Some Thoughts on Banks

And a look at First Republic

Quick note: I am not a banking expert although I know generally how banks work. Thanks especially to my friend Searching for helping me along that learning process. Given I am not an expert, this post will stick to discussing broad issues. The situation is currently complex and fast moving, and instead of covering every single possible eventuality, I’m going to leave out some nuance. I’m happy to discuss via email or over twitter DMs @willis_cap

From the moment Silicon Valley Bank announced that they were selling stock, I thought we could have some issues. I hadn’t followed the bank closely but was aware that they had negative tangible book value if you marked the securities book to market. That wouldn’t be an issue given marks to regulatory capital were fine due to the treatment of held to maturity (HTM) securities, but if depositors start to leave and the bank is forced to sell securities to meet those redemptions, you can run into trouble quickly. Flipping through the presentation and seeing the deposit flight YTD signaled a real issue here. I figured the stock would be down an enormous amount but hoped they’d be able to complete the equity offering to plug the new equity hole in their balance sheet. That obviously didn’t happen.

Once it started to seem likely that Silicon Valley Bank might fail on Friday morning, I thought that it was possible that we could have a bank run on any non-systemically important financial institution (SIFI) over the weekend unless the government stepped in to provide liquidity. I have some clients with uninsured deposits at First Republic Bank, and on Friday morning I spoke with them, and we agreed to wire enough funds out in order to get under FDIC insurance thresholds.

The decision-making calculus was, unfortunately, easy. If Silicon Valley Bank did fail either on Friday or over the weekend *and* the US government didn’t step in with a liquidity backstop and protection for uninsured depositors, then uninsured depositors at any non SIFI would rush for the exits over the weekend and on Monday morning. I thought that the odds that the US government would not step in were very low although non-zero.

I heard people describing this as a classic prisoner’s dilemma, but the difference between this situation and a prisoner’s dilemma is that there is *no* benefit to keeping deposits in these institutions. Moving it into t bills gives you complete safety and a higher rate of interest. If you have money that needs to stay in deposits at a bank (perhaps operational business deposits) then moving to a SIFI also completely removes your risk. The decision to move deposits was, and is, simple and rational.

This is, in my mind, the main issue with the banking system right now – it doesn’t make any sense to keep uninsured deposits anywhere but a SIFI or in t bills. The US government has an implicit guarantee for uninsured deposits at these huge institutions, as any default by those banks (let’s say that includes Chase, Wells Fargo, Bank of America, and Citibank) would clearly become a systemic issue, which would trigger a government bailout of depositors.

This implicit guarantee gives an easy destination for any uninsured depositors deciding on where to move their money, and it furthers deposit outflows at smaller banks. I don’t know the long-term solution to this issue, but I hope someone smarter than me can figure it out.

The US government’s decision to accept banks’ securities as collateral for lending provided necessary liquidity and almost certainly prevented the failure of numerous non-SIFI banks this week. But they haven’t solved the entire problem. They have solved the most pressing risk, which is that of inadequate liquidity, but they haven’t solved banks’ other two main risks – risks to earnings power and solvency.

Liquidity risk relates to a bank’s ability to fund deposit requests. If enough depositors ask for their money back, many banks would not have enough liquid assets to meet those deposit requests. If that happens, a bank needs to be put into receivership so that the bank can sell its assets to fund deposit requests in an organized way. The US government giving banks access to the fed window largely solves liquidity risk, as banks can post securities as collateral, borrow large sums of money, and pay out customers.

But this borrowing feeds into the risk to banks’ earnings power. Most banks were paying close to 0% on their deposits, and with short-term borrowing rates around 5%, the banks that post assets as collateral and borrow from the fed are shifting part of their balance sheet from borrowing at 0% to borrowing at 5%. With enough of this switching, banks’ earnings will be seriously impaired or even negative. If these banks lose enough money, they will eat through their capital buffers and need a capital injection, sale, or other strategic alternative.

The third risk for banks is solvency risk. In my opinion, most investors misunderstand solvency risk for banks, assuming that any bank that is insolvent will fail, and vice versa. But if you mark all assets and liabilities to market (MTM) and find a bank insolvent, that won’t matter in and of itself, as long as its regulatory capital ratios, which ignore some MTM activity, are sound. A bank can keep operating if it can fund deposit requests and earn money so that it doesn’t deplete its capital base.

The main risk, as I see it, to a bank that is insolvent relates to the perception of its safety. Banking is a confidence game, and when depositors panic, the game is up. Being insolvent clearly increases the chances of a bank run in the first place.

A second risk exists for an insolvent bank for investors that own the company’s securities. If deposits flee and the bank is forced into a sale, potential buyers will look at the balance sheet on a MTM basis, and if it is insolvent (or even just undercapitalized), the bank will require a capital injection to reinforce its equity position, sometimes making a deal to sell the bank impossible as the earnings of a reorganized bank may not be high enough to justify the necessary capital injection.

I think these issues can be illustrated in real-time if we look at First Republic, which finds itself at the center of the crisis. Full disclosure: I bought First Republic stock a few years ago during the market drawdown during covid. I sold some of that stock as it went higher over the next couple of years, although I didn’t sell it all at the highs (I didn’t want to pay all of the taxes. Dumb.). I still held some shares into last week before selling the rest of them and I currently have no position. All that to say, I’m quite familiar with the bank as a former investor.

The most pressing issue for a bank is always its liquidity position. Looking at First Republic, we need to know the bank’s sources of funds if depositors draw down balances. We don’t need to look at their balance sheet too closely to answer this question (although we normally would), as they have secured significant liquidity from a credit line from JPM, they have access to borrowings from the fed, and they have received an emergency $30B of additional liquidity in the form of additional deposits from the nation’s big banks.

There is a worry that even all that liquidity might not be enough as they don’t have large amounts of the right types of assets to use as collateral for the fed borrowings, but I’ll shelf that for now as the $30B in deposits plus the credit line from JP Morgan should be enough. Let’s just assume that the liquidity is there so that we can examine the next issues.

The next issue is the bank’s earnings power. From First Republic’s recent press release announcing the emergency deposits from many of the nation’s largest banks, we can see that the bank has lost *at least* $30B of deposits given disclosures on incremental borrowing from the FHLB and the fed.

How would their earnings power look with this new liability mix? Based on their YE 2022 balance sheet and some estimates on incremental loan rates and deposit costs, before the deposit flight I had estimated that they would earn ~$3.5-3.75B of net interest income (NII) for the year. Analyst estimates were much higher, but this included benefits of growth from new deposits and loans, which will now not be happening given the current environment.

That level of NII would lead to ~$600-800mm of pre-tax earnings with an incremental $160mm of preferred stock dividends below the earnings line. Knowing that at least $30B of deposits have fled, we can say that at a minimum their EBT would drop by $1.2B ($30B multiplied by the difference between the old funding costs of ~1% and the new one of ~5%), and likely somewhat more as that $30B number is the absolute minimum. With these incremental costs, the bank would likely be losing at least $500mm / year. First Republic’s tangible equity base for regulatory purposes is roughly $14B, so they can’t go too many years of losing significant sums of money before they’ve eaten through their regulatory capital.

Many people have said that to stem this deposit flight and stabilize the franchise that another bank could purchase First Republic. This is where solvency risk comes into play. If you mark First Republic’s securities and loans to fair value, tangible book value goes from $14B to roughly *negative* $10B. A potential purchaser would need to inject $10B to plug the equity hole and another ~$10B to recapitalize a business (somewhat lower than the $14B of current capital as we’re assuming the balance sheet has shrunk) that would earn roughly $1.5B per year in normal circumstances.

The future outlook for First Republic depends mostly on the outlook for deposits. If depositors decide that all is well and they can move deposits back, then they could replace the current expensive borrowing with lower-cost deposit funding, move back towards profitability, and eventually earn higher returns once old loans at low, semi-fixed rates start to mature.

If deposits stabilize here or keep flowing out, then they will almost certainly keep losing money and inch closer to eating through their capital cushion. My personal guess is that those deposits that already left are gone for good and they aren’t coming back as there’s no real incentive to do so.

So if First Republic stays operating as is, they will likely see more deposit outflows, which they’ll have to fund with expensive borrowings that will lead to more and more losses. If First Republic tries to cut expenses, they’ll hasten deposit outflows as the fact that they’re a high-touch bank is the primary reason to bank there in the first place. If they try to sell the business, the buyer will need to inject roughly $20B to recapitalize it (even if they were to purchase the equity for a nominal sum), while the earnings would likely not be high enough to justify the amount of equity injected. They could try to let the asset side of the balance sheet run down in order to delever, but loan paydowns are relatively low given negative convexity and relatively long duration in the loan book. They are in a tough spot, and I hope that there’s some sort of way to resolve this besides what the situation clearly implies.

I think we can view the challenges for other non-SIFI banks through the same lenses that we used to examine First Republic. First Republic has longer duration in its securities and loans than most banks, but you can see similar issues. The current government borrowing program will solve temporary liquidity issues, but any banks that replace enough low-cost deposit funding with these borrowings will see significant hits to net income, while many of these banks would need fresh equity capital in order to recapitalize them if they were to try to sell.

The extent of the pain is primarily a function of the outlook for deposits. Using round numbers, a bank might earn roughly 1% on its assets, so if 20% of their liabilities reprice from ~0% cost to 5% cost, earnings power will be wiped out. If they’re forced to pay somewhat higher rates to keep some depositors from leaving, then you will have similar headwinds to NII and earnings. This brings us back to the implicit guarantee for the SIFI banks. That issue is not solved, and truthfully I’m not quite sure how it can be solved. Rationality isn’t the only thing that drives business decisions, but I’m not sure how tenable the business models for these entities remain given an environment where rational depositors would not keep uninsured deposits with them. I hope that regulators and policy makers can figure out how to fix these issues, but I’m not holding my breath.

Note after publishing - it has been reported (although not confirmed) in the WSJ last night that First Republic has seen $70B in outflows. If this is true and these outflows don’t quickly reverse, then I don’t see how the bank can continue. First Republic would be losing something like $150-200mm / month. I will also mention that First Republic’s news release said that the $30B of deposits from the nation’s big banks came with terms “at market rates.” This is very vague but if the rate is anything less than short-term risk-free interest rates, then that’s a direct transfer of money from the big banks to First Republic. I doubt that that’s the case, but I might be wrong. Unfortunately, if the $70B number is close to correct then it will not matter as incremental borrowings of $40B at ~5% will swamp earnings power even before factoring in some costs to the $30B from the big banks.

Thoughts for My Investment Process

As a long-only, buy and hold equity investor who still owned some First Republic stock coming into this, I have to say that I now find bank stocks uninvestable. Others have understood this better than me and I give them kudos. People will look at First Republic for the first time in the wake of this and talk about how obvious all of this was and how poorly the bank was managed. I think that’s unfair – First Republic served its customers well, grew its business over time, earned enough money to fund its growth, and stuck to writing sound loans. People will call them stupid for keeping the duration of its loans and securities portfolios too long, but that’s not really what killed the bank. I can almost guarantee you that those depositors who pulled their money weren’t studying the financials and saying, “oh if I MTM the loans then there’s a $10B equity hole here, I need to pull deposits!”

What really happened was a bank that, if you squint, looked kind of like First Republic failed and uninsured depositors got scared. First Republic was always a high net-worth bank, so the implicit guarantee on deposits at the SIFIs hurt them especially hard. They had the wrong deposit mix for the wrong type of crisis, and the bank was impaired in one fateful day. Lower duration on the asset side of its balance sheet wouldn’t have changed that.

You could argue that if you own the securities of a bank with a low mix of uninsured deposits then your risk of a run is much lower, and you would be partly right. But I can think of situations where even a bank like that could see a run. The bottom line is that the health of a bank is always a function of the confidence of its depositors, and depositors can get spooked for any number of reasons, instantly evaporating the value of that bank.

Now, I was aware of the uninsured deposit mix, the negative equity if you MTM the loan book, pressure on NIM given rising deposit costs and all the rest. If depositors hadn’t rushed out, First Republic would still be able to go about business as usual. Earnings would be down somewhat but over time they would be able to improve NIM and go back to earning decent returns. But it isn’t business as usual. Depositors pulled money and the bank is in crisis.

Going forward, I find it impossible to hold the equity of any non-SIFI bank (I don’t want to own SIFI bank stocks either, but that’s a story for another day). All banks have to take some credit risk or some duration risk (or a mix of the two), or else they won’t be able to earn money. Right now there’s a crisis for those that took duration risk. At another time, there will be a crisis for those who took credit risk. Even if banks can still earn money, the franchises will be instantly impaired if depositors lose confidence, and my lesson from First Republic is that it's very difficult to confidently and accurately predict the banks that will never have depositor panic.