a beer at a baseball game: markets aren’t static.
  • United reportedly upped the ante to 800 bucks and still found no takers. At this point, the point that irks me more than probably anything else in this story, they stopped bidding.
  • In plain language, United Airlines didn’t think they should have to pay more than 800 for a seat on that plane.

In equally plain language: THAT’S BULLSHIT! Here is a short list of people who are routinely called socialists/communist for making the exact same claim that United Airlines implicitly made when they refused to pay more than 800 large to buyback a seat they needed (i.e. “the price is more than I am willing to/able to pay, but the good is something I need really bad so somebody needs to do something!”):

  • Bernie Sanders supporters who believe affordable health care is a right
  • College students protesting the high cost of tuition, and rapid rates of increase in tuition, at universities across the U.S.
  • Affordable housing advocates who want to live in expensive zip codes without paying market rates.

I know this last bit is touch off-topic but felt compelled to include it because I’ve seen a few argument with the flavor of, “Well sure, United handled the situation poorly but…they had every right to remove passengers from the plane because they offered fair compensation according to their contract of carriage and nobody accepted.”

Again I say BULLSHIT! If we are going to insist on defending markets when markets are pricing marginalize people out of important things like prescription drugs…we shouldn’t get too put-out when market conditions occassionally cut into corporate profits.

If you only favor markets when markets favor you then you’re not a free-marketeer, you’re just another asshole trying to rig a game in your favor. That’s fine, I don’t expect anything more or anything less from corporate enterprise. But we all need to be honest and clear about what happened on UA flight 3411:

  • A large corporate entity invented a system, part of which included a secondary market for airline seats, that allowed them to simultaneously profit from uncertainty and sell the same excludable good to multiple individuals.
  • We (society) generally viewed this system without much skepticism because…well, I’m not sure why exactly.
  • United found, in this specific instance, that the system THEY CREATED spawned a spot market they didn’t like and didn’t anticipate.
  • Instead of taking their loses like normal market participants are expected to do when they misjudge market conditions, they shut the game down, punched a guy in the face, and dragged him off an airplane.

And now they are getting royally FUCKED and I’m loving it!

Tangent #1: rational preferences

I’ve seen a few media outlets suggest that ticketed passengers in UA flight 3411 were acting ‘irrationally’ by not accepting 800 bucks cash money and an offer to take a different flight. This line of reasoning seems to be based on the fact that some passengers probably paid less than 400 for their tickets…so the buyback offer was more than 2 times the face value of some peoples’ tickets.

That’s irrelevant. I’m gonna have to be the jargon police here: rationality, as it relates to preferences, only requires completeness and transitivity. The most exhaustive discussion you are likely to find on the axioms of choice viz-a-viz preference relations is in Mas-Collel, Whinston, and Green, a book that every economist I know hated…but is actually pretty sweet when the pressure of an impending qualifying exam is removed from the environment.

For those that enjoy the compact, efficient language of mathematics I also highly recommend Paul Samuelson’s Foundations of Economic Analysis, Chapters 1-3.

Transitivity means that:

Suppose I’m asked to form a preference relation between three beers Green Flash West Coast IPA, El Segundo Brewing Mayberry IPA, and Knee Deep Brewing Citra Pale Ale.

  • If I prefer West Coast IPA to the Mayberry IPA, and
  • I prefer the Mayberry IPA to Citra Pale Ale, then
  • I should prefer the West Coast IPA to Citra Pale Ale

As long as that simple condition is met than my preferences are transitive.

Completeness simply means that I am capable of forming a binary preference relation between each pair in the choice set. Here is the really dope part about completeness: indifference is a totally valid option. That is,

  • I prefer the West Coast IPA to Mayberry
  • I prefer the West Coast to the Citra Pale, and
  • I am indifferent between the Citra Pale Ale and the Mayberry IPA

is a complete and transitive set of preferences.

Putting this into the terms of our airline example: no passengers were willing to give up their seat for 400 dollars. This means all passengers preferred travel from Chicago to Louisville on flight 3411 on April 9th, to 400 dollars. Since the price was upped to 800 bucks without any takers, this means that all passengers also preferred travel from Chicago to Louisville on flight 3411 on April 9th to 800 dollars. Rationality has nothing to do with the price paid for the ticket. As long as:

  1. none of those passengers who turned down 400 and 800 would have accepted 600 then preferences were transitive, and
  2. as long as, theoretically, those passengers were capable of saying ‘yes’ or ‘no’ to buyback amounts between the low offer of 400 and the high offer of 800, then preferences were complete.

As long as preferences were complete and transitive, they were rational.

This is important because, to me at least, one of the great strengths of an economic approach to understanding human behavior is that economists don’t like to make normative judgements about people’s preferences. Whether someone preferes brussels sprouts to strawberries or beets to blueberries is of absolutely no consequence to me: as long as they act on those preferences in some kind of predictable manner then they are a completely rational economic actor, making it possible for me to mathematically represent their behavior.

Tangent #2: Why do beers cost so much at the ballpark?

The answer is both interesting and pretty mundane. In the simplest, most complete terms, a 16 oz Sierra Nevada at AT&T Park in San Francisco costs roughly double what 16 oz of Sierra Nevada purchased at the Safeway right outside the ballpark would because people at the ballpark are willing to pay more.

While it’s true that there is limited competition inside the ballpark, not even monopolists like to lose money. Bud Light costs 50 cents per oz and Sierra Nevada costs 75 cents per oz at AT&T Park for this simple reason that people will pay that much to drink beer and watch a Giant’s game….that’s the mundane part.

Here’s the interesting part:

There is some interesting mental calculus that goes on in order to arrive at the different valuations people can have for the same good in different context. For example, for me, beer and live baseball are so complementary that my willingness to pay for live baseball without beer is pretty close to 0. Stated another way, the cost of beer is ‘baked in’ to the ticket price for me. If the price of the ticket compromises my ability to purchase two beers of a quality no less than Sierra Nevada, than I won’t purchase the ticket.

Now, lemme take a second and tie this back to the United Airlines case: when you are at the ballpark you have selected into a group of people that, on average and in that setting, have a higher willingness to pay for beer than they would in another setting. Vendors at the ballpark know that about you and charge you accordingly for beer. United Airlines, when they oversold their last flight of the day from Chicago to Louisville, put themselves in a position where it was worth a lot to them to get one of those seats back. The individuals on that plane knew they were holding a good that, at that moment and to the prospective buyer, was worth a lot…and they priced the good accordingly. When the ballpark does that to us, we complain to our friends and the other people in line then we:

  • don’t buy beer then complain about how much beer costs at the ballpark
  • buy Bud Light instead of Sierra Nevada then complain about how much beer costs at the ball park
  • buy Sierra Nevada then complain about how much beer costs at the ballpark.

What we DON’T do is offer the concession stand guy a 10-spot for the 12 dollar beer then punch him in the face and start a riot when he says no.

For anyone interested in a shallow dive into some of the ways economists have extended classical micro-economic theory to address complex choice occassions here are some suggestions:

Simonson and Tversky, 1992, Choice in Context. Whether is is beers at a ballpark or seats on airplane, context matters. If I prefer Coke to Pepsi it is tempting to think that when faced with a choice between a 12 oz can of Pepsi for 1 dollar and a 60 oz bottle of Coke for 5 dollars, I should prefer the Coke because they are the same unit price. But it’s pretty easy to imagine some circumstances under which I would chose the Pepsi (presumably because I just don’t want 60 oz of soda period). Context matters.

Richard Thaler, 1985, Mental Accounting and Consumer Choice. Richard Thaler, Amos Tversky and Daniel Kahneman, are 3 pretty big names in modern micro-economics. They’ve all made substantial contributions to the integration of behavioral psychology into economic models of choice. You could probably Google-Scholar any of those names and read the first thing that comes up and find yourself into something pretty interesting. I like this paper because I think it relates pretty directly to how an individual in possession of airplane seat might price that good for a buyback auction. From Thaler:

Consider the following anecdotes: 1. Mr. and Mrs. L and Mr. and Mrs. H went on a fishing trip in the northwest and caught some salmon. They packed the fish and sent it home on an airline, but the fish were lost in transit. They received 300 from the airline. The couples take the money, go out to dinner and spent 225. They had never spent that much at a restaurant before…

All organizations, from General Motors down to single person households, have explicit and/or implicit accounting systems. The account systems often influence decisions in unexpected ways.

Stigler and Becker, 1977, De Gustibus Non Disputandum Est. I like this paper because it’s a little bit of a defence of classical microeconomic theory. The theory of Bounded Rationality has certainly revolutionized how we think about complicated choice processes but, in my opinion, the recent explosion of behavioral psychology/experimental economics has led some people to underestimate the power of classical micro-economic theory. Here Stigler and Becker demonstrate how phenomenon (persistent drug use/addiction, susceptibility to advertising) once thought to violate important assumptions of micro-economic models can actually be explained quite robustly using the generalized calculus of utility maximizing behavior.

Dhar and Simonson, 1999 have a paper that is a little thin on theory and empirics but does present a lot of nice examples of complementary choice in ‘consumption episodes.’ In particular, one way to think about how people like me make choices at the ballpark is by thinking about how we ‘balance’: If I want to buy a ticket for a the ballgame, generally I can buy:

  • A really good seat and 2 cheap beers
  • A pretty good seat and 1 expensive beer and 1 cheap beer
  • A mediocre seat and 2 good beers

Point #2: This isn’t an isolated incident

I don’t have a ton of meat to offer here but I do want to throw out a few examples of other firms/industries that have successfully conned us into creating, or lobbied hard for, systems where they get all the rewards of a market but are insulted from the adverse market conditions:

‘Too Big to Fail’ Financial Institutions.

Our bank bailout was basically indistinguishable from the system where United Airlines gets to cap bumped passenger compensation. Financial institutions in the years leading up to 2007 made ass-piles of money placing risky bets on assets with really poor financial fundamentals. When mortgages started to default and mortgage backed securities started to crash the market consequences of incorrectly pricing risk fell squarely on the heads of some of the largest banks in the U.S…..those consequences were then promptly distributed out to taxpayers in the form of a ‘bailout.’

California Ag

In California, for a surprisingly long time, agricultural firms were not required to pay farm-workers the minimum wage. And I’m not talking about paying undocumented workers cash-under-the-table for work that ‘on paper’ doesn’t exist. It was perfectly legal to pay farm-workers a wage below the state minimum wage. There was acramonious debate when this regulation was changed. The California Farm Bureau fought really hard to keep agricultures exemption from minimum wage laws.

This wouldn’t sit so poorly with me except for the fact that California Farmers had no trouble accepting a 2 billion dollar gift from the government in the form of the Central Valley Project (the CVP cost around 3 billion and irrigation districts served by the project were expected to pay back 1 billion within 50 years of construction…in the form of a principal-only loan of course).

I know this seems like a reach but I don’t really think it is. In arguing against imposing minimum wage compliance on farms, the California Farm Bureau made the case that farmers shouldn’t have to bear the cost of artificial distortions to the labor market…that the agricultural labor market should be unregulated and the free-market should be allowed to efficiently allocate resources. Ok. BUT…in accepting a massive government handout to build irrigation infrastructure the exact opposite case was made. Specifically, if forced to participate in capital markets as they were in 1933, California Farmers would never have been able to afford the critical infrastructure that delivers water (their primary productive input) to the farms. So on the one hand they wanted to be free to profit from an unregulated/poorly regulated labor market but, on the other hand, they wanted protection and relief from a capital market that put necessary infrastructure totally out of reach. In short, they wanted exactly what United Airlines wants: to participate and profit from the parts of a market that favor them while reaping the benefits of intervention in the parts of the market that don’t.

Side Note: I know it probably feels like I beat-up on California Ag a lot. I’m not at all antagonistic to CA agriculture and most of the irrigation districts I work with are centered, balanced business people. However, thier lobbiest have a habit of picking fights over things like minimum wage compliance and overtime pay that I find hypocritical.

Written on April 19, 2017