Description

I own shares in two illiquid US community banks that I am very excited about.

1) The first one is a California bank with a conservative management team. The bank is owned 50% by a charity that is reliant on the firm’s dividend payments. The bank has very low deposits costs and good fee income through an attractive trust business. It also holds a large securities portfolio that is underwater and marked as held-to-maturity. While there are unrealized losses on the securities, the company’s low deposit costs will still facilitate a healthy ROE. Last year, the bank earned a 12% ROE. Given the current environment, I project that the bank’s funding costs will rise from basically 0 and that the ROE will fall to 6%. The stock seems excessively discounted at 44% of its $363m Q1 tangible book value. Put simply, you are buying a 6% ROE at 0.44x of book value.

2) The second bank also has very low deposits and good fee income. They recently invested a lot of liquidity into securities yielding 4%. Its combined loan and securities book is generating a yield of 4.6%. Last year’s funding costs were below 0.1%. Assuming this goes up to 1.0% in 2023, net interest margins will still be 3.6% and result in an ROE of 23%. The company is trading at only 3.4x my estimate of 2023 earnings. Put simply, you are buying a 23% ROE at 0.69x of book value.

Ok now for the big reveal: they are the same bank, Exchange Bank (EXSR), under two different accounting standards. In the first example, I use held-to-maturity accounting for the securities book. This is the accounting that SVB used and that most large US banks use. I assume most people on this board are familiar with this method. In the second example, I fully marked to market the securities book. This is called available-for-sale accounting and most community banks already use this. However on top of this I also mark to market the income statement by using the yield to maturity that is implied by the marks on the balance sheet within interest income. This accretion is normally accounted for within other comprehensive income (OCI). I believe this is the main disconnect I currently see in the market. Most market participants use mark-to-market for the balance sheet, which lowers book value but they forgo using mark-to-market accounting for the income statement where it would increase interest income, net interest margins and net income.

The market is ignoring the accretion of OCI losses

A simpler way of illustrating this is the following. Imagine that Exchange Bank (and all other banks with underwater securities) would sell all of their securities at current market prices and then buy them right back at those same market prices. If they used AFS accounting, book value would not be impacted. If they used HTM accounting, book value would decrease. However, in the following reporting period, interest income and net income would increase under both AFS and HTM accounting. Would this sale and repurchase increase the value of the bank? I don’t think so.

But SODAI!? You don’t care that EXSR management lost $141m (post-tax) in 2022 by buying long-dated securities at rock bottom interest rates?? I do but this mistake is already captured in book value. And given that most market participants thought that we were in “lower for longer” environment, I don’t think they need to be punished beyond that decrease in book value. One could even argue that P/B should increase since the lower book value will dramatically increase the ROE. Below you can see 2022 earnings and my estimate of normalized 2023 earnings under these 3 different accounting standards. If interest rates were to stabilize, the three accounting treatments would eventually converge as the AOCI losses run off. But for now, I think the far-right column is the best illustration of current earning power.

Overview of Exchange Bank

Ownership: EXSR has been written up twice before on VIC so I will keep this brief. The bank opened its doors in 1890. In 1948, the son of the founder, Frank Doyle, put his ownership stake into a perpetual trust to assist in the education of “worthy young men and women attending Santa Rosa Junior College.” Since then, the scholarship fund has provided over $100 million to more than 140,000 students. Today, the Doyle Trust still owns 50.44% of the bank and is reliant on the bank’s annual dividend to fulfill its mission. The bank doesn’t file with the SEC but provides financials and an annual report on its website.

Branch network and trust business: The bank operates 18 branches in Northern California, primarily in Marin and Sonoma counties. The bank also owns an attractive trust administration and investment management business. In 2022, the bank generated $24m in fee income, equal to 25% of net interest income. So from a fee income perspective, EXSR resembles more the larger US banks than most community banks.

Very low deposit costs: In 2022, the bank paid $3m in funding costs, equal to 0.07% of deposits. As of march, 38% of deposits were non-interest bearing. 68% of all deposits were fully insured by the FDIC. The low funding costs can be attributed to the bank’s long history, its branch focus, an attractive customer mix between consumer and small business accounts, a large amount of checking accounts, a lower amount of money market accounts, and management that is not interested in chasing deposits. 

Securities portfolio ($1.6bn): The bank holds a very large securities portfolio of primarily agency MBS, GSE debt and municipals. In 2022, EXSR reported a pre-tax loss of $200m ($141m post-tax) on these securities within other comprehensive income (OCI). This represents a loss of 11% on the portfolio and implies a duration of 4 to 5 years. Management told me the portfolio has a duration of 4.5 years. Small management mistakes can be deadly in banking so I can understand when investors are skeptical of EXSR as a result of this mismanagement. However, I am making a bet that this was a one-time error and that it is not foreshadowing larger problems. And given that these losses are fully captured in book value and market cap, I view the accretion of these losses over the next 5 years as a major hidden earnings stream.

To be conservative, I assume the portfolio has a duration of 6 years (MBS will have longer duration due to lower prepayment rates). In 2022, the portfolio generated an income yield of about 1.7%. So assuming a duration of 6 years on the portfolio losses as of Q4 ($200m ÷ 6 years) implies that at current marks, the portfolio is generating an additional $33m in income. This would equate to a total interest income from securities of $63m or a 4.0% yield, which seems more in line with current market rates.

Loan book ($1.5bn): Due to the large securities portfolio, the loan book is a much smaller percentage of assets and equity than at most banks. Over 85% of the loan book is in real estate. Around 30% of the loan book is non owner occupied commercial real estate. Management told me that they have very little in office. The CRE book is well diversified with sectors such as industrial and storage. Geographically the loans are located in the regions of the bank’s branches (Sonoma and Marin county) and in Sacramento. There isn’t much exposure to San Francisco. 8% of the loan book is in construction, with a large portion of that being in single- and multi-family. I see credit risks as being small compared to most banks due to the small size of the loan book, the strong capital position and the conservative management. The bank did have some credit issues in 2008, but they weren’t severe enough to cause dilution to shareholders. In terms of duration, management has told me that 80% of the book reprices every 5 years.

Deposit run stress test

While I don’t believe that EXSR has any of the ingredients that would cause a bank run, I still want to illustrate the strong capital and liquidity position of the firm. If we assume that $1bn of deposits would leave tomorrow, the bank would still have a CET1 ratio of 14.2%, significantly above peer banks and above regulatory minimums. In this scenario, I assume the bank would sell $1bn of its fixed income securities that carry a risk weighting of 20%. Assuming all the deposits that left the bank were non-interest bearing, the bank would still generate positive net income. I don’t think there are many banks that could survive this stress test.

Conclusion

I view EXSR as an attractive investment due to its low valuation, its strong capital position and the attractive returns on equity. The strong ROE is driven by the higher-than-average fee income and the rock bottom deposit costs. Due to the sleepy customer base, deposit costs will reprice slower than at most banks, which could mean that 2024 earnings will decline again. Nonetheless, I think earnings and ROE will still be attractive at today’s valuation of 0.7x tangible book value. Given the accretion of AOCI losses, I see the bank doubling book value over the next 5 years in my base case even after paying the $8m annual dividend. The dividend should eventually increase as AOCI losses run off and reported earnings increase. Assuming that valuation reverts to 1.0x tangible book value, the stock should be close to a triple on a total return basis. 

Catalyst

- Runoff of AOCI losses / growth in book value