Description

We continue to believe that large-cap financials offer intriguing opportunities for long-term investors. Let me start by saying that if you believe the US banking system is permanently impaired by disruption from fintech, regulation, or a decade of ZIRP hiding massive buried credit losses, this idea obviously isn't for you. As I have discovered on several other threads here in VIC, it is impossible to argue these points with much success (on either side). Either you think the US banking system looks about the same in a decade, or you think it is doomed. There doesn't appear to be much middle ground. Submitting US Bank (and M&T before this) should tell you where we stand. There are no guarantees... for either side of the debate.

US Bank is a $682B (total assets) bank based in Minneapolis. They operate ~2,200 branches in 26 states. Last December, they completed the acquisition of Union Bank from MUFG (MUB in this writeup), expanding their presence on the west coast (perhaps poorly timed here, but likely works out over the long run). The acquisition added ~$80B of assets to the USB balance sheet. There are ongoing merger & integration costs running through the financials currently, and that will last for a few more quarters. When we talk about earnings here, we're excluding the merger costs.

The investment thesis is simple -- USB is a fairly conservative (I'll address the word "fairly" later) Midwest bank with a diversified loan portfolio and a diversified line of other banking/financial services for which they earn fees (things like payment processing, trust and investment management, credit card fees, corporate payments, etc). Based on recently released 1Q23 earnings, I dare say that you wouldn't know there had been a banking crisis judging from just the USB financials. Deposits were roughly in line with normal seasonal patterns, loan growth was reasonable, and most of their non-lending businesses are performing roughly in line with what you'd expect given current economic conditions (for example, mortgage banking fees are down, but that's to be expected given the rise in rates).

The end result is that USB likely earns about $4.50 per share in 2023 (in the VIC table above, I used consensus estimates), so the stock is trading at ~8x earnings with a 5.7% dividend yield. That P/E is low, so what is the market concerned about? Capital levels. Capital has been impacted by two things -- the MUB acquisition and the duration of their investment portfolio (they got caught up in the same duration issue that so many others are facing). As of Q1, CET1 is 8.5%. Prior to the MUB acquisition, CET1 was ~10%.

In AOCI, USB has $10.2B of unrealized losses from their securities portfolio (one year ago, that was $6.9B, at 9/30/22 it had ballooned to $12.5B). The underlying securities are Treasurys and MBS (so most of the portfolio is guaranteed by the US gov't), so this is a rates issue not a credit issue. Duration of the portfolio at Q1 was 3.8 years. They have put in place some hedges to guard against a spike in rates, but otherwise, these current unrealized losses will likely burn off over the next 3-4 years supplementing capital ratios.

This is a critical part of the thesis -- both normal earnings and the burn-off of these unrealized losses will dramatically improve capital levels. Indeed, management says CET1 will improve by at least 25bps per quarter (could be faster if rates fall, but I'm not expecting that). I'm wrong on USB if they need to issue equity to bolster capital levels. For now, I do not think that will be the case, but strange things happen in banking.

Before I get to what this might be worth, let me talk a little bit about credit quality. Earning assets totaled $604B at Q1. The earnings asset mix was 11% cash, 26% investments, and 63% loans. CRE loans are 14% of gross loans, and inside CRE, office represents 2% of USB's loan portfolio. That seems to be the asset class most worrisome to investors, and USB has relatively low exposure to office (they do have quite a bit of exposure to residential mortgages, if that is a concern this idea is not for you). The MUB acquisition required a significant boost to credit provision (they added $1.2B to expected losses during the acquisition quarter (4Q 2022) and ~$400M during the first quarter of this year). The current allowance is ~$7B, or nearly 2% of gross loans. Charge-offs are low compared to history, but moving up now that COVID stimulus payments have ended. We are modeling an increase in credit losses over the next few years.

What is this thing worth? During the Q1 call, the company guided to 2023 average earning assets of $600-610B ($607 for Q1), 3.00-3.05% NIM (3.10% Q1), total revenue of $29.5B (midpoint), and $17.5B of expenses (high end). If that is about right, the variation of expected EPS depends mostly on credit losses. We are modeling ~$2B for the year ($427M in Q1). We have no particular insight here, but are trying to find a reasonably accurate estimate. Obviously, here is another place where we'll each have our own view of the US economy over the next few years, and the resulting credit losses. In our model, we think credit losses run between $2-2.5B for each of the next five years. Then, they slowly get better -- we're not expecting much asset growth over the next 5 years either, so provisions are higher than historical levels relative to loan growth in our thinking.

Pretax, preprovision income is going to be in the neighborhood of $12B in 2023. One can work in their own credit loss estimate, tax the output, pay preferred dividends, and get their own EPS figure. We're at about $4.50.

As I said before, USB needs to increase CET1 levels. We are assuming that they'll pay out the current dividend and retain the remainder of their profits for additional capital for 2023 and 2024. We have modeled very modest share repurchases after that, but they make very little difference to the intrinsic value of the business given our view of the repurchase size. We think they'll act more prudently and continue to retain capital while growing the dividend. 

We also think NIM falls over time, but not radically so. Here is where we disagree with most bears who believe that NIMs will be far below historical levels permanently. We simply do not have that view. If we did have that view, we wouldn't own any banks. We also think that credit gets worse, but not catastrophically so. Our framework for that point of view is that this recession will be focused more on white-collar workers on the coasts (which is why the MUB acquisition may not have been ideally timed). USB just doesn't have significant risk to a white-collar coastal recession. We are more than willing to change our mind on that framework, so that's a very fluid opinion.

To wrap it up, we think investors earn a mid-teens total return for quite a few years with today's purchase price. USB is riskier than MTB (our other regional bank investment) because there is a higher probability of a capital issue at USB. That said, we think the probability is low so we believe we are getting paid for that extra risk. We have a smaller USB position (compared to MTB) because we prefer the more conservative option. 

Catalyst

Earnings growth, AOCI burn off lead to higher capital ratios, and a re-rating. This will take several years.