Points, profits, and packed planes, with Gary Leff

This week I'm joined by Gary Leff, an expert in airline mileage economics. We talked about genesis of airline and hotel loyalty programs, the substantial overlap with U.S. credit cards rewards programs, why these are some of the largest contracts in capitalism, and some quirks of the world which cause particular planes to fly almost empty while others are consistently overbooked.

Sponsor:

Vanta automates security compliance and builds trust, helping companies streamline ISO, SOC 2, and AI framework certifications. Learn more at https://vanta.com/complex

Timestamps

(00:00) Intro
(01:10) History of airline loyalty programs
(04:36) The economics of airline loyalty programs
(08:01) Parallels and differences with hotel loyalty programs
(14:01) Tax implications and legal considerations
(19:50) Sponsor: Vanta
(21:08) Tax implications and legal considerations (part 2)
(34:02) Credit card reward programs
(46:56) The David Dao incident and its impact on airline policies
(47:29) Airline compensation strategies and Delta’s unique approach
(49:08) The economics of airline miles and accounting practices
(53:15) Airline debt and the role of frequent flyer programs
(55:21) The impact of airline policies on customer loyalty
(01:08:11) The value of partner miles and airline alliances
(01:13:11) Creative ways to earn miles and exploit promotions
(01:20:05) The infamous coin promotion and other mileage hacks
(01:30:39) Wrap

Transcript

Patrick: Hideho everybody. My name is Patrick McKenzie, better known as Patio11 on the Internet. And I'm here with Gary Leff, who is an expert in airline economics, specifically with regards to airline mileage programs.

Gary: Well, hello Patrick. It's great to be here with you.

Patrick: Great to be here with you as well.

So we're going to do a bit of a deep dive into the wonderful wild world of airline mileage and loyalty programs. But before we do, credit cards are a big part of this story, and as many listeners of the podcast know, I previously worked for Stripe, which does an awful lot of business in credit cards.

Credit card issuers are a different part of the industry than credit card processors. But because the issuers get very, very persnickety about this particular thing, for reasons we will discuss, it's important for me to say that nothing that I say is based on non-public information. And it comes from just my perspective as an industry observer. And it's not necessarily representative of the views of any other party.

That disclaimer out of the way: let's chat.

History of airline loyalty program

I have a particular point of view on why airlines and hotels evolved these loyalty programs while most corporations didn't. But I would love to hear your point of view on that first.

Gary: Yeah, so first of all, you step even a little bit further back. It would've been the case that an airline frequent flyer program would've been illegal prior to 1978 and the Airline Deregulation Act.

People often misunderstand what airline regulation was all about. It wasn't this pro-consumer protection sort of idea. It was meant to prevent competition in the industry and to protect airline profitability. The government decided who flew where and what prices they charged. And there was more or less a commodity product.

There was even discussion at one Civil Aeronautics Board meeting over regulating the thickness of sandwiches on board aircraft because when prices are set above a market clearing level, they're making abnormal profits and they want to compete for customers. They can't compete on price, they have to compete on other things. And so we have this idea that back in the day, service was so much better. The food was so much better. It was like that for a reason. They had high margins and then wanted to compete for them.

After deregulation - and the reason that airline frequent flyer programs would've been illegal then is that the Supreme Court has said that frequent flyer programs are effectively rebates on the price of a ticket. And that is even true when you're earning the miles for something other than flying. And even if you're spending the miles for something other than flying, the program itself is basically a rebate on a ticket. And you couldn't provide customers with rebates because that would've been a workaround for their price regulation.

Shortly after deregulation, we started to see frequent flyer programs of one kind or another. There were rebate programs of various kinds, even before deregulation. Southwest Airlines started as a regulatory arbitrage. They were an intra-Texas airline. Because they only flew within the state of Texas in the 1970s, they were not subject to federal regulation. They could choose where they wanted to fly and what they wanted to charge. And so they were undercutting prevailing rates.

By 1976, a couple years before deregulation, we started to see the Civil Aeronautics Board allow experiments in price competition as they called it. National Airlines was competing against Southwest and even undercutting their pricing. In order to protect their pricing, Southwest adopted a two-tiered fare structure. They said you can pay our full price, say $26 for a segment, or we'll give you the $13 price. If you're on somebody else's dime (and I think we'll get to this question later about principal-agent problems), you can pay the higher price and what we'll do is we'll give you a bottle of liquor. In 1977, Southwest Airlines was the largest liquor distributor in the state of Texas.

[Patrick notes: The airline makes this claim publicly.]

The birth of the modern program as we think of them now was May 1981 with American Advantage and two weeks later, United Mileage Plus. They began as a way to differentiate what's otherwise a commodity product. Maybe you would wait a couple of hours to have a less convenient flight because you were brand loyal. Same price, one schedule is better, but you want your miles. It was the original way of differentiating these products, especially in a world where there were more airlines competing then than there are today.

The economics of airline loyalty programs

At the time there were quite a few unsold seats on planes. The marginal cost to provide a seat that's going to go unsold to a consumer is practically zero. It isn't actually zero. There are some processing costs associated with an incremental passenger, some additional weight, which has implications for fuel, but it's really just a few dollars a trip. And once that seat takes off empty, the airline is never, ever, ever going to get a single dollar of revenue for that seat again.

They can't. So it costs them almost nothing to give away. They don't wanna undercut the price that they're gonna charge to other passengers just by lowering price to almost zero. Somebody who would pay $300 might get it for 50 bucks - it's costing them revenue. And so they had this opportunity through their frequent flyer program to liquidate unsold inventory at a deep, deep discount in this opaque way.

Consumers value the product they're getting at something approximating the retail price while the cost to the provider, the airline, is next to nothing. And so there's incredible leverage in offering the traditional notion of that saver award. People also have romantic and emotional attachments to travel. So you get this combination of the emotional attachment and the incredible disconnect in value between the cost to the provider and the consumer. And they become really popular.

Patrick: Emphasizing the point that you made, part of the notion was to protect the perceived value of the service being exchanged. Because if one did the quote unquote orthodox economics thing and just attempted to maximize for the capacity usage of plane by allowing the price to go down to someone's marginal propensity to pay for things, there would be people who would be trained that, "Oh, you can actually get between city A and city B for $40 at least some of the time." Maybe I should rearrange my schedules or only fly at certain times of the year to get there for the $40 price that I know will be available in the future.

And then we'll talk about seigniorage mechanics later because airlines are one of the few private industries in the world that allow themselves to print currencies with the government not objecting to that. But printing a currency that a user will value at more than the face value printed on the currency is a really nice trick if you can pull it off. [Patrick notes: The industry with the most comparable seigniorage mechanism, exempting crypto for the moment, is likely video games. When I played World of Warcraft, Blizzard could print me a gold piece any time it wanted, and I valued that gold piece at more than the face value of a gold piece, which is (of course) nothing. Western video game companies spent years staring the economic logic of this in the face and flinching from the conclusion, which is a topic for another day.]

Because of people's emotional attachment to travel and because of the frequent user behavior where people will accumulate reward miles or accumulate reward dollars or whatever the magic currency is, they accumulate points over a year or some longer period of time and then spend them on typically personal leisure travel in a way that will be emotionally memorable to them. As leisure travel is frequently when people do it on yearly basis, or even people will spend 10 years saving up for their honeymoon, for example, and then “splurge” on the honeymoon. [Patrick notes: Guilty.]

And that allows the airline to denominate those points, not in the dollar and cents value of them - and we can talk a lot about calculating dollar and cents value 'cause some effort has been put into that - but in the emotional weight of that experience that is even higher than the actual economic transaction that is happening.

Parallels and differences with hotel loyalty program

So many ways to take this forward. Hotels also have a similar problem. And so, I'll provide a bit of a voiceover on that one for you, if you don't mind. The hotels are more obviously a real estate business than airlines are, but I like to think of airlines as also being in the real estate business.

There's a fundamental constraint on the number of slots at airports. Airlines pay a lot of money to lease a slot for a period of typically years at a time. The slot is the right to fly out of a limited number of fiscal resources in the world. The exact fiscal resources matter if you're in airline operations, but don't matter to listeners of this podcast.

But there are physically limited resources in the world. When you have a slot that implies based on the nature of your operations, that you will be able to fly a certain number of planes per day out of that slot. And so you multiply the number of planes per day, times the days left in your lease times the number of seats that you have on the airplane that you can physically fly, and that is your inventory.

And some portion of that inventory goes bad every day. Like you said, when the flight takes off with a seat empty, you can never recover the revenue from that opportunity. Hotels have the same problem where they have a number of rooms every day, and those rooms can either be rented to someone for the evening or they can sit empty. And if you don't rent it out on Tuesday, you never get Tuesday back even if the room is available for service later.

And so that I think is one of the reasons why airlines and hotels cottoned on to this revenue maximization opportunity where other commodity service providers largely don't have nearly as much built up work with regards to this. We see loyalty programs up and down the economy, but they're at a sharply different scale for airlines specifically, and then hotels as a great step after that.

Gary: Yeah. With regard to real estate, in the US context there's only a handful of slot controlled airports. There's National Airport in DC, JFK in New York, LaGuardia in New York. There are schedule facilitated airports that are especially busy. The binding constraint at a lot of the other airports is gate space. So again, different form of real estate.

You're absolutely right that hotels have spoiling inventory as well. They used to lean into that a lot more in terms of liquidating it through their frequent guest programs than they do today. Generally, it used to be the case that if you wanted to use your points for a room, they may or may not be available at any given hotel. It was based on if they're gonna sell out or not.

Starwood Hotels was the first to introduce what they called "true redemption," but the idea was no capacity controls. If there was a standard room available at the property, you could have that, and they modified the economics.

The economics of hotels is really quite different than airlines. Airlines usually own the planes that carry you. You as the passenger are the customer of the airline. With the hotel chain, for the most part, they've moved into… 

Patrick: An asset-light model. Sorry, I'm anticipating your next sentence. [Patrick notes: This was often misunderstood by people early in the life of Airbnb. Many of the large hotel chains were already at or moving in the direction of the own-no-property place Airbnb started.] 

Gary: Exactly. They don't own the properties, they're mostly a marketing engine. In some cases, they are a hotel management company too, and they have brand standards, but they market their brand and customers are attracted to the brand. They have this frequent guest program, which then delivers the customer to the owner.

There's some hotel owners that refer to the members of the loyalty program as prospects. They're prospects for filling their rooms. And the loyalty program is buying the room from the hotel owner, giving them revenue in most cases for room nights that would've otherwise been unoccupied at a very deep discount.

And in some cases, they're taking up inventory that could have otherwise been sold. And so they have a model where they compensate the hotel owners for the night, based on their actual occupancy for that night. And usually roughly speaking, and it varies a little by chain, when a hotel is more than 90% full on a given night, they'll get something closer to their average daily room rate for that night over the course of the year. But if they are less than that, they'll get a much lower amount that is a little bit over their cost of servicing the room.

Some hotels that regularly fill up don't like this at all because they think, "Well, look, on nights that we could be selling at two times our average rate, we're having to just take our average rate." And there's an interplay between the hotels and the hotel chain in terms of managing the available inventory and playing games with that precisely over whether or not to make the inventory available.

But it's roughly the same idea, but also because of some of these differences, hotel programs to a large extent came a little bit later and matured somewhat later. And early on, they would just partner with the airline programs. You would check into a hotel and you could opt for the airline in which you would collect their miles through your stay. So the hotel is buying the miles from the airline and giving it to the customer as a partial rebate.

And of course the customer likes this when they're the business traveler. In the early days, the programs were marketed to somebody on someone else's dime paying for their travel. They're getting something personally out of it. So they can take their family on vacation by redirecting their employer's dollars.

Absolutely. The early phases of it were elements where commonly expensed items. If you were in a different industry, a consumer's gonna say, "Well, why would I overpay you for something in order to get a rebate back when in both elements, it's my money?" There isn't the element of redirecting a rebate on somebody else's resources that in the travel and entertainment space is especially common as well. So that was kind of the early notion of the development.

Tax implications and legal consideration

Patrick: There's also an element of moving money from one's left pocket to one's right pocket. So there's a famous IRS court case that establishes the service's usual treatment for this, and I will link it in the show notes. [Patrick notes: Here’s the controlling guidance [PDF].]

But as a gloss, and I'm not an accountant and not your accountant if you're listening to this, there are many taxpayers, particularly self-employed people or people who control firms where they will pay for travel spending out of the business and expense that against the income of the business, but receive the personal benefit of travel.

The IRS has said basically, officially it's a rebate on your costs. And so if you spend $2,000 on business travel (subject to all the other requirements), that $2,000 is certainly deductible against the business's revenue on the business's tax return. However, if you receive a rebate, you should adjust the $2,000 by the amount of the rebate.

However, if we actually enforce that against every small business in America, every congressperson in the country would be calling us daily saying, "Why are you busting my local small businesses over tens of dollars of tax advantage that they're getting?" So what the IRS has effectively decided is, look, it comes out in the wash. Businesses are taxed at the rate they're taxed at. We are okay with you taking a moderate amount of value out of a business into your personal consumption through this mechanism. As long as - what's the phrase? Pigs get fat, hogs get slaughtered. There is a particular threshold there.

And the threshold was established in one case where a person arranged his personal affairs, as I believe, a traveling consultant to pull out six figures worth of frequent flyer miles every year by doing every game you could possibly imagine on the frequent flyer thing. And the IRS effectively told this person, "Look, if you were taking a vacation every year based on frequent flyer miles, we'd let that slide. But you are not allowed to pull out a six-figure income that is untaxed out of your business, just 'cause you're a business owner." And so they hit them with the hammer pretty hard, and now people try to keep the degree of redirection and the degree of tax minimization to a dull roar.

Gary: There was also a case where businesses were being pitched by salespeople for one of the big card issuers with an arbitrage play where they take their business, the rebates from their business credit card in the form of points, open up a second personal card that allows them to more or less convert those points into cash personally. And that this would work out as a way of taking untaxed income and therefore a much higher rate of return than they would've otherwise received.

There were ways of paying to move money where the rebate was greater than the cost of doing so. And that card issuer was ultimately fined and agreed that it shouldn't have happened. That the salespeople were to be sternly talked to over that. There are definitely thresholds, it looks really bad when you go beyond the behavior of the ordinary business traveler. And when you draw that extra attention, then we have to pay attention to it.

But yes, as a general matter, if you're a personal traveler and you buy your own ticket and you're getting a rebate, it makes clear, logical sense, you're getting part of your own money back. And so there may not be a taxable transaction, but when it is your employer's money, you are receiving some benefit. Just as employers have limits on how much non-monetary benefit and compensation they can give you without taxation occurring.

But it's one of these things where the frequent flyer programs became so popular. It used to be that federal government employees could not take the miles from their own travel. They were supposed to take the miles and then use the miles to offset their business travel for the government.

That changed during George W. Bush's first term. That administration changed the policy there and they said, look it was simply too cumbersome. It was very unpopular. No federal employee was bothering with this. Either they weren't in compliance with the rule or they weren't bothering to collect the miles because why collect the miles for the government's benefit? And then they've gotta look for award tickets that are going to match their business travel needs for the government. And it's harder to do.

[Patrick notes: The administrative difficulty of policies a) determines whether those policies are successful and b) is heavily dependent on infrastructure which may change overtime. For example, for many, many years, some businesses had travel rules which obliged employees to essentially comparison shop on behalf of their employers. “Too much work, not going to do it” plus “Not worth losing anyone over” mostly made those dead letters.

Then some travel management SaaS providers started building in both the capability and the incentive. I’m a previous user of Navan (no direct affiliation!), which had a really innovative model. They mathed out the company’s allowable expense target for each trip then, if I found a ticket/room which beat it, would pay part of the difference out as an Amazon gift card. It’s probably a bit irrational for me to spend much time looking to save $200 on airfare, but it’s worth a minute or three, and e.g. $60 in Amazon gift cards was enough incentive for me to be slightly more picky than I would otherwise be.]

Gary: These miles are in the same account as your personal miles that you might be earning from other activity with the airline, with a credit card or hotel or whatnot. And just the accounting was not worth it - the juice was just not worth the squeeze. And so the federal government said, "Look, we're going to change that longstanding policy. It's not reflective of any reasonable benefit to the government or of how people think about these things."

Patrick: We mentioned the principal-agent problem earlier, and one of the reasons that the principal-agent problem doesn't get aggressively acted on by the largest and most sophisticated actors in capitalism is that partly there's an emerged set point in our culture that people that are in the professional managerial class who do a lot of flying on behalf of work are simply allowed to do this form of transaction. And if you tell them they're not allowed to do it, they get upset.

Partly it's because there is a certain amount of friction with regards to travel and with regards to booking travel, and some businesses have internal or external travel vendors that handle the selection of flight and similar. But very many businesses think that, "Okay, ultimately, this is a fairly small expenditure on the business with regards to a particular flight. And any savings that we could get with respect to that flight are quickly eclipsed by employee costs in negotiating on exactly which flight Tuesday is the most convenient for an employee when you factor in unobservable things like their daughter's violin recital that they wanna be home in time for, et cetera."

And so we'll push the work to finding exactly the right flight and booking it and similar to our employee. And we will allow them to capture some amount of benefit for that. And as long as they don't skin us alive on which vendor they choose, we'll be basically okay.

Gary: So even in a really managed travel environment where you've got very clear travel policies and you're booking through a portal, and they have contracts with a given airline as well as hotel... Look, employees game this stuff all the time. They say, "Well, what time, what flight time really works given your schedule?" Or, "The airline that you want is more expensive outside of policy. So I'm just gonna wait to book because my plans aren't really firm yet for this trip." And maybe the price of the less expensive one goes up and now all of a sudden the company is paying more.

The airlines compete vigorously for the managed travel business. And part of that is in taking care of their travelers and in knowing that the travelers have incentives that they wanna align with the company incentives to direct large scale travel budget to a given vendor. They extend frequent flyer benefits to the travelers even when they notionally have the contract locked in. Because it is a lucrative business to the airline.

One of the ways to think about a frequent flyer program early on, you used to hear commentators say, "Gosh, these things have become monsters. They're so expensive, the airline, nobody, no airline can stop doing it or they'd lose out. But if you asked any of them, they would just not start a frequent flyer program today." That's wrong. And it's crazy.

If you go back to the early days of frequent flyer programs, they were cost-effective and measurable marketing vehicles compared to what came before them. So what would airlines do in the old pre-frequent flyer days to market their flights? They might take out an ad in a magazine and there was like a three, four month or five month lead time in placing the ad. They're launching a new route. Here's an ad and a market: "We're gonna fly to this place."

Okay. How do you measure the effectiveness of that ad? How do you tie specific business to that ad? These are expensive and slow moving. Now all of a sudden, for the cost of a really small rebate, you are directly able to measure your ability to move the needle on travel, on specific routes, on travel with specific people. And you also develop a one-to-one permission-based marketing relationship.

So it was a more cost-effective way of marketing. So you have lower expense and more effectiveness in your marketing through this loyalty marketing vehicle. So it's not just about a pure transfer of extra spending relative to the alternative. They're gonna spend money marketing anyway. It's a more efficient way of doing it.

And so, as you say, there's an expectation, but if you go back to this mismatch in what the consumer values the product versus what it costs to provide, the company is sort of outsourcing some of the comfort that they might otherwise need to provide on the road, some of the compensation to the traveler, to this loyalty program. "Look, gosh, I hate being on the road so much and away from my family, but at least I get my miles." And so if you didn't have that, the company might have to compensate travelers on the road in some other fashion. So there's a bit of a countervailing force there too.

Patrick: The people who spend the most money every year on traveling as business travelers typically have a total compensation which is far in excess of their travel compensation. And that is because capitalism has determined that moving this brain from office A to office B is extraordinarily lucrative.

Given that it remains lucrative to continue flying that brain in a metal tube through the air, the company would strongly prefer that the employee not say, "Oh, please don't send me to San Francisco this week," but rather leap at the opportunity to serve the business need that needs to send someone to San Francisco this week.

And so, as you said, the company could do that by giving them a salary differential, et cetera. It's very low cost for the company to say, "Oh, travel on one of the nicer airlines. It will treat you right. We will make sure that the airline takes care of that."

Speaking of this happening in the early 1980s, there were a confluence of factors I think, that caused that timing. One is the deregulation that you pointed to in the late 1970s. One was the technical substrate available for doing this, in that it didn't become viable to have personal one-to-one CRM (customer relationship management) systems until the computing revolution made that more viable for firms.

And so the ability to track people on a one record in the database is to one person in the physical world was not uniquely created during the era, but became much more within the capabilities of firms to put together in the early eighties. And that is partly why some of the early adopters in CRMs - airlines are one example, casinos, oddly enough, are another and for similar reasons - they have a long tail of business, which is very long.

But servicing the people at the top of the distribution is extremely lucrative. And so it's important to know when someone arrives at the casino today. Two patrons might look like each other when they're checking in, but you really have to know which of the one is going to spend a million dollars this year, and which of the one is going to spend $300 this year.

So, the co-evolution of it and the airline industry helped to create these programs. The change in how airline tickets are booked also is an interesting differential between airlines and hotels. It's become less the case that hotel booking is distributed all over the world, but a larger percentage of hotel transactions have an intermediary in the middle.

Most commonly these days, the OTAs (online travel agents) like Priceline and similar. And as we mentioned, hotels are operating asset-light businesses. And one of the things that they most like is Marriott.com, their outpost on the internet, and they want - there are signs in basically every hotel: "Book direct, book direct. Please book direct" - because they do not leak tens of percentage of the transaction to the online travel agent in the case of where you book direct.

Gary: Although, in fact, they don't pay the OTA, but they do pay more to the chain in those cases. The chain absolutely wants it. In the case of hotels where "book direct" crowds out your Expedia of the world, it's a better deal for them. In cases where "book direct" crowds out people that would book through some other lower-cost channels or with the hotel, or might book through some unknown way but sign up for the program, it doesn't work out as well.

It's interesting the hotel loyalty programs have grown exponentially over the last decade, and it's not because people are more interested in hotel loyalty, and it's not because people are more loyal to hotels. It's because the chains - you go to Marriott.com and you save 2% if you're a member of the program. And so you join the program, it doesn't mean that person is ever gonna come back to Marriott and they don't disclose their active members, let alone the number of people with multiple transactions in 12 to 18 months.

So you get these numbers that are 200 million members of a Marriott Bonvoy or a Hilton Honors program. But it doesn't actually mean that 20 million people ever think about "I'm a Hilton Honors member."

Creative Low-Cost Benefits and Credit Card Economics

Patrick: I think the program managers were also helped by creativity and how they could take things that were very low marginal costs for the hotel to provide and make them extremely high salience points to get someone to sign up for the loyalty program. The biggest one is simply wifi. Members of a loyalty program get free wifi or you can pay $14.95 a night, your call.

Gary: Right. And Delta now provides wifi on their aircraft if you join the SkyMiles program. And JetBlue was providing it free already.

About 13 years ago I had written - and people thought I was crazy - that eventually wifi in the air is gonna be free. That what has to happen is there has to be enough bandwidth to allow everybody to be using it. And then at that point, it makes economic sense to effectively be bundling it with the ticket when there's not a user trade-off among the users. There's not a higher marginal cost to provide it to one more user.

People said, "Oh, this is crazy. We're never gonna have free wifi in the sky." Well, JetBlue provides it free. Anyone who provides Starlink provides it free, and the first carrier in the US to do that was JSX out of Dallas. But Hawaiian now has that, and Qatar has that, and United's installing it in its aircraft and it's faster than anything.

In the sky you have the low Earth orbit satellites, and so there's just less latency in the communication. But Delta said, "Okay, look, we're going to use this to drive signup in our loyalty program."

The reason they want this, of course, is because then they can market their credit card. That's what makes them profitable. And just about anybody that actually flies Delta regularly, who isn't new to it, has already been marketed that enough times that almost everybody who's gonna get it has gotten it.

And so one of the great values in anything that drives new signups - they may have a lower propensity to ultimately convert, but it's a new pool to fish in. And similarly they benefit from their Starbucks partnership in that way as well. They get more people into the program they can market to because they want to earn Delta miles for reloading a Starbucks card in addition to their Starbucks transaction points, and then they can market those conversions.

Patrick: Yeah. The attributability of this is pretty major. Travel is one of the largest spenders on advertising. But as you mentioned, you pay Madison Avenue to produce a wonderfully lit shot in Hawaii. And who knows how many seats at the Marriott that drives, whereas there's presumably some business analyst deep in the bowels of Delta, who can show you a Tableau dashboard with exactly how many dollars were spent on the Starbucks co-branding opportunity, and then trace that to an actual revenue number aspirationally. [Patrick notes: As someone with some experience of working in marketing, this is… aspirational more than it is the lived experience of marketers.]

Gary: The interesting thing is they have all of this data. They're not always actually great in practice about deploying it, answering questions, and testing their marketing in a way that informs future decision making in a real way. But certainly better than they were.

I mean, you mentioned the confluence of technology developing along with the advent of frequent flyer programs. Some of the early programs actually were managing individual accounts on paper. This is hard to imagine that 45 years ago they were like, your frequent flyer account is this paper ledger.

That had changed pretty quickly, but the technology that was in place in the early days leads to some interesting things today with regard to the nomenclature. If you're a frequent flyer of United Airlines, you've got status with United Mileage Plus. "I'm a Premier 1K." The legacy was, it was the hundred thousand mile flyer was called a 1K. That's weird. 1K is 1000. Right?

Well, the database back at the time at the advent of the program and when they had these status levels allowed for two characters. So you had Premiers and Premier Executives and they had two character indicators. And so this was never even meant to be a customer facing name. It was just in their computer system, the two characters to get the status, and customers began to see it and talk about it that way, and was ultimately folded into the branding. But it's purely an artifact of the database limitations from the 1980s that influences how we call the status level today.

Patrick: Yeah. We see similar economies and status in a variety of places. And they're things that happen organically versus a company understanding that anytime we create a new schema, anytime we create a new namespace, we're creating value in the world. Maybe we should think about how we distribute that.

Low length Twitter names and Twitter names that are particularly sought after - there's an entire shadow economy in that. For geeks of a certain generation, there was Slashdot which had auto incrementing user IDs, meaning they were assigned 1, 2, 3, 4, 5 instead of being assigned like social security numbers randomly over a specified number of digits. And that's an underspecified specification of social security numbers, but bracket that.

The people that had low user IDs were the OGs, as it were, of the forum. And after a while people were like, "Oh, having a five digit user ID is like a mark of status in both this forum community, but also the broader geeky community around Slashdot." And so there would be people selling five-digit user IDs to other people for them to use for their own purposes.

Gary: Yeah, I mean, years ago - maybe this was 15 years ago - in Dubai, the license plate "1" sold for $14.2 million.

Patrick: The city of Chicago also has a number of license plates reserved for the mayor, the cardinal of the archdiocese, et cetera. And they're very serious about that reserved list. Humans are weird. [Patrick notes: You’ll have to go waaaaay into the Chicago Tribune archives to find the citation, but it’s also on the Internet as a scan.] 

Credit card reward program

But scoping back from the great nature of humans to attach value to any namespace that we can create - anthropology is such a fun field - credit card reward programs are an enormous, enormous business. I have my own point of view on these because I think they're probably the largest single contracts written in capitalism. Why is that? Why are these such a great business to be in?

Gary: Right. So, first in terms of the scope - Delta talks about their co-brand deal with American Express representing (there's a lot of rounding up involved in this), but they talk about it as representing something on the order of 1% of GDP on their particular co-brand credit card. I think it's closer to two-thirds of a percent. But that's still an incredible amount.

And going back, say seven years, you would've seen each of the three largest airlines in the US with a hundred billion dollars plus charges across their specific co-brand products. And those have only grown since then.

Patrick: For the benefit of people who don't understand how credit cards make money, I will link Bits about Money’s explanation "How Credit Cards Make Money." 

But we can do the voiceover right now.

Reiterating the disclaimer: I am using no non-public information. Presumably Delta has struck an agreement with American Express, in this exact example, to do a revenue share. American Express makes money on their credit cards in a variety of fashions - charging people interest, but also charging businesses what is called in the industry "interchange," which is a fee for credit card acceptance.

American Express's interchange rate is relatively high in the industry. The transaction that they encourage businesses with is to say: just like airlines like having attributable marketing dollars, you also like having attributable marketing dollars. We will bring you the best, most well-heeled customers into your business, and they will be spending American Express money. And to accept the American Express money which your well-heeled, very desirable customers very much want you to take in return for goods and services, you'll need to pay us a percentage, and that percentage might be, for example, 4.5%. Get a quote from a webpage, not from me.

[Patrick notes: One might suggest reading this page.]

Patrick: So when someone spends, indicatively, 4.5% in interchange on a thousand dollars transaction, that's $45 revenue. And Delta and American Express have pre-committed that if you bring us $45 of revenue via one of your desirable customers having signed up for an American Express account because a Delta flight attendant pitched them on the desirability of doing that, we will pay you some portion of that $45, and we will pay you some portion of their interest charges, and we might even pay you some portion of their annual fee.

Gary: Right. There's a lot. These deals have gotten incredibly complicated as they've grown with a lot of different elements to them. But roughly speaking, if you go back in time to the modern frequent flyer program born in 1981, it didn't take very long before the airlines realized that people absolutely loved these points.

At the beginning, they were an experiment. A promotion. The original American Advantage program in 1981 was conceived of as a two-year promotion. It wasn't even a permanent feature of American Airlines. But they were selling points to travel partners because people wanted to accumulate the points, and those travel partners could reward their customers with points to encourage business.

The very first airline co-brand was the Continental TravelBank MasterCard from Marine Midland Bank. That's effectively through acquisitions and whatnot, HSBC today. But that was just five years in. And very quickly thereafter, we saw co-brand cards with United Mileage Plus (which was with First USA at the time) and with American Advantage the next year with Citibank.

The banks realized that they could get this well-heeled business, high spending business by taking a portion of the money that they were making and rebating it to the customer. Look, they have a huge marketing expense to attract customers because these customers are very valuable. And if we gave some of that money effectively as a rebate in the form of miles that customers really value - value at a higher level than they're even costing us - we are going to get an outsized return on bringing in those customers.

We may make less on every individual transaction if we're spending more on marketing, but we're getting more transactions through our network. And a lot of the value comes from scale in transactions at a lower marginal cost anyway, so it's profitable. And what we're really buying in many cases, especially as the value of the rewards goes up, is "revolve."

So some portion of the customers - we're not just earning the transaction and making a smaller spread. We are going to make money by acquiring a lending business, as people don't always pay off their credit card in full every month. The demographics on this is gonna vary by card. Some of the cards that rebate an especially large portion of the interchange, even all of it, turn out not to get very much revolve at all. And so the banks bet wrong and they wind up losing money.

Look at Chase with their Sapphire Reserve product. It was a huge hit. They introduced it in October of 2016. They were spending a huge amount of money on the customer. Customers loved that, and they filed a series of 8-Ks about how much more expensive it was than they were expecting. Ultimately, they've lost cumulatively a couple billion dollars based on publicly available information on the product, in part because with the criteria for the customer that they're looking for in this product, they're not actually getting all of that lending that you might otherwise expect given the volume of transactions.

Patrick: I have a somewhat minority point of view with respect to how good of an idea the Chase Sapphire product was, but it's complicated. I have to talk around some things here. 

[Patrick notes: I will disclaim, out of excessive scrupulosity, that I am a shareholder of Chase because I bank at Chase and always try to get a few shares in every bank/etc I use to use Investor Relations as an escalation pathway. ]

Patrick: Just like there is cross-subsidization within a particular product in that you could rebate more interchange but hope to make it up on the revolving business attached to a credit card, certain banks have the full ply of financial services to sell to someone. And they might make the decision that there are certain customer cohorts where some products for the bank are sold at very thin or even slightly negative margins to a large number of customers.

The classic checking account is very thin or negative margins to probably 80% of Americans. But then you hope to get the entire banking relationship and then make your money in other places, such as in mortgage origination and similar. And so credit cards were historically not one of the quote unquote "loss leaders." But some banks have some products which might or might not have been intended as loss leaders at various points.

Gary: Yeah, so that's certainly the justification that Chase provided after the fact with regard to Sapphire Reserve, when some of the numbers started to come out. I'm not entirely sure how well that's actually worked out for them in terms of generating, say, mortgage business, judging from the outside.

They had, early on, some promotions where they tried to stimulate Sapphire Reserve cardholders earning the Ultimate Reward points attached to that card for other sorts of transactions. And we haven't seen them go back down that well quite as often. And my sense is that they didn't quite work as well as they'd hoped.

But at the same time, if you're a bank the size of JP Morgan Chase, $2 billion sounds like a lot of money, but when you're talking about over the course of a decade, it's not super material. And it is a prestige product, and it's reasonable to make all sorts of bets.

When Citibank took the Costco card away from American Express a decade ago, this actually was one of the impetuses for the cost of some of these deals for the airlines beginning to really skyrocket, setting a new bar for the price that banks were willing to pay for co-brand deals.

There was a bet Citi was making, and they were clearly going to be overpaying just on the basis of the transaction volume they were getting. There was a bet on revolve. There was a bet that they wanted to acquire these customers. I'm not sure that the execution's been as great as hoped in that particular product. Just the frictions involved - for quite a long time, to get a Costco credit card, you had to be a member of Costco, and at the beginning of this deal (and it's a multi-billion dollar a year deal), if you wanted a Costco credit card but weren't already a Costco member, they would tell you, "Oh, we can't give it to you. Go over to Costco and sign up and then come back to us." They couldn't even get the technology going to do it all in one transaction.

So there are problems attached to it. But there are bets, and not every bet works out as well as you hope. But you can make a reasonable argument that there's positive expected value to be created.

Patrick: And it's even a multi-party bet too, because (using public information) one of the factors involved in the decision to take a swing at Amex for customers that are on the upper end of the American socioeconomic spectrum...

Amex does business substantially worldwide, but they were most interested with respect to US home customers. Amex competes with Chase as an issuer of cards, but Amex competes with Visa as a card scheme. And so Visa would like there to be more Visa cards in the upper echelons of American society than currently prevail.

And Chase said to Visa, reportedly, "We can do that, but we can't do that under the current constraints of Visa's rules and price tiers." [Patrick notes: You can read them, but be forewarned, trying to understand interchange pricing has driven smarter engineers than me mad. This PDF is an infohazard.]

Patrick: Every credit card is assigned on the backend - and if you are extremely adept at reading disclosures, you can probably reason out which one it is - but it's assigned a tier by the credit card brands, and the tier determines the interchange rates that get applied to merchants under the hood. And that sometimes might or might not be the price that the merchant actually pays because there could be someone in the middle that is paying it on their behalf and marking it up, but there are price schedules. You could find the price schedules on the internet.

And Chase told Visa, "The top line in this price schedule just does not have enough interchange in it for us to make this worthwhile to the best customers. American Express charges much more than your most exclusive cards do. We want a new top line - make an entirely new category, and there will be one entrant in it when you start out. It will be ours and we will go head to head with Amex for this." And Visa did that.

And then some jostling has happened, and I have limited ability to comment on the jostling, but the Chase Sapphire Reserve literally caused a change in how Visa cards are priced worldwide.

Gary: And of course, Chase can do that too because they're predominantly a Visa issuer. They have a deal referred to as "leasing Visa's network," which allows them to drive much higher levels of spend without much higher levels of cost. So they're collecting more of the interchange themselves and funding some of the efforts that they're undertaking. So lots of different things in the deals that make this all work.

Patrick: Chase is also one of the largest credit card processors in the world. And a mistake that people sometimes make outside of the industry is to view a credit card issuing program only within the four walls of the issuance itself.

But there might hypothetically be a team at Chase involved in issuance and another team involved in credit card acceptance. And you can hypothetically link the complex multi-year, multi-billion dollar commercial discussion about credit card issuance with a particular large consumer of credit card services with the multi-year, multi-billion dollar discussion about who do you use to process your credit cards. And then play games around where the margin is made. So that is the thing that might or might not happen in the world with regards to some issuers.

Gary: Well, and when you look at some of the deals, say with the airlines, it's not just around the credit card spend. So with Delta, there are a lot of ways in which American Express and Delta go back and forth and embed themselves into each other's ecosystems.

Delta has an American Express card that they use for the purchase of jet fuel as part of their deal with American Express, so that American Express can earn interchange on the purchase of fuel, a good portion of which then of course gets rebated back to Delta in the deal. Delta disclosed the credit line on this credit card for jet fuel at $1.1 billion in an SEC filing.

The David Dao incident and its impact on airline policies

Delta's the most generous with overbooking compensation. If you go back nine years, airlines for the most part had more involuntary denied boardings than they do today. There was a famous incident about eight and a half years ago with David Dao on the United Express flight where United needed to take him off because they wanted to put employees on the plane to operate another flight. And so they said, "Look, we don't have enough seats and we're gonna ask you to get off." And he refused. Chicago aviation police came on. This was a very, very unpleasant situation. The man was taken off and bloodied, and there was a worldwide outcry and worldwide discussion about overbookings.

Airline compensation strategies and Delta’s unique approach

Airlines would take volunteers to try to limit this, but they wouldn't give you very much compensation because the most they ever had to do was what the DOT said they would give you in cash if it was involuntary. They wanted it to be voluntary. They wanted to pay you less than they'd have to pay you in cash. At the time, I want to say it was $1,350.

Shortly after that, it became very bad for airlines to have involuntary denied boardings and they would go to extreme lengths, paying out in a couple of edge cases close to $10,000 to be voluntary, not involuntary. Now that began as a cost center to get clamped down on by airlines like American and United, just before the pandemic. The pandemic really tightened things except at Delta, where they really do still make efforts to almost never have an involuntary denied boarding.

The other thing is that airlines usually give out travel credits, not cash. Delta gives out... Can you guess what they give out?

Patrick: Is it gift cards with an issuer attached to them? It might be. Can you guess what issuer is attached with them? I think it might be - not that I've ever received this in the recent past - I think it might be American Express.

Gary: Yes. So there's all of these ways in which they get up to some of the really big numbers. The credit card company is paying for checked bag fees for card members, and that's part of why people get the card and it puts it in the wallet. And so it's part of the acquisition cost. They're paying for lounge access in the case of some of the premium cards with the airline. There may be dollars attached to the earlier boarding.

And then the accounting on this gets really interesting. If I just buy an airline ticket and I earn miles for that ticket, the airline is the party to both of those transactions directly with me. And the accounting rules under what's called ASC 606 say that they have to make a provisioning for the value of the travel that's going to be provided.

So they're gonna recognize most of the revenue upfront for the ticket, but they're going to provision (and the big airlines are roughly speaking the same on this) - they're provisioning about a penny a piece in value per mile for future travel that they provide.

The accounting is entirely different when they sell a mile to a third party. It's not splitting the transaction with themselves anymore. And the rules are much, much, much looser. And so when they sell the mile to a bank, they say, "Okay, we're gonna split this out between how much money are we getting for renting access to our database to market the card? How much are we getting for use of our brand? How much are we getting for those bag fees?" And we're kind of deciding ourselves. We have to have a plausible model that passes muster with the auditors that we hire. So there's industry standard ways of doing it.

But by the time you split across all that revenue, the amount that actually gets booked as a liability for future travel is on the order of an eighth of a cent. The exact amount's a little bit different by airline, but roughly speaking, there's this eightfold difference in the value of the same mile in the books of the airline.

The airlines really, really like the sold miles. In fact, the head of the Avianca LifeMiles program (Avianca is an airline in Star Alliance) - I was at a talk once where he said that they describe the miles that you earn from flying as "bad miles" and the miles that they sell to partners like credit cards or that you buy directly from them as "good miles."

Patrick: And that's just in the year of issuance. After the miles are in the bank, theoretically speaking, they are a liability rather than an asset. However, we've discussed some ways to do accounting gamesmanship, and accounting gamesmanship can be useful if you're having, for example, a soft quarter or a soft year. You might push a little more on various people to get more revenue recognized this year so that you can make your numbers, get your bonus, et cetera.

In a future year, when you owe customers literally billions of dollars of this deferred service provision, you can do things to cause that cost to go lower and then recognize the difference between the current valuation of that and your new restated valuation as a revenue opportunity for this year. Can you describe that process?

Gary: Right. So there's a couple of pieces to this and a couple ways that they do this. For one thing, fewer airlines expire miles in the US now than used to. But changing expiration policies was one way of moving the liability off the books.

So first, you have these miles as a liability on your books, and you have an assumption about how those miles are going to be used in the future. Small changes in how often they're used or how many miles go unused have huge impacts to revenue given the scale, so they can make changes to their assumptions about future use or how many are going to go unused.

They can expire miles, literally moving miles off the books and into revenue that had previously been assumed would be redeemed. You can also change the price of redemptions.

During the Great Recession, airlines - one of the ways that they generated liquidity was through what was structured as loans, but they were pre-selling miles to their co-brand credit card partners for large amounts of money. You would pre-sell 500 million or a billion dollars worth of miles at a discount to your credit card bank partner. And that raised a lot of money, but nothing compared to what they did during the pandemic.

First, as part of the first CARES Act, American Airlines accessed federally subsidized loans backed by their frequent flyer programs to the order of about $5 billion. But eventually they paid those back and borrowed more by going into private debt markets. Delta and United went to private debt markets. These three airlines each raised between six and a half and $10 billion against the future revenue streams of their frequent flyer programs, which is largely selling miles to their banking partners.

One of the ways that they marketed this to the debt investors is explicitly to say, "Look, one of the reasons you can be sure of getting your money back - now you're at the front of the line, frankly, even ahead of our customers - because we control the currency. We control how it's earned, we control how it's redeemed, and we control therefore the pricing. So we can adjust those levers to manage our expenses to ensure that you're prioritized in getting your money back."

So adjusting the prices is another thing. Now, there's always the risk that you harm your future revenue if you alienate your customers. And historically, my sense is that some airlines have alienated customers more than others, even when taking similar steps.

For the most part, Delta had not suffered from the same reduction in spend volume that followed devaluations at United and American. This is partially because of a stronger brand and stronger operational reliability, partially because of the role that Delta plays in some of its hub cities. If you're in Atlanta, you're probably flying Delta and you associate that with the credit card you need to have.

Most consumers aren't thinking through the challenge of, "Well, if I actually get a Membership Rewards earning card that transfers to Delta, I can even earn more Delta miles than if I had a Delta card. And I have other options as well." But Delta hadn't suffered in the same way until, interestingly, perhaps a year and a half ago.

The impact of airline policies on customer loyalty

Delta announced major changes to their elite status program and their lounge access for their credit card customers. And perhaps because it was centered on their most fervent, most loyal customers, the backlash was tremendous and they very quickly backtracked on those changes - not entirely all the way back, although they also added some sweeteners about a month later.

I even expected that they would make a little bit more of a change a year later. And they didn't. By all outward indications, it seemed clear that they got scared because of things they were seeing in their relationship with American Express.

People were certainly saying, "Look, I'm canceling my American Express card," or "I'm sock-drawering the card. I'm not using it anymore." But there was a huge outcry and a bigger risk perhaps than the benefit, which was that much of the changes were meant to drive spend on that American Express card.

Changes to how you were earning elite status meant you were going to increasingly earn it through things other than flying. And the simplest one was by spending on the co-brand American Express card. You were going to retain the amount of access to their lounges that you already had by hitting spend thresholds with their premium American Express card. And without doing that, the stick was you were gonna get less. But it was to such a degree that they may have pushed customers a little bit too far.

There are also other revenue recognition games on the flip side. When you have extra seats that are going unsold, there is a built-in reason why you want to make those available for points redemption. You want to encourage that redemption because when the points are redeemed, you recognize that liability off the balance sheet as revenue. So instead of getting nothing for those flights, you are actually recognizing revenue at the point of redemption. And if it is not costly to do so.

I know during the financial crisis, getting award seats was like the easiest thing. I helped someone one time - they said to me, "I need to get my family of nine from Mexico City to Frankfurt." And I said, "Well, how would you like the Lufthansa nonstop flight? And how would you like all nine passengers to go not in business class, but first class? Because that's available on your preferred date of travel." That's not so easy and much more expensive now.

Economic Shock Absorbers and Market Dislocations

Patrick: This also functions as a bit of an economic shock absorber for the airlines. Since airlines are exposed to a relatively small section of industries for a large portion of their travel business - we skew towards tech and finance in the United States, tech, finance, consulting - those are three very large travel spenders.

During the financial crisis, go figure, a lot of the finance and consulting firms stopped spending so much money on business travel. The airlines are able to say, "Okay, we would still prefer to earn revenue for flying airplanes this month. One way we can do that is by going to our book of customers and say, 'Alright if you're paying us in our own currency, we're going to be extremely solicitous of taking that business right now.'" We would be less solicitous if McKinsey and all were spending as much money as they did a quarter prior.

Gary: Yeah. Airlines have gotten a lot better now at filling planes than they used to be, certainly better than at the beginning of the programs. Whereas you might have had 60-65% load factors 40-45 years ago, you often have 85-90% now. And that's overall, which means there are some routes that they're selling out virtually every plane, every day, every seat. So there's almost never the case that there's that extra seat that costs them almost nothing to provide. And there are a few routes where there's lots of extra seats.

Patrick: I have a hilarious anecdote about what causes at least one route in the world where things are routinely under provisioned. A large automobile manufacturer in central Japan considers it to be very positive for its interests for there to be a certain number of flights between central Japan and Tokyo, but central Japan is their number one concern here, to Detroit specifically.

Japanese leisure travelers don't open up their atlas to the world and say, "You know where I really want to go on vacation this year? Detroit." So that's a plane that would not have a huge amount of utilization normally. But this automobile manufacturer doesn't want to, on any given Tuesday, need to charter private jets. So on any Tuesday, there must be a certain number of committed flights for that reason.

Also, because the same flight that is flying a very small number of brains between factories of this automobile manufacturer and their American counterparts and also their American factories, there are some relatively low-weight, relatively high-dollar value cargo things that are flying on the plane routinely. And so that particular flight would be economical with five people on it.

Thus I have occasionally been on that flight because Detroit is very close to Chicago. It's like, "Wow, this is really nice to be able to spread out. Can I expect this flight to still exist next year?" The attendant said, "Oh yeah, this one will be around for a while."

Gary: There used to be a United Airlines flight from San Francisco to Nagoya for a very similar reason. And it was always precisely the case that if you could find award seats in business class on no other flight, you could always get it on the United San Francisco-Nagoya flight.

On the other hand though, some corporate deals for travel are so big. Before the pandemic there was some information that was unfortunately from United's perspective released by some of their salespeople that they were selling 50 business class seats a day between San Francisco and Shanghai for a company that is often referred to as the Cupertino Fruit Company, also known as Apple.

Patrick: Apple was extremely displeased about that information leakage.

Gary: They absolutely were, but the story of how airlines have gotten better at filling seats over the years, meaning that there are fewer available seats that are virtually costless to offer, has tremendous implications for the pricing of frequent flyer miles awards.

When - and some of this has to do with having fewer airlines, some of it has to do with better technology and marketing, some of it has to do with just who runs the airlines and how they're run - but given that they're filling planes...

I wrote something 22 years ago which basically said, "Here's why you can expect the cost of frequent flyer mile awards to go up." Just think of it in the simplest monetarist framework. Think of miles as a currency, and understand nothing more than the simple formulation of MV equals PQ. You have the amount of money, or in this case, miles in the economy and the speed at which they're redeemed on one side, and you have a quantity of available award seats on the other and its price.

You're printing more and more miles, but the quantity of available seats to offer is not going up. In some cases it's even going down. So that alone is going to suggest an imbalance. The imbalance either means that you have a shortage, and in fact, there was, dial back 20 years to David Spade doing commercials for Capital One as they built up their own branded miles program that weren't frequent flyer miles, but a currency to spend on travel. And his idea was that nothing is available. You call up wanting the award and they say, "Not available, no, no, no." That was the mantra. And so customers were frustrated even then.

But you either had the shortage idea or prices needed to go up in order to ration the available seats. And what airlines have done is they said, "Look, we know we have to deliver to our customers. We can't just tell them nothing is available or they're gonna stop accruing the points, and we don't want that to happen because this is really richly valuable to us."

Patrick: Right. You need to protect the perceived value of the currency.

Gary: We have a problem, and what airlines have mostly done with their own flights is (there's still sometimes outsized value, less so with Delta), but what they do is they say, "Okay, we're not just making these seats that are going unsold available. We're making more seats available - at some level, almost every seat at some price."

And the frequent flyer program is going to buy those seats at some internal price. And when we're actually displacing paying passengers, the value per mile - that perceived difference that we talked about between what it costs versus what the consumer values - that shrinks. So the value per mile shrinks and you're pretty much just spending miles as a currency to buy the thing that you want from the airline at a lower value.

That's why you often get the best value using miles from one program for seats with its partners, because usually in most cases that is still only dealing with the seats that are liquidating excess inventory.

And it's also the case in the US context that with foreign partner airlines, there's often a lot of excess inventory. I used to live in DC, I lived there for 18 years, and this was a great market for it because out of Washington Dulles airport, you have flights on Qatar Airways and you have Saudi, and you have Emirates, and you have Etihad, and they're all flying pretty near each other.

The economics of the market doesn't necessarily support that, but they all have this idea that they need to fly to the capital of the United States. For a while, ANA had their flight to Washington Dulles. It was ANA flight one for a reason, representing the relationship between Japan and the United States. And as a result, all of this excess capacity meant a lot of unsold seats. It was a great place for me to fly using miles.

One of the best places to fly right now internationally, if you want business class seats to Asia - and Asia flights have gotten really, really hard with miles, and part of that is because of US relationship with China and there's just a lot fewer flights between the US and China than there were before the pandemic.

There was this land rush of Chinese airlines trying to get into the United States. China had a rule that Chinese-based airlines could not compete with each other on the same route. So it was only one airline allowed to fly a given route. They all wanted to be first to absolutely stupid routes that they might eventually wanna fly later - cities that most people in the US have never heard of, but that meant empty planes that they were flying every day in some cases with the hope of eventually needing it, but so that nobody else could have it.

And then you could buy tickets on these flights and connect elsewhere in Asia. China Eastern via Shanghai, you could often buy for $400 round trip in coach. But with all this excess capacity into Asia, through Asia, it meant it was drawing passengers away from other flights too. And there was pricing pressure, and you could get award seats.

Well, without the return of these flights, which by the way, US airlines have lobbied against allowing - China originally cracked down on them during the pandemic, but the US airlines didn't want it because they don't want the overcapacity - it hasn't been as easy.

You get plenty of award seats in business class flying between Seattle and Taipei because you have three different Taiwanese Airlines and Delta flying the route between Seattle and Taipei, which is sort of crazy. And so it's the one route that you can sometimes get good business class award values with SkyMiles because they don't wanna give up the route.

Patrick: I don't even know that that one is necessarily crazy on the fundamentals given fruit companies and other people who might want to take advantage of exactly that route.

Gary: But that there are four airlines doing it and not all of them have the relevant corporate deals is what I'm saying. And what that does is it depresses prices. If I've got the chip business and I can make the economics of my flight work, but they don't all have it.

Patrick: Makes sense. So, wow. I feel like we could go for a million miles on other places for customers to benefit from quirky dislocations in the global economy or nation state level politics with each other, airlines doing politics with each other, which is also a thing.

The value of partner miles and airline alliance

I think one of the reasons that partner miles maintain their value in a way that first party miles sometimes don't, is it's no longer an airline just negotiating with its customers as a class - you're negotiating with another very sophisticated partner that might have you locked into a multi-year contract where there are expectations of continuity. In airline programs, those expectations are typically not contractualized with an individual user. Does that match your expectations?

Gary: And in fact, sometimes those expectations are contractualized between the airlines in terms of their multi-year deals. In other words, what they charge to a consumer may be in the deal. And these are multi-year deals.

So when American Airlines changed the award chart for its own flights a couple of years ago, I asked, "Well, what are we gonna see changes to award charts for your partner flights?" And I got told, "Oh, that's hard, because each of the partners was negotiated bilaterally. And so it is such a long process that that's not on the immediate horizon."

Patrick: Right. And it's not merely just negotiated bilaterally, but they're on different start-stop dates, different cadences and so on. Multilateral coordination of organizations is hard. The alliances are difficult aircraft carriers to turn, et cetera. And so, there will likely still be some alpha to be had by savvy customers with respect to looking for partner flights.

Gary: At the same time, these are negotiated differently depending on different alliances as well, and which airline is making the change matters too.

Interestingly, for earning status with an airline within oneworld, there used to be a rule that said not only did you have to meet the flying criteria of your airline for their status, but you also had to have a minimum number of segments on that airline. So you couldn't just earn British Airways status only flying American Airlines. And that was a Oneworld rule.

American decided they didn't wanna do that anymore, so they just stopped. The rule went away, because in fact, oneworld headquarters moved from New York to Dallas to American Airlines headquarters. So the influence in the alliance was such that the rule mattered and constrained all of the parties except for American Airlines and subsequently Alaska Airlines, which is also a member of oneworld, has gotten rid of that rule for their own program. But it's no longer an alliance rule. Some of them still have the rule because they wanted it or because of historical legacy reasons. But there are rules and they apply differently to different players.

Patrick: I'm curious, what was the American Airlines reason for wanting to go away from a flight-based qualification? Was that to sell more credit cards or just because the contours of their business are a bit different than the contours of some others?

Gary: During the pandemic, they made a big change to how status was earned. They no longer said you had to fly a certain number of miles and spend a certain amount on tickets. They said, "Look, you're gonna earn what they call loyalty points. And loyalty points are earned from most activity in the program."

Because ultimately, selling miles to partners has a lot higher margin than earning miles from flying and selling a ticket. In fact, even selling a ticket is the lowest margin activity that the airline itself has when it interacts with a passenger. Selling a seat is higher margin, selling an upgrade, selling bags, any upsell is gonna be higher margin.

And so the thing that they were rewarding frequent business in was their lowest margin activity. And they really want to use this most powerful part of their program to incentivize growing the high margin parts of their business.

So they said, "Look, in the past, maybe you spend on airline tickets and you fly a certain amount and you earn status. But actually it's a very valuable customer to us, somebody that doesn't fly very much at all, but puts a couple hundred thousand dollars of spend each year on our Citibank credit card. Citibank is our biggest customer. We want them to be an even bigger customer. We wanna encourage you to do that. What do we consider valuable?"

And so they reconceptualized that and restructured their status program. And increasingly, all of the US airlines have had a growing role for at least their co-brand credit card and sometimes some other activities, but none had gone quite as far as American said, "Look, you can just flat out earn status this way."

And American wasn't the first, to be sure. Frontier Airlines before the pandemic said $1 of spend on our credit card is a status mile in our program. They just didn't have quite as much to offer as American. It didn't take off the same way.

So they said, "We wanna have full rewards for this and not have a minimum number of flights. And if you wanna earn your status through us, by the way, it doesn't even cost us very much to deliver status benefits if you're never flying us. Like, go for it."

Creative ways to earn miles and exploit promotion

Patrick: I have to make an aside about large credit card transactions. So obviously you can get to $400,000 a year in spend on a co-branded card in a variety of fashions. One is to just put all the business's expenses, from office paper to et cetera, on that card. There are other and faster ways.

Once upon a time at a credit card processor, there was a young engineer who was faced with one of many design decisions that one makes when coding up a computer program, which is: what is the largest number we would accept as input here? I won't say what the largest number was, but this was an engineer who didn't have a huge amount of experience in the world, asked to decide what's the largest amount you could conceivably do in a single credit card transaction. And they picked some number X.

Not the next day, not the day after, but pretty soon people started complaining to the credit card processor. "Hey, your maximum charge is X. What's up with that? We want to do transactions that are much larger than X."

And the follow-up questions: "Why? You understand we charge you a percentage times the dollar value. And that makes a lot of sense when the typical charge is $40 or $80, but at X you're paying an awful lot per transaction. You know this, right? Why aren't you doing this over a wire or something?"

And the customers would say, "We've got reasons. They've got reasons. It's not your job to second guess this here - bump up the limit." And the limit has gotten bumped a number of times in response to that.

There are various transactions in the world that you would never maybe expect if you've never had to make them. But the example of the American Express fleet card for - "Okay, I've got a 747 in Paris, fill her up." That's a bit of money. [Patrick notes: I would not expect that exact transaction to be done on a credit card, but could I imagine buying a jet worth of avgas on a credit card? Er, yes, yes, I could imagine that.]

Patrick: And not using anyone's private information here, but captains of ships, inclusive of super tankers, have substantial necessity to be able to transact anywhere in the world they find themselves in. As you know, many credit card brands have statements like, "Visa, it's anywhere you want to be." And so in lieu of having like a safe full of gold bars or hard currency, although they also have those wildly, they will often have a fleet card that has - okay, how much oil does it take to get a super tanker from one part of an ocean to a different part of the ocean? The card must be able to pay at least that much. Wild, but it works.

Gary: A couple quick stories. Ten years ago there was a lot of news worldwide around a wealthy Chinese man who won teacups at auction in Hong Kong and earned 2 billion Membership Rewards points. Now isn't quite as much money as it sounds like because the Hong Kong Membership Rewards program is sort of an inflated number - divide by eight basically, and you get an equivalent.

But the story was about how he paid with an American Express card and earned his points, and you think, "Wait a minute, auction outside of the Chinese mainland taking place by credit card." And you immediately know where this is going. What is the incentive there? And it wasn't the points.

Patrick: Yeah. This is evading capital controls, which is a business that many credit card issuers would not officially like to be in the business of. But you know, you're not required to understand the entirety of your user's economic life. [Patrick notes: As Bits about Money has covered many, many times, the financial industry is deputized as an arm of the state, including in more freedom-loving places than China, and being in the business of offering products contrary to the goals of the state is a rough place for it to be.]

Gary: There was a plausible reason for the transaction and frankly, with the auction itself.

Now I've gotten merchants having their accounts shut down because of the size of transactions I've put through them. In 2009, US Airways ran a holiday shopping promotion where they were offering a 250% bonus in the miles that you earn, provided that you made between six and 10 different transactions during a particular period of time, and no more than one qualifying transaction per merchant.

And they had one partner, a company called Track It Back that would sell you stickers. The idea was you put these stickers on your stuff and they would help you find the thing if you lost it - put a sticker on your laptop. There's a number to call, they're gonna provide a reward and take care of shipping if you lose your laptop and somebody calls in.

Now it turns out what the reward was when the person called was very low cost to them - it was stickers. Normally they would give out 20 miles per dollar, and it was this great way of, if you could expense stuff like that, you could get a bunch of US Airways miles.

Concomitant with this holiday shopping promotion, they decided to offer double the miles or 40 miles per dollar. And they thought this is a big marketing push for the holidays and people are just gonna buy, and this is great. And they didn't really do the math or understand the implications of this promotion.

So 250% bonus on 40 miles per dollar is 140 miles per dollar. You're buying miles at about seven-tenths of a cent a piece. And if you're really ingenious and decide to donate the product that you receive to a qualifying charity for the tax write-off, maybe you reduce your cost basis to about half a cent.

And with the cost of awards at the time, say you were buying business class round trip tickets to Hong Kong for $450. So this is like a pretty good deal.

So for myself and others, Christmas morning 2009, I'm sitting in a hotel room effectively with a stack of Amex cards. And just one after another buying stickers from Track It Back. Now they're used to doing $40 transactions, $100 transactions. But I'm aiming for earning 16 million miles, so we're doing like $100,000 transactions.

This gets denied. I'm on the phone with the president of Track It Back on Christmas morning, and he is trying to get his Amex merchant account reopened. And offhandedly he says, "No, I gotta say that we're out of product. But don't worry, we'll have some shortly after the new year."

And I'm like, "Okay, stop. Here's the thing you have to understand," because he didn't quite still get this: "Nobody actually cares about your stickers. Traditionally merchants are going to post the transaction with the miles when they ship the product. I said if you do that after the first, when the promotion expires, you are getting everything returned. You do not want that. Post the miles now, and ship fulfill your end of the deal later. Everyone's gonna be good with that. Trust me."

He never got his Amex merchant account back. They just - it was just too much. He could never come back. So we were using Visas. But in one day, one promotion, we generated 16 million miles across different accounts.

The infamous coin promotion and other mileage hack

Patrick: This is bringing to mind one thing that I just have to mention here, which is occasionally both the financial industry and the airlines have been surprised by the degree of adverse selection that has happened with respect to promotions.

In one case, the United States government was surprised by this. The Mint had a wonderful idea: "We have new commemorative coins. US currency is worth what US currency is worth. We like to protect our perceived value of currencies that we generate seigniorage revenue on. So we will sell you $1 of commemorative coins for $1. With no additional fee charged."

Gary: And no credit card charges.

Patrick: Yep. And the thing that folks like yourself discovered very quickly was that you can buy any number of dollars from the Mint. Really. It's not like they're gonna run out. And you received the dollars in the mail, walked the dollars over to your friendly local federally chartered bank, which will credit you with $1 in your account for every dollar of US currency you bring them, earn the points, and then round trip this as many times as possible based on shipping speed and similar. And people were racking up - I think there were reports of deep into the six figure spends on commemorative coins.

Gary: I'll tell you a bit of the rest of the story on this. There's a frequent flyer trip that I used to help with called a Mega Do, where a bunch of frequent flyers would literally charter a plane from an airline and do tours of stuff with airlines and hotels and kind of behind the scenes sorts of things. And on the Star Alliance Mega Do Trip we'd get some media to come along too.

Scott McCartney, who I know really well, came on the trip. He was the airline reporter for the Wall Street Journal, and he gets to talking to a guy on the trip (I won't name him in this conversation, but he's public about this stuff). He gets to talking to a guy who did $2 million worth of coins. This is interesting and became a Wall Street Journal story on these dollar coins.

And so once it was a Wall Street Journal story, we talked early on about how things, when there's too much attention on something, it becomes very awkward for a lot of people and they have to pay attention and begin to clamp down. And that's a lot of what triggered some of the clamping down.

But there's another piece of it that lived on in the broader culture. It was this Wall Street Journal story apparently that inspired the idea of the trillion dollar platinum coin being deposited at the Federal Reserve as a solution to the debt ceiling. So it has all sorts of implications, but it was this frequent flyer charter trip 15 years ago that triggered a lot of the attention to that particular opportunity.

Sometimes, by the way, the deals come more directly at the expense of a party. You think this works out for a lot of the players. You say, "Okay, the government wanted these in circulation. It wasn't actually getting the circulation, but their metric was how many are they shipping out? And so they were happy - we're able to report great growth in these dollar coins."

The credit card company is getting their piece on the transaction, and the customer is getting their miles, and the airline is getting paid for the miles, and everyone's pretty happy until there's too much attention. Sometimes it actually does cost one of the parties involved.

So in December 2021, American Airlines has a venture with MasterCard called Simply Miles, which is card-linked transaction offers. You sign up, you opt into offers from a merchant, and when you transact with that merchant with a MasterCard, the transaction detail flows through and triggers this offer and you get rewarded with miles.

They ran a year-end holiday promotion of sextuple miles. And MasterCard was funding an offer with a charity they were supporting called Conservation International with 40 miles per dollar. And they were also funding this bonus effectively - they were buying miles from American as a joint venture, and American was offering them a lower price on the miles for this.

But at 40 miles per dollar and sextuple miles, now all of a sudden you're buying miles for about 45 basis points, 43 basis points. Well, this sounds pretty good, and it was supposed to be available only while a certain supply of miles lasted, till they hit a cap, because MasterCard had a budget for this.

Well, it exploded on a few places, including my blog, over a weekend. So it wasn't maybe monitored as closely over the weekend. And they immediately hit that cap, but they didn't remove the offer until Monday. And by my estimate it cost them several tens of millions of dollars more than they had planned.

And so there was this period of limbo where nobody would confirm that they were actually honoring the deal. I had said, "Look, they're going to honor it. I mean, worst case, they're gonna do some refunds, but there's no way that MasterCard isn't going to stand behind an offer to reward donations to their sponsored charity at Christmas. There's no world where this is not working out. So take the operational loss, move on. We've learned our lesson."

But somebody had to go to their boss or go to their boss's boss and say, "We have an oops" and kind of recognize this. So it took a few days to work through that. But I did 7 million American Advantage Miles on that. I had a reader that I think did 50 million. But there were lots of people who did quite a lot.

And because American had just started to reward transactions other than flying with status, a bunch of us became Concierge Key members at American Airlines - their super top secret unpublished level that George Clooney was in the movie "Up in the Air." Because all of a sudden, American was recognizing probably $70,000 of revenue just from MasterCard alone on me. So I looked like a valuable customer.

A bunch of eyebrows were raised when everyone saw that I had become a Concierge Key from this thing, and I didn't get renewed and I knew that I never would be, but it had lots of benefits to it. But it did come out of at least one pocket there.

Famous Mileage Promotions and Conclusion

Patrick: I think the most classic example of an airline, I think it was also American, getting heavily adversely selected with regards to a promotion was way back in the day. They decided to do - I don't know if it was an attempt to raise capital or whether it was a marketing stunt that went horrifically right - but they sold lifetime tickets for, if I recall, about $250,000 after having mathed out what the expected usage over a lifetime for airline services was.

And there were people in various roles such as finance or traveling consultants or similar, who are very good at doing that math and had more knowledge of their current intentions or planned course of their life than American Airlines did, and leapt at the chance. They've continued to honor their obligations under those lifetime passes. While it's been reported every few years in the media that they're not revoking those contracts because contracts are contracts, they've attempted to tighten on some of the edge cases there.

Gary: Well, the edge cases - people who bought, there were different versions of these passes. Some people bought for themselves and a companion, so whoever was gonna fly with them. And there were people apparently who would sell the companion travel, or they would just accompany the companion wherever the person needed to go. And they would make money by flying. This was obviously against the rules and there were people who had their passes canceled.

In terms of how much some people did with this - the most frequent flyer of United Airlines is a man named Tom Stuker, who has flown over 24 million miles with United and has two planes named after him. The incremental flights don't cost him anything. He earns the miles which he turns into gift cards. He literally makes money flying. So it didn't quite work the way that was intended.

I thought you were gonna say the probably most famous mileage transaction. It was the pudding cups and Dave Phillips who earned over a million American miles. But it was memorialized in the PT Anderson/Adam Sandler film "Punch Drunk Love." And they actually did other things, worked on other promotions as part of the filming and scripting of this that didn't make it to the final cut.

There was a program called Latin Pass that was awarding a million miles that included, you had to fly on all of the different partners in order to earn those miles. And so some of this was actually flying on planes to earn the miles. But the plot was that in order to long distance keep up a relationship, Adam Sandler's figuring out ways to cover the tickets.

And he discovers in the aisle of a supermarket that he can earn these miles from the pudding cups. The miles were per scan, per transaction. Individual pudding cups were coming with their own barcode. And then he figured out he could donate the pudding cups to a food bank and get the volunteers there to peel off the barcodes. But then of course he had to send them in. Because back then this was all by mail, and you'd ship them in with no real tracking, and hope that they reach there. But it did work.

And there've been all kinds of things like that. Healthy Choice was another one over the years. I had years ago, when I had a lot more hair than I do today, Bosley Hair Restoration was offering 20,000 Delta miles for a hair loss consultation. And I went in for one, my wife went in for one. She has very nice long hair. You just go in and do these things.

With British Airways miles, you could earn enough by test driving Jaguars. Because they have family accounts, you get four people's accounts linked together and you have the sales person sign four forms and you tell them, "Don't worry, I'm not gonna waste your time actually making you do the test drive 'cause I'm not buying - just sign my forms." So it was going to a Jaguar dealership to fly to Europe. A lot of these great arbitrage opportunities.

Patrick: Well, I'm sure we could continue talking on this for hours, but do have to end it somewhere. And this is as natural a point as anywhere. Where can people find you on the internet?

Gary: Well on social channels, I'm @GaryLeff and you'll find me on my site viewfromthewing.com.

Patrick: Thanks very much everyone. We'll see you on Complex Systems next week.

Gary: Thank you very much. Pleasure and an honor.