So uh, I hit the send button on the banks before Ally reported earnings, and so there’s no specific update it in that post.

I’ve got to rectify that (a) for those of you who care about the facts and the data, and (b) because we want to stay intellectually honest and open. I admire CEOs who increase their disclosure when things aren’t working exactly as expected, not CEOs who always say everything is hunky-dory when it’s impossible for conditions to always be that way.

So if you’re curious about progress on the thesis, read on. If not, I hope you liked the Citi-focused post recently and are as excited and hopeful as I am to see the bank regain its footing under Jane Fraser and her team.

Below, we’ll talk through Ally while weaving in some insights about banking and businesses in general.

Progress on the Ally Thesis

Qualitatively

I know of a guy who does a green/yellow/red light system to track their portfolio holdings.

I think the Ally thesis is working, but it’s taking longer than I thought, and I don’t think the bank’s industry position is as good as I originally thought.

So, green/yellow?

I’ve gained more appreciation for just how easy it is to be Bank of America and JPMorgan and how much harder it is to be Ally. I also think I should have spent more time researching Capital One and taken a larger position in it, splitting the investment more evenly between the two. Capital One clearly has a superior credit card franchise, and card is a great business for every scaled, mature player. All these banks are automotive lenders, too, but auto lending is one of the more competitive parts of banking.

(Remind me to talk more about credit card one day.)

Still, Ally is better than them all at auto lending, and arguably one of the best in the US. It will produce good results because:

1. Ally cross-sells more products to the dealers and gives some business tools away to them for free. That makes the dealer relationship stronger. Ally gets more (and preferential) loan flow from them as a result. They’re also an “all weather” rather than “fair weather” lender, where some lenders will bow out of the market when things aren’t good, then show up again later when they are, which strains the business relationship with the dealerships.

2. Ally’s credit analytics and collections are better than anyone else. The big banks generally try to play it safe and lend only to super-prime borrowers (people with really, really high credit scores). Those loans have low interest rates and low default rates. They have sticky, low-cost deposits,1 so they can lend at low rates and still earn a good margin. Ally doesn’t, so they have to play smarter, and they do. In our report, we showed they lend to near-prime borrowers, that big banks avoid. They then cherry-pick the best credits with really granular data that most others don’t have. Both big banks like Wells Fargo and small credit unions often paint the market with a broad brush. Ally can also be the best at near-prime lending because they’ve invested more in collections. The typical near-prime borrower defaults on their loans several times more often than super-prime borrowers, so you need to have a lot more people and much better processes to contact them and “work out” the loan or otherwise repossess the car much more often. Even JPMorgan doesn’t have the infrastructure to do this, and has no incentive to make the investment.

None of this has changed, so Ally has competitive advantages here and still deserves to win and to earn excess returns on capital, at least in this part of the market.

But… Ally has fewer avenues available to it to deploy a lot of capital at good rates of return. The money center banks can put capital to work wherever they see good growth and returns on capital. In the posts about Citi, you saw Citi has 5 major business groups, for example. There are many places for it to reinvest. The other money center banks are the same. Plus, they have loads of relationships already built, where customer demand is just kind of tepid today, but is likely to accelerate (recall that loan and deposit growth was our thesis on the money-center banks). With Ally, it needs the used car market to recover, and it needs to keep growing its distribution (relationships) in corporate finance, otherwise it won’t be able to keep reinvesting capital. It doesn’t have any other businesses.

(Admittedly, if it can’t reinvest, then it won’t grow its deposit base as much, either, meaning it won’t have to pay up as much for deposits. Which is good for margins.)

Furthermore, it’d be about as good to buy back shares with the company’s excess capital instead, but it can’t. Because its securities portfolio is ~$3-4 billion underwater after interest rates rose, it has to hold back a lot of capital until bank regulations are finalized and it knows whether it has to subtract this unrealized loss from its regulatory capital, or not. If this wasn’t an issue, I estimate Ally would already be buying back ~$1.5 billion this year, ~13% of its stock. But it can’t. So this problem created a big missed opportunity to take a lot of stock off the table at a big discount to intrinsic value. Alas :(

Quantitatively

The auto lending business is more than half of earning assets and is the big driver of our “NIM” (net interest margin) expansion thesis, where we think the “spread” between Ally’s deposit costs and its auto loan yields is going to keep widening out. This is still playing out. It should keep happening in anything but a sharply-rising-interest-rate environment.

(That said, it looks like the NIM may turn out lower than I originally modeled. Growth too. Tweaking our financial model to account for things like the sale of the credit card business and other changes, it looks like Ally can do $1.5 billion or ~$5/share in earnings in 2-4 years, the bottom end of the $5-6 we have been talking about over the last 1.5 years.)

On the deposit side: Since the Fed started cutting rates late last year, Ally’s deposit costs also ticked down to 4.01% for the quarter, from 4.23% in 3Q. The Fed’s cut short-term rates by 100bps (1.0 percentage points), so I expect deposit rates across the industry to keep falling and for Ally’s rates to fall with that. There’s a lag effect, as most deposit rates don’t just fall automatically: it comes via competition among the banks. Already, Wells Fargo said on their earnings call that they are pulling back pricing on products like CDs (certificates of deposit, also called GICs in Canada). That is the stuff that competes most closely with Ally’s big-ticket savings accounts. So we know it’s happening. We should see sub-4% rates on Ally’s deposits next quarter. In its supplemental filings, Ally also shows that its mix of deposits has been improving over the last 5 quarters (the latest quarter is on the left, and 4Q23 is on the far right):

Brokered deposits are the crappiest kind of deposits, where you don’t even do your own customer acquisition and you pay a broker to find people’s money in a competitive auction setting. CDs are the second crappiest. Then MMAs and OSAs (money market accounts and online savings accounts, respectively) are the best of the three. So you can see Ally’s deposit quality is improving, which means its deposit pricing should also look better going forward than in the past. That is good for the NIM thesis.

On the loan side, it’s a little trickier.

The company had record auto loan flow of 14.6 million applications in 2024, up 5.8% y/y even though the used car market is still soft (+~1.8% volume). Ally had successfully pushed dealers for more flow. Loan approvals are mainly automated, so Ally ended up with more high-quality loans than it wanted. S-tier loans increased to 49% of total loans, a really big bump up in the quarter:

This caused a small problem. If you have been keeping abreast of our NIM expansion thesis, you know that in this chart, we need the following to happen:

1. The purple/white bars need to go up. That’s the existing portfolio’s average interest rate. They need to go up by…

2. … closing the gap with the teal line as the loan book “turns over” into market interest rates that are higher than the existing loan book.

Clearly, if the teal line goes down, you can close the gap that way. But that’s no good for us. The loan portfolio yield won’t have risen. So, too much of that, and the “spread” Ally makes between deposit costs and car loan yields won’t go up as much as we want it to. That means Ally’s profits won’t go to where we hope they will. Which means we won’t make the money we thought we would when we bought the stock.

The company claims that the ~0.8ppt drop in new loan yields is only a little bit due to (a) falling rates in the market as short-term rates fell, and (b) increased competitive action. E.g., from following other banks, we know the super-prime market is getting more competitive again. E.g., among others, Wells Fargo has stepped back in:

Instead, Ally’s said that it’s mostly due to (c) that 49% mix of S-tier loans, where they got more than they wanted. Those loans have lower interest rates because they are to higher-quality borrowers with lower expected loss rates. That’s not Ally’s sweet spot, though, since we pointed out above that their competitive advantage is the strongest when they are lending to near-prime borrowers, where competitors don’t have the infrastructure to beat Ally.

Management said they’re already increasing pricing on these higher-quality loans so that the “capture rate” falls. Clearly, by increasing your prices, the probability the customer takes your loan offer at the dealership goes down, and the odds of it going to one of your competitors goes up. Going forward, they should see just as much loan flow and just as many loan approvals, but fewer funded loans as people take other offers instead, which will push the company back down into its sweet spot.

That will push the teal line back up.

And if the teal line stays up, then the purple and white bars will keep marching upward. The loan book, currently yielding 9%, will keep marching toward 10%. It’ll march upward while deposit rates are coming down. That means the “spread” should continue widening out.

Which means our thesis is still directionally correct.

(Or, in a falling interest rate environment, it it will at least fall slower than deposit rates are falling, so the “spread” will still widen. In a rising rate environment… well… we are screwed for a couple years if that happens.)

Management said on the call that they expect net interest margins to continue increasing from 3.19% a year ago to 3.30% today to ~3.70% in the medium-term, and this turnover of the auto loan portfolio is the main reason. That’d get you ~$5/share in earnings. Our modeling implies something similar.

Finally, the charge-off problem they pointed out earlier this year, which sent the stock -17%, is also going away. It’s from loans made in 2022 when the car market was in short supply and used car prices spiked. More of these loans are delinquent vs. what the company expected when it underwrote them, even when it tried to be a little bit more disciplined knowing we were in a hot market.

For each year of loans made — each vintage — the chart plots 30 day loan delinquencies each month. As a loan vintage “seasons,” the % that go delinquent rises. Nobody defaults on their car loan the day they buy the car. It happens over time.

The ‘22 vintage is doing worse than what the company underwrote. The ‘23 vintage is about in line. But… the ‘24 vintage is actually doing better than Ally expected. The loan book turns over into these loans. By the end of this year, ‘22 loans will be only 10% of the portfolio, while ‘24 and ‘25 loans will be ~25% and ~40% of the loan book, and most of the portfolio’s losses will be from loans made in ‘23 and ‘24, which have been performing in line with or better than what Ally underwrote. Therefore, the charge-off rate will normalize (in the absence of a severe recession).

So, in summary, it still looks like the thesis is progressing, but a little worse than the mid-point of the base case we underwrote when we bought the stock mid-2023.

— Chris