How to Evaluate Bank Stocks Using Metrics That Matter

Most of the valuation habits people bring to stocks fall apart the moment they open a bank. A bank's revenue is scattered across interest income, fees, and trading gains. Its balance sheet is mostly loans and deposits, not factories or inventory. And its profit comes from the spread between what it pays for deposits and what it charges on loans, which is a mechanic that simply doesn't exist in a normal company. So the usual toolkit gives you the wrong read.

Here's the trap I see most often. Banks tend to trade somewhere around their tangible book value, which looks cheap next to a software company trading at many times book. That comparison is noise, because the two balance sheets are made of completely different stuff. If you want to judge a bank, you have to use bank metrics. Below are the ones I actually look at, in the order I look at them.

One practical note before the metrics. Almost everything here comes straight out of the filings on EDGAR. The 10-K and each 10-Q carry the balance sheet, the income statement, and the credit and capital tables, and the parts that actually move the thesis usually live in the management discussion and the footnotes rather than the summary tables up front. Banks also publish a quarterly earnings supplement or investor presentation that lays out NIM, the efficiency ratio, credit metrics, and capital ratios on one or two pages, which is the fastest way to build a peer comparison. Pull two or three comparable banks side by side and the outliers jump out.

Net Interest Margin: Where the Money Comes From

Net interest margin (NIM) is the gap between what a bank earns on its loans and investments and what it pays on its deposits and borrowings, measured against average earning assets. If a bank reports a NIM of, say, 3.5%, it's clearing about 3.5 cents of interest for every dollar of earning assets it puts to work.

NIM is the engine of the whole business, and a few things move it: the rate environment (higher rates usually help, as long as the bank can reprice loans faster than deposits), how competitive the fight for deposits is, what kind of loans sit in the book, and how the bank funds itself. When NIM is climbing, the core business is getting more profitable. When it's falling, something is squeezing the spread, whether that's deposit pricing pressure, a drift into lower-yielding loans, or a flatter yield curve that narrows the gap between funding costs and lending rates.

Don't read one quarter and call it a trend. Pull at least two years of NIM so you can see through seasonality and one-off items, then compare against banks of similar size and mix. A regional lender leaning on commercial loans usually runs a wider NIM than a big money-center bank with a large capital markets arm, so comparing the two tells you nothing.

Efficiency Ratio: How Disciplined the Operator Is

The efficiency ratio is non-interest expense as a share of total revenue (net interest income plus fee income). Lower is better here, because it means the bank spends less to produce each dollar of revenue. It's the closest thing banking has to a clean read on management discipline.

As a rough map, the best large banks tend to run below 60%, community and regional banks often sit somewhere in the 55% to 70% range, and anything well above 70% is a flag. A high ratio means the bank either has a cost problem or a revenue problem, and usually you can figure out which by looking at whether expenses or revenue moved.

What I really watch is the direction. A bank growing revenue faster than expenses will see this ratio drift down over time, which is positive operating leverage and one of the better signs that a management team knows what it's doing.

Loan Quality and Credit Risk

A bank makes money by lending, and lending means taking credit risk. All that interest income only reaches the bottom line if the loans get paid back, so loan quality decides whether the reported earnings are real or borrowed from a future write-off.

Start with non-performing loans (NPLs) as a percentage of total loans. These are loans where the borrower has stopped paying, usually 90 days past due or more. As a rule of thumb, an NPL ratio under 1% is healthy, north of 2% is worth digging into, and up around 5% points to real trouble in the book.

Then look at the allowance for loan losses, sometimes just called the reserve, which is the bank's own estimate of losses baked into the current loan book. Compare the reserve to non-performing loans (the coverage ratio). Coverage above 100% means the reserve more than covers current problem loans, which gives you a cushion. Below 100% suggests the bank may have to top up provisions later, and provisions come straight out of earnings.

One more cross-check I like: net charge-offs, which are the loans actually written off minus any recoveries, as a share of average loans. Set that next to the provision expense. If provisions run ahead of charge-offs, the bank is building reserves, which is conservative. If provisions lag charge-offs, the bank is effectively releasing reserves, which can flatter current earnings and is worth a second look.

Capital Adequacy

Capital is the shock absorber that lets a bank eat unexpected losses without tipping into insolvency, and regulators set floors so that cushion can't get too thin. The ratios to know are Common Equity Tier 1 (CET1), Tier 1 capital, and total capital.

CET1 is the one that matters most, because it's the highest-quality capital, mostly common equity. Under Basel III, large banks have to hold CET1 of at least 4.5%, plus a 2.5% capital conservation buffer on top, plus any surcharge for the biggest systemically important institutions. In practice, banks you'd actually want to own carry CET1 comfortably into the low teens.

A bank sitting near the regulatory minimum has almost no room to absorb losses, pay dividends, or grow the loan book. A bank carrying capital well above its requirement has options: it can return cash to shareholders through dividends and buybacks, which is usually what you want to see. So capital works as a safety metric and also tells you how much flexibility the bank has.

Return Metrics

Return on assets (ROA) tells you how well the bank turns its whole asset base into profit. For a bank, a ROA around 1% or better is generally strong, and drifting below about half a percent suggests it isn't earning its keep on the assets it holds.

The return metric I lean on more is return on tangible common equity (ROTCE), because it measures the return on the equity shareholders actually own, stripping out goodwill and intangibles picked up in acquisitions. The strongest banks tend to compound ROTCE in the mid-teens or higher, which is the level where a bank is clearly earning more than its cost of equity and genuinely creating value.

ROTCE is especially handy when you're comparing banks with different acquisition histories. A serial acquirer carries a pile of goodwill, which inflates total equity and drags reported ROE down. ROTCE strips that accounting artifact out and gives you a cleaner picture of what the operating business actually returns on shareholder capital.

Deposit Franchise Value

People skip past the deposit base because it rarely shows up in a single headline number, and that's a mistake. Banks with large, sticky, low-cost deposits have a structural edge, because their funding is cheaper than banks leaning on wholesale money or high-rate promotional deposits to keep the lights on.

To size up the franchise, look at three things: the share of non-interest-bearing deposits (checking accounts that pay the depositor nothing), the average cost of deposits, and how deposits are trending. A bank funding a big slice of its book with non-interest-bearing accounts is effectively borrowing at close to zero, and that flows straight into a wider NIM.

Stability matters as much as cost. During the 2023 banking stress, banks whose deposits were concentrated among large institutional clients or a single industry saw money leave fast, while banks with broad, diversified retail deposits held up. That episode is why the share of uninsured deposits, meaning balances above the FDIC insurance limit, became a risk metric people now check first. It's worth pulling from the filings before you get comfortable with any bank.

Valuation

Price to tangible book value (P/TBV) is the workhorse valuation metric for banks. It's what you're paying for each dollar of tangible shareholder equity. At 1.0x, the market is pricing the bank at its net asset value. Above that, the market is betting the bank will earn more than its cost of capital. Below that, the market is either expecting weak returns or suspecting there are losses hiding in the book.

The trick is to read P/TBV against ROTCE, because the two are linked. A bank compounding a mid-teens ROTCE against a high-single-digit cost of equity deserves to trade above book. A bank grinding out returns below its cost of equity should trade at a discount, and if it doesn't, you're paying up for nothing. That ROTCE-to-P/TBV relationship is about the most dependable anchor there is in bank analysis.

Put a hypothetical to it. Say two banks both trade at 1.4x tangible book, but one earns a mid-teens ROTCE and the other earns high single digits. They look identically priced on the screen, and they aren't remotely the same buy. The high-return bank is arguably fair or even cheap for what it compounds, while the low-return one is expensive for what it delivers. Reading valuation without the return next to it is how people end up owning cheap-looking banks that stay cheap for good reason.

Always sanity-check against forward price-to-earnings using consensus estimates. When a bank looks cheap on P/TBV but expensive on P/E, that gap is often a tell that book value is overstated, usually from unrealized losses on the securities portfolio that haven't fully hit reported equity yet. That's a real risk when rates rise, and it's exactly the kind of thing that lives in the footnotes rather than the headline numbers, so it's worth reading them before you get comfortable with the valuation.