May 31, 2022 - 8:29am EST by
2022 2023
Price: 1.00 EPS 1 1
Shares Out. (in M): 1 P/E 1 1
Market Cap (in $M): 1 P/FCF 1 1
Net Debt (in $M): 1 EBIT 1 1
TEV (in $M): 1 TEV/EBIT 1 1
Borrow Cost: General Collateral

Sign up for free guest access to view investment idea with a 45 days delay.

* Idea not eligible for membership requirements


Quick synopsis

I recommend a short position on American Express. My view (shockingly) is that the company's financials have deteriorated significantly over the past ten years due to the squeeze of the competitive environment and discount pressure killing margins. What many investors believe to be (and is heavily touted by management to be) a spend non-capital intensive earnings power has become essentially just another lending business in decline. The company's non-interest income declined 90% by 2021 from its $6.4 billion peak in 2007 because of a 1700 basis point increase in reward/variable costs. The most significant decline happened in 2020-2021. They are currently at 2-4% margins on their non-interest revenue base, which means a 5% increase in rewards as % of their fees causes them to dip negative besides interest income. They are trading at 20 times 2021 ebit, which is entirely from interest income. 

Management just guided 2023 to higher than mid-teen eps growth and raised LT to mid-teen eps growth. There is NO chance they will hit their 2023 or LT targets; in reality, their income will start dropping soon. In 2022 they expect variable costs to rise another 200 basis points, and non-interest margin for 2021 was 2%-do the math-all income derived from interest... Their interest yield is at a peak and will probably decline by 20% in a normalized environment. Their non-interest income will dip negative soon (perhaps 2022), and the top line will contribute to losses as its margin is negative (operating leverage is stretched thin from the 2017-2018 restructuring). They will have to increase their loan book by 40+% annually to hit their eps targets, something they could not do at even 5% in the past.


The thesis is predicated on a couple of things.

  1. A way to high valuation.

  2. A misunderstanding of the business earnings profile by investors.

  3.  They will for sure miss their growth targets soon (2022 is debatable) and will have to reset expectations.

  4. They are very liberal in their accounting practices, with about $900 million in 2021 of reduced expenses from netting non-realized gains and reserve releases (billions taken out like that in the past). In addition, there was an undisclosed reserve release in the 1st quarter where their income was higher than expeted because of lower OTHER EXPENSES (the most suspicious line). They have large reserves disclosed as other ($10 billion+), and they determine their reward liability and costs because of assumptions embedded.

  5. Their income will start declining soon as interest yields normalize and non-interest income gets squeezed more, leading to secular decline valuations.

  6. They will perform poorly in a recession.

To give a layout of the write-up, I will discuss the following topics in the following order: 

  1. General view on the deterioration of the spend business,

  2.  A more granular financial review of the different segments in the business,

  3. Interest income and the threats to it

  4. The competitive environment and how it may get worse.

  5. Short term financial outlook and valuation

  6. The story of the mis-selling by management of the Amex story and their efforts to hide the business deterioration.


The general story.

2007 was, in our opinion, the peak of American Express's business. Amex generated about $6.4 Billion pre-interest income and an ADDITIONAL $880 million in interest income, bringing their income to $7.28 billion and was, in reality, a spend non-capital intensive business. From there, the GFC hit, and their non-interest income hit a low of $3.5 billion in 2010 attributed to the crisis. 


Amex fared the best in the industry with higher quality loans and a small balance sheet and recovered the fastest, with loan losses peaking lower and sooner than the industry. They recovered to a post-GFC peak of $4.6 billion in non-interest income (down almost 30% from peak) but generated a higher $3.1 billion interest income for a total of $7.7 billion. 


This recovery happened while the banking industry was in tatters; however, two shifts were happening. First, the multitude of new laws governing what banks could do to generate revenue together with interest rates at zero, was essentially killing a consolidating industry’s earning base, and they were desperate to get it back. One of the most effective solutions to this was credit card volume which translated into credit card loans.

 Credit card loans are extremely valuable to banks because the competition does not revolve around rates- after all, no one thinks they will rack up balances. The critical selling prop is their rewards financed through interchange fees which banks started giving aggressively. Their willingness to give back all interchange through rewards came because banks did not care much for the interchange revenue because that's not how markets valued them, and it was not their business. The whole point was to get the high ROI loans and the sticky bank relation.


 In addition another change was that Visa and Mastercard had recently gone public and were now pushing hard to grow credit volume on their networks, of which Amex held a US 25% share in 2007. Before this, they were essentially just a partnership between different banks, did not have a separate interest in growing volume, and had a more challenging time raising interchange fees.


I want to quote Ken Chenault from his 2016 investor day, where he himself laid out the environment: 


"Competition across payments has also accelerated. Margins are being compressed at the customer and merchant level in a number of businesses and across several markets, most visibly in the US where both issuers and networks are looking to expand volumes and are willing to give up economics to do so.

On the issuing side, a number of broadscale banks have seen returns in some of their financial businesses decline. As a result they've reallocated investments to their payments business which typically generates higher returns. This has made the competition for premium customers more intense with issuers offering ever richer level of rewards to acquire and keep customers.

The competition has been particularly evident in the cobrand space. One change in this segment has been the bid process itself. Competition for some cobrand accounts in the US and Canada has become more aggressive due to the solicitation of independent bids from issuers and networks."


‘’On the issuing side, it's clear that competition is more heated because of the focus on growing balances and cobrand portfolios have loans that can be bought with a high enough bid. Increased competition for cobrands was a marketplace change we saw coming.’’



"Regulatory activity has also increased. The implementation of new regulations across the EU is putting pressure on our discount rates. The regulatory activity in Australia is intensifying.

The US dynamic is different. It's a large market and as you know with very intense competition.

As I mentioned earlier, many issuing competitors have been willing to reinvest all their spend revenue into richer rewards to drive higher loan growth.‘’


‘’’This is changing the industry dynamics, and because of this, I believe success in our US consumer business will depend even more on the growth of lending. We've done very well at growing balances and have good momentum."

Starting from 2013, their business went on decline because of those reasons laid out: their non-interest Income dropped practically consistently. They hit $3.8 Billion of non-interest income in 2018 (down over 40% from 2007 peak with volumes way higher). From there, the continued competitive environment and their new CEO Steve Squeri's focus on not losing volume to Visa and Mastercard caused a super steep drop to a low in 2021 (yes, worse than 2020) of $645 million in non-interest revenue (down 90% from 2007). They made up the balance from $5.3 Billion in interest income from a record low write-off rate. They essentially made $6 billion of ebit in 2021 (what we believe will drop in the future) or $7.9 ebit per share, which practically all was supported by loan income. The stock trades at over $160, over 20 times 2021 ebit on a loan-funded earnings power.


So what exactly caused the company's profitability decline? 

It came from increasing rewards-member service costs and marketing costs. In addition, it came from an increasingly competitive environment for cobrand cards where the co brand partner pushed harder for the economics, and the issuers obliged.

The top three variable cost lines, including marketing, rewards, and rebates to cobrands (includes some other things), went about 1700 basis points of revenue from 2011-2021 to 63% of revenue with a consistent uptake. In 2016 margins on non-interest revenue to costs were at an 11% profit margin. In 2021, after a 12-point increase in variable costs, the margin was at 2% due to much expense tightening. In 2022 they expect variable costs to rise over two points-you can do the math. Without getting into details, investors largely ignored this because their balance sheet business has done so well in the past 5+ years.

Now besides competition, there was another thing that was pressuring margins heavily-Discount rate declines. Amex Discount rate went from a 2.58% average in 2005 to the current 2.3%. The story management sells is that as long as volume increases faster than the decline in the rate, their revenue base is intact. The issue is that the company has enormous variable costs that increase primarily as % of spending (or members). So what you get when you have rewards going from 1.5% cashback with no sign-up bonus and a 2.58% discount which becomes 2% cashback and a $200 sign-up bonus and a discount rate of 2.3% and a more competitive marketing environment is unit economics in the red.

The cause of the discount decline is from multiple places, including their work on getting smaller merchants. They did this through a program launched in late 2013 called Optblue. It essentially gave acquirers who historically were just a salesman for Amex to merchants, and they now gave them a wholesale rate where the acquirer then marked it up to the MDR-essentially Visa and Mastercards business model. In addition, the EU and Australian interchange regulation, although not directly applicable to Amex, caused competitive pressure on their rates. Then was the general forces from scaled companies who can negotiate separately with Amex, whose relevance was falling and now holds only a 7% market share in worldwide payment volume against its 13% in 2006 with their US share moving down too. Then, the last aspect that was pressuring them was the decline of spending from T&E, where their discount is higher.


Now to get a bit more granular into the different businesses here and their respective declines.


The consumer group.

This segment first turned negative on non interest earnings in 2015 as Costco ended Amex issuing and exclusivity. Of course, the cause was about 1600 ADDITIONAL basis points from variable reward-etc costs in 5 years, rising from about 59% of non-interest revenue to 75% in 2019. This would have been deep in the red and killed the spend-centric story on the consumer group first. They, therefore, reshuffled corporate expenses out of this segment, reporting a positive non-interest income of $700mm in 2017. However, that fell apart as the decline continued, and they reported a billion-dollar loss on non-interest in 2019, almost a two billion dollar income decline from peak. Essentially all the reported top-line growth that investors gush at results in more losses unless balances go up!

In 2020 as the segment continued to deteriorate, management decided to change the revenue mix among segments taking high margin revenue from both the business segment and, more importantly, their network segment and allocating it towards the consumer segment. This made the billion-dollar loss in 2019 become a $200mm loss. However, this failed too, with the segment hitting another peak loss of $750mm in 2021 even with all the shifts.


The commercial segment.

The next segment is not as bad as the consumer group but is catching up. Structurally the competition is less fierce in the commercial segment because it is harder to get businesses to keep high-cost credit card balances and is, therefore, less attractive to banks. However, that does not mean they are not going after the business at a slower pace. In addition, although the competition in this space is more manageable, the gap between payments/rewards on the business cards cannot fall too far behind the consumer cards because they are relatively interchangeable in the more profitable SMBs.

In the four years from 2016-to 2019, the segment contributed about $3 billion in no interest profits. The 5+% points of rewards margin squeeze were mostly offset in revenue growth, leaving it a profitable but no growth segment. During covid, that changed dramatically as the competition heated up with the company giving up 600 basis points of margin from rewards in ONE year. This caused the segment's non-interest profits to plummet to under $1.5 billion in 2020 and 2021, giving up about $1.25 billion in earnings. Non-interest margins in the segment now have descended from the 26+% they had in 2016 down to 15%, almost a 50% decline. This segment is critical because if it turns red, the company will be in a dire position to keep up its story.

The merchant acquirer/network segment.

Last but not least is the golden negative growth not talked about-part of their business-their Network and merchant acquiring business. The 3rd party network part, we believe, is less than a billion-dollar revenue business; this is a no-growth business. The reasons here are multifaceted; the first is simply that it is a harder sell to banks competing fiercely with Amex. In addition, in many international markets, specifically Australia and, more importantly, the EU, there is the regulation on 4-party payment rails, and Amex had to exit this not to get classified as such. In addition, there is a constant competitive pressure to give up more incentives and rebates in pricing in this business, with Visa and MasterCard incentives as % of gross going up 1,000 basis points in the last decade. Now for them, this is offset because the business is so leverageable that their revenue growth makes up for this with little margin pressure; however, for Amex, this puts margin pressure because they have no top-line growth here.

The rest of the segment is their merchant business which is just a slice of the discount revenue. In addition, this includes payments to both processors and merchants for taking and perhaps encouraging Amex card usage. So many pressures here are causing this highest margin segment which was relatively stable on its profit, to get walloped from 2019 and onwards, taking out some $700mm+ from its $2.3 billion profit in 2019.


Their interest income dynamics.

So now that we have established that the non-interest spending part of their business is gone and will likely dip substantially negative, the focus becomes the balance sheet business-Lending.

There are three topics to address here:

  • The ability to grow balances.

  • The amount they will make on their spread.

  • The losses from write-offs.

First, the most important is Amex's historically hard time getting consumers to use their cards for revolving balances. Amex's key target (super prime) causes this because they carry a smaller amount of lending than others. In addition, Amex generally captures a smaller portion of their revolving balances relative to their spending share. After all, that is not their bank. In general, although all in all balances for super-prime are growing slowly, the balance per revolving card is relatively stable because AND more importantly, the share of revolving cards has been shrinking over the last couple of years. The reasons for this can be debated, but I believe (though not core to the thesis) that it is because credit cards are an awful way to borrow, and lenders see that. The new fintech wave is looking to take out every part of personal finance they can get their hands on, and credit card debt is the biggest target, especially for super-prime. BNPL and marketplace lending are prominent examples of areas growing very fast and primarily from credit card debt. Another big issue here is that co-brand cards historically for Amex have been a much bigger % of loan balances against spending share. These are under massive pressure as retailers try to keep much of the economics.


Regarding the spread between funding costs and interest revenue, Amex worked here a lot over the last decade. In reality, Amex is at a disadvantage because it must fund more than double its lending book because of accounts that are not interest-paying but can be held for two months. Historically they had a beta on funding higher than their revenue beta. However, post the GFC, they sustained an effort to build a proprietary savings account base to lower and diversify costs. In 2020 they accelerated this hugely by doing what they said they never would and acquired Cabbage and used their checking capabilities to launch a proprietary one. As a result, their funding sources went from 25% proprietary accounts and 60% debt in 2012 to 53% accounts and 33% debt saving them all 2+% in net interest costs. This was done, in our opinion, because they understood the need for interest to hit bottom-line objectives. Even with all the recent moves here, they believe they are still negatively affected all else, even from rising rates.


Now in regards to the loss rate here, too, they are at a disadvantage because they need to absorb about 1% of the loan book in receivable balances. Now covid pushed write-offs to historical lows from a historically low write-off decade down to 1.2% for Amex against a decade average of 2.3%. However, this will not last and will likely worsen as the macro environment is unfriendly. In addition, they will have to keep funding a 1.1% loss on the receivables segment, so a 3.4% average loss ratio against a likely 10.5% net interest income nets about 7% net interest income, which is better than that pre GFC but much lower than the 8.75% seen in 2021. In addition, another interesting note is that Amex losses as % of JP Morgan, Capital One, and Citi jumped from 2018-2020 to an average of 72% from 50-55% averages in the past-(are they compromising on their core?).

Is the business under secular pressure?

Now I would like to argue that not only is the business not as high-quality today as it was, it will get even worse, and we are dealing with a secular pressured industry. A passive viewer would see the industry getting increasingly more competitive both in the consumer and more importantly the commercial SMB spaces. Chase is offering for the first time a $750+ sign-up bonus on its no annual fee business card- ludicrous and more generous than the historical $500 Amex never competed on. In addition, Citi was the only Card offering 2% cash-back against the industry norm of 1.5% in the old days. Even Citi never offered sign-up bonuses for this. However, now many companies are offering this, most notably Wells Fargo, which has never been a fierce competitor, has started a 2% plus $200 cash back card. Amex, who hates cashback because it is purely rational, has not matched this yet but likely will. The economics from a spending view is negative here with the 2% giving back practically all the discount and dips negative when considering other Card services-marketing -sign up bonus. The unit economics here are terrible unless you get good at lending.

Another prime example is Capital One, which, IN 2021-2022, revamped its business cards with the no annual fee of 2% spark cashback with a $3,000 sign-up bonus. In other words, the economics on the Amex Commercial segment is way too high to support competitors who are giving up every dollar of their fee revenue as rebates.


Now to quote some of the emerging and existing competitors.


First from the WFC CEO at the end of 2021," Our rewards were not competitive. Our customer service was not competitive. Our credit lines were not competitive, our fraud experiences... So what we said is, let's figure out what's going on here. And we've renegotiated our deal with the networks, moved primarily towards Visa, basically taken the economics that we've received, put that back into the product, etc.."

And from JP Morgan some very interesting comments from the most recent investors day on business cards and in general-’’we're pursuing several areas for growth. While we have a healthy 17% share in business card and we're a clear #2, our #1 competitor share is decently higher and their share is our opportunity. To that end, later this summer, we are launching Ink Business Premier, a new pay and full product designed to meet the needs of larger businesses. This is testing well with our customers today. So we're confident in the product-market fit……less than $7 billion of gross cash marketing across CCB, up $1.4 billion over the last 3 years. As you can see, the lion's share of the spend and the growth is in Card………the credit card industry and Chase within it have undergone radical transformation over the last decade. On the bottom left, fierce competition in the premium card and travel space, in particular, has driven cost per account up and net interchange materially down as more value has been introduced to attract and engage premium customers. And the competition for scale and distribution of a strategic co-brand relationship has also been heightened. 

And from Capital One on their venture X launch.

"Pricing continues to be mostly stable. Rewards offerings have definitely become richer, and we continue to watch that closely. We saw some modest increases in upfront bonuses, mainly in the form of limited time offers, especially in the travel space. The earn rates for rewards have increased with some new product structures introduced recently, particularly in cashback. And then you also, let's not forget, have the fintechs and the buy-now-pay-later effect and various things that are going on there….. This is really a matter of being all-in after that marketplace. A subset of all the big card players have chosen to go after that business intensely. Capital One is one of them."


Besides increasing competition, they will have to increase rewards to stay competitive as things stand today.

JP Morgan currently spends about 78% of all card income (and merchant processing) on rewards and rebates up since 2019 but not hugely. This is against AXPs well under 70 just in the consumer and business segment is running well under 50. This is likely another 2,000 basis points increase just to hit parity with their biggest competitor, who pulled in front of them on the consumer side and is trying on the business side. This is besides the fact that JPM themselves are ramping it up significantly. Capital One, we estimate to be at high double digits on rewards payouts over Amex and is well in the red on the card business pre-interest revenue. Citi gives back over 100% of its credit card fees in rewards and partner payments, again hugely higher than Amex does. 

Now there is a general perception that Amex has the highest interchange fee, allowing them to have better economics. That is not so true anymore, as Visa and Mastercard have raised their rates over the years and Amex lowered. It is hard to make an exact comparison of Apples to Apples here. Still, Amex's average discount is 2.3%, and Capital one had in 2021 an average interchange rate of 2.25% (besides network rebates). They are generally in lower-tier cards, and I believe Chase is likely higher with the highest Visa e-com discount at close to 2.5%. 

All in all, Amex still has room to go in keeping their share, and the ambitions they outlined in their last investor's day about their top-line growth and share growth are laughable. Unless they ramp spending higher, especially as the business platform comes under attack. Another issue is that banks gain more from credit cards than their economics. It is the easiest way to get a consumer onto the platform, which lowers CAC costs by upselling them all their products that superbanks are doing.

Key here is that all you need is competitive pressure to add 500 basis points in reward spending for the non-interest revenue to dip negative and their top line growth means more losses.


Financial projections and valuation


In the first quarter of 2022, their margin on non-interest revenue went up a bit to 5% (from 2021 2%), driven by operating expense leverage. This leverage is highly suspect as other expenses dropped significantly from previous quarters and they disclosed a reserve release. However, besides that, historically and per management guidance, they expect reward expenses to go considerably higher as the year progresses. We, therefore, see this as peak earnings of about $480mm from non-interest revenue and believe they will generate significantly lower income for the rest of the year. The company is guiding towards about $9.5 in eps for the year or roughly $7 billion in post-tax earnings. If you assume they can keep the $480mm of non-interest ebit for the rest of the quarters (which we do not believe), then they net about $1.9 billion from that and must make up about $7 billion in pre tax earnings in interest income. Current quarter average loans are about $88 billion. Assuming normal step-ups, average year loans should be near $90 billion; the company would need to hit average yearly yields of 8% net of all receivable write-offs to hit the annual eps target in the mid-range. 

The thesis does not stand on them not hitting this year's targets as there is a good chance of that. However, much has to go right for that to happen. First, write-offs must stay near pandemic lows; second, they must keep their non-interest income which we believe will doubtfully happen. The chances of both write-offs increasing and non-interest income going to 0, is high, but it's their 2023 and LT targets that there is no chance they hit.

Management is projecting 10+% revenue growth LT and mid-double-digit eps growth. They are currently at a 5% margin on non-interest income if and when they increase rewards and marketing by 5+% of fee revenue base, they will be at negative ebit pre interest. WHAT THAT MEANS IS REPORTED REVENUE GROWTH WILL BE MEANINGLESS BY THEN AND ACTUALLY MEAN BIGGER LOSSES. They will have to make up for it in interest revenue which is at a peak in its yield. Therefore, they will have to grow their balances 20-30% a year to hit earnings growth, something they have struggled to hit 5% in the past and are now under bigger pressure. In addition, management's change in projecting 10+% revenue growth over historical mid-single-digit will only be hit by Xtreme competitive pushing, which will cause them to go non-interest negative in significant numbers.


The current valuation is ridiculous and comes from the perception of their business. The current market cap stands at higher than $125 billion. If they hit their target this year, they will likely have an ebit of $1.8 billion (if you are super generous) in non-interest ebit. If you assume, that is a stable cash flow over the next many years (which it is not even next year) and put it at a generous 15 times ebit, and it is worth $27 billion. Then they can generate on this year's projected loan balance a normalized 6.3 billion in ebit at a generous ten times total company is worth $90 billion, a 30% decline at generous valuations. In reality, they will likely dip negative soon, and the ebit will probably fall significantly under $6 billion a year. A secular decline (to a point) balance sheet cash flow where new technology is running after it is worth five times ebit, so $30 billion is at a non-dramatic bear case. There are significant catalysts to bring out there-rate from potential macro loan issues. More importantly, the company is missing guidance over the next two years and needs to reset expectations.

Amex tried keeping the narrative going.

Now Amex tries very hard to keep up this narrative that they are a spend-big moat business. First, they have always reported gross revenue, including rewards, while almost everyone in their segment classifies it as net. In addition, they changed in 2018 to include anything they do contra and started reporting it as gross. They also changed allocation in 2019 between segments to make their consumer segment better than it was.

Now another thing they do is consolidate many expenses into other expenses, including gain on sales which get Netted out of expenses. That is ludicrous with them taking out about $750mm in expenses in 2021 because of unrealized gains (marked to market themselves) from Amex ventures and have not written them down in 2022 yet. We estimate that they netted out almost $900mm of expenses in 2021 from gains/and reserve releases. In 2022 they disclosed that they released a reserve which got netted out of expenditures but did not disclose the amount. We believe that is likely one of the reasons other expenses dropped so much in 2022 and explains the higher than expected profits.

There are many ways to play around with numbers here, with both reward liabilities up in the air. They disclose that they did not change the assumptions on the amount that will be redeemed (96%) in the past few years. However, the cost per point has not been disclosed or any changes. In addition, they have $10 billion as reserves classified as other. This includes basically whatever they want and can be used to get their desired numbers. We are still researching any precise earnings manipulation but believe there may be some.

Management's preaching about their closed-loop system is ridiculous and does not warrant discussion.

Now I would like to quote their CFO from their 2019 investor's day, which is a typical way of mis-selling their story.


"Now as Anna showed you on an earlier slide in the context of the commercial side of the house, when you look at the company overall, we remain heavily spend and fee centric, 81% of our revenues coming from spend, fees. But I also know that I talked to many of you in this room who say, yes, I hear that Jeff. But boy, when I think about the growth you're getting in lending, what does that mean for the future. Let me just remind you of the basic math given a very small portion of our revenues that net interest income is today. If you go back more than a decade. In 2006, net interest income was 19% of our revenues, about where it is today, 12 years later. And because we're getting great growth across all 3 of our revenue lines, if you think about the next decade and project out a decade and let's assume, for a second that net interest income, as well as card fees grow at about 150% of discount revenue, it actually makes very little difference. Our model remains very spend and fee-centric."


I do not hold a position with the issuer such as employment, directorship, or consultancy.
I and/or others I advise hold a material investment in the issuer's securities.



Missing targets and resetting expectations.

    show   sort by    
      Back to top