Community Banks and the Benefit of Boring

Before March 2023, Silicon Valley Bank had a clear depositor niche (startups and the venture firms that fund them), a quickly growing balance sheet, and a great story to tell investors. Then deposits fled - quickly - and the bank failed within days. Signature Bank and First Republic followed. Now, the stories told about them are mostly the cautionary variety.

Nassim Taleb makes the point that between the Latin American debt crisis in 1982, and then the crash in 2008, the banking industry lost more money than it had made in its entire history to that point.¹

To an outside shareholder, bank balance sheets are opaque, and especially with the use of complex financial derivatives, it’s not always clear that the bankers themselves know what risks they’re taking.

On the other hand, banking is one of the oldest financial professions. The core banking business is simple and widely replicable: take in cheap deposits, lend money at a higher rate, and apply leverage to get an acceptable return on equity after expenses. It’s a time-tested model.

As a matter of caution I’ve generally stayed away from investing in bank shares, but the thought occurred: am I being unfairly prejudiced?

Some smaller banks, operating at the regional or community level, might be quite different (and simpler to understand) compared to the global mega-banks. I wanted to dig in closer to the roots of banking, and see if any exceptional operators or business models stood out.

Here’s what I found.

How I Screened

Banks don’t fit the analytical model I use elsewhere. Free cash flow, return on invested capital and debt-to-EBITDA are close to meaningless for a business whose raw material is borrowed money (for a bank, the deposits are the debt, and also the raw materials needed for making loans). So I built a bank-specific template within Claude and ran each company through it:

1. Financial condition. Ten years of credit quality, capital, funding and profitability, built from company filings and FDIC call reports. Asset quality and funding come first. Face-level profitability is less important, because a bank can look very profitable right up until the loans go bad.
2. Franchise and growth quality. What does this bank do unusually well, could a competent large bank copy it, and is the advantage growing or eroding? This phase also covered insider buying and governance.
3. Valuation and thesis. Price against the bank’s own history and its peers, a thesis with falsifiable indicators, and a verdict on whether to pass, add to the watchlist, or advance to deeper due diligence as a potential position.

The verdict rules were mechanical and set before I started. Advancing to a deeper look required either a strong franchise or an attractive valuation, and a weak financial condition was an automatic pass.

To find names to research, I started with a well-known Regional Bank ETF, and kept going through the list of holdings until the market caps looked too low to be investable-liquid. After running 55 names through my screening process, I pooled all the phase-one tables and asked questions across the whole group. The statistics below were computed from those tables.

A few terms defined:

• Net interest margin (NIM) is what a bank earns on its loans and securities minus what it pays for funding, as a share of earning assets.
• The efficiency ratio is operating expense divided by revenue. Lower is better: 50% means the bank spends 50 cents to earn a dollar.
• Nonperforming loans (NPLs) are loans that have stopped paying, and charge-offs are the ones written off.
• Tangible book value per share is shareholders’ equity minus goodwill, per share, roughly what a bank would be worth if wound up completely.

Where Are We in the Banking Cycle?

The short answer: profitability among small U.S. banks is at the top of its ten-year range, and that’s coming mostly from interest margins rather than superior loan quality.

Source: company filings and FDIC call reports for 55 US banks; medians computed by me. NPLs use nonperforming assets where a bank doesn’t report NPLs. Data through Q2 2026.

Median return on assets matches the decade’s previous highs (2018 and 2021), and both of those years had one-off help: a tax cut and reserve releases. This time it’s the margin. Median NIM is the highest in the series.

Credit, meanwhile, is normalizing from an unusually low base. Median NPLs are about double their 2021-22 trough, and 89% of banks have more problem loans than in 2022. Reserves haven’t kept up, so the cushion covering bad loans has thinned from 3.8 times NPLs to 2.3.

That being said, it doesn’t look like a turn yet. The share of banks with rising charge-offs went from 85% in 2023 to 41% in the latest twelve months, and margins are still widening at more banks than not.

Bear in mind that I’m writing this in early October 2026 based on data that stops in June. You’re reading in the future and so this is more of a quaint historical snapshot than actionable data. It will be interesting to check back in a year or two, repeat this study, and see how the banking profitability cycle has evolved.

Margins Move Together, Credit Doesn’t

I was curious to learn whether banks get into trouble in lockstep. How much does picking a superior operator actually help? To get a first impression, I compared each bank’s year-on-year changes from 2017 to 2025 against every other bank’s.

Margins move together (average correlation between banks of 0.44), which makes sense, since interest rates drive them. Charge-offs barely do (0.13). The gap in charge-offs between the better and worse halves of the group tripled from 2022 to 2025, and a bank’s credit rank in 2019 told you little about its rank today.

The point: the margin is a macro bet you take on with any bank. Credit is where the individual bank matters. Today’s worst credit among community banks is concentrated in a handful of specialty lenders rather than spread across the group. Where property stress is showing up, the common thread is multifamily loans (Austin apartments, affordable housing, New York rent-stabilized buildings) more than any one region.

The Lure of Cool

How valuable is a niche business model for building shareholder value within a small bank?

I sorted the banks into specialty lenders, where the niche is the business model (e.g. banking for fintech apps, lending to plaintiff law firms against their case books, financing trucking companies’ unpaid invoices) and everyone else (mostly construction, commercial, and residential real estate). Then I measured value created per share as what I’ll call tangible return: annual growth in tangible book value per share plus dividends.

At first glance, specialty banks won. Their median tangible return was 13.2% a year against 11.4% for the rest, and the gap is unlikely to be chance.

But here’s what that comparison hides:

1. Wider tails. 6 of the 16 specialty banks had at least one year of charge-offs above 1% of loans. Only 4 of the other 39 did.
2. Survivorship. My data only includes banks that still exist. SVB, Signature, First Republic and Silvergate were all niche models, and none of them is in the sample as they don’t exist anymore.
3. No recession. The ten years I studied contain no real credit downturn (2020 was cushioned by stimulus). Relatively young banks with a newly-developed niche may not have been tested by a full economic cycle.

The academic base rates point the same way. Fahlenbrach, Prilmeier and Stulz found that US banks in the top quarter for loan growth underperformed the bottom quarter by about 7.4% a year in stock returns over the following three years, as losses that hadn’t been provisioned for showed up later.² Meiselman, Nagel and Purnanandam found that banks reporting unusually high profits before a crisis tended to do worse in it.³

There’s an economist’s explanation for why. A novel lending edge works while it’s small. Scale it up and you’re increasingly winning loans because you were willing to say yes when others said no (“the winner’s curse”). Niches may make excess returns, but the usual route to them is fast growth plus concentration in a narrow sector, and that combination invites trouble.

The Boring Variable That Mattered

The single best predictor of tangible return in the group was the efficiency ratio.

Source: company filings and FDIC call reports for 55 US banks, 2016 to Q2 2026; calculations mine.

The rank correlation is strong (-0.63), and it holds when you throw out the specialty banks entirely. Efficient banks didn’t get there by taking more credit risk either: efficiency had no relationship at all with peak charge-offs. Every quartile contains four specialty banks, so this isn’t niche-versus-plain in disguise.

My first mental model had focused on the direction of the efficiency ratio, on the theory that a bank already running at, say, 45% (highly efficient) has little room to get better. However, a good starting level predicted later value creation. Every 10 points of lower efficiency ratio was worth about 1.5 percentage points a year of extra tangible return. The only clearly bad group were banks that started inefficient and then stayed that way.

Limitations: efficiency explains only about a fifth of the differences between banks, so it’s a filter, not a ranking tool. It’s also a ratio, so revenue moves it. A bank with very cheap deposits and a fat margin can post a low efficiency ratio without a lean cost base, and the lean cost base is the more durable of the two. And the market easily notices efficiency; good operators tend to trade at full prices.

Are Deposits the Moat?

The last thread is where I’ve ended up on what actually makes a bank durable.

Deposits gathered through brokers or rate-driven internet channels belong to the most mobile customers a bank has. They move the moment someone else pays more, and if AI agents get good at shopping yields on people’s behalf, that kind of funding will be permanently in jeopardy. A clever lending edge gets competed away once it’s big enough to notice. A great deposit franchise is very hard to copy, though I suspect it still erodes, just slowly.

So I re-ranked the group on deposits alone: total cost of deposits, the share that pays no interest, reliance on brokered or wholesale money, and my judgment of where the deposits come from.

Some examples of banks with very low deposit costs, and where those deposits likely originate from:

Source: company filings and FDIC call reports, Q2 2026 or latest available; ranking mine.

The strongest deposit franchises are built on business operating accounts and dominant local share. That money may sit in an account to make payroll and pay suppliers, not specifically to earn a yield, so better rate-shopping tools have the least impact.

We can also invert this screen. What about banks that pay full price for deposits, but still show high profitability? A bank that can operate well while also giving depositors a competitive rate on their money has less to lose if customers become yield-hunters. That line of inquiry leads back toward underwriting specialists and other niches that aren’t always obvious at a first pass. But this piece is long enough already.

Where Things Stand

The result of this research is that I have a shortlist of roughly a dozen community and regional banks that are standouts in some fashion: between deposit franchises, superior underwriting, or unusual economic niche, I’ve identified some names to track for future follow-up.

For the community banking sector as a whole, we’re clearly not at a trough in the credit cycle but it’s not obvious we’re at a peak, either. The figure that would demonstrate strain in local economies - non-performing loans - is a lagging indicator. But, with some exceptions, community bank balance sheets mostly look healthy.

Interest margins are near a 10-year high. Over the last few years, banks have been rewriting and repricing loans at higher rates, which lifts what they earn. Meanwhile, deposit costs (which rose more slowly than market rates on the way up) have been edging back down at several of the banks I looked at. I’ll be interested to see whether that gap holds, or whether margins narrow as depositors shop around for higher yield.

Of the banks I looked at, it was rare to observe both an excellent depositor franchise and excellent loan underwriting in the same place, and where I did, the market certainly saw the same thing and so shares were at a premium; no obvious bargains to be had.

Important Information

Banks included in the cohort statistics (55 US banks): AMAL, BANR, BOKF, CASH, CASS, CBC, CBSH, CCB, CFR, CHCO, CLBK, CUBI, CVBF, EFSC, ESQ, FBIZ, FFIN, FHB, FHN, FSUN, FULT, GABC, GBCI, HOMB, HOPE, HTH, HWC, IBOC, INBK, LOB, MBIN, NBN, NRIM, NWBI, OBK, OFG, OZK, PNFP, QCRH, RBCAA, RF, RNST, SFBS, SRCE, SSB, STBA, SYBT, TBBK, TFIN, TRMK, UBSI, WAL, WSFS, WTFC, ZION.

Limitations: the decade I studied has no credit recession, so I can’t tell how banks diverge under real stress. The sample contains only survivors, and only the banks I chose to look at on a fairly unsystematic sampling method, which probably tilts the overall sample toward quality. Tangible return measures value built inside the bank, not shareholder returns; it ignores changes in what the market pays for a dollar of book.

Note: Bank stocks carry risks beyond those of most equities. Banks are highly leveraged by design, so modest loan losses can consume a large share of equity, and deposits can leave far faster than assets can be sold, as 2023 showed. Credit quality, interest rates and regulation can change quickly, and the figures here run only through Q2 2026. The banks named in this article are illustrations of a research framework, not recommendations to buy or sell any security. A concentrated position in a single bank or region adds risk that diversification would otherwise reduce.

Positions and scope: As of last review date, the author, Nick Bormann, holds none of the named banks in his personal accounts but may buy or sell any security discussed at any time without notice. Bormann Wealth Management, LLC is a registered investment adviser and holds no positions in the securities discussed here. This article is general research, not advice for your circumstances. See Disclosures and Form ADV Part 2A.