How to Analyze a Bank

Why a bank is not a normal company

A bank is a business built on borrowed money, and that changes every rule of analysis. A typical company borrows to buy factories or inventory; a bank borrows as its raw material. It takes in deposits and other funding at a low rate, lends that money out at a higher one, and keeps the spread. Debt is not a warning sign on a bank's balance sheet, it is the inventory, which is why the usual alarm bells about leverage do not ring the same way.

That leverage is enormous by design. JPMorgan Chase, a useful anchor for the whole sector, carried assets of roughly twelve times its equity at the end of fiscal 2025, according to Tenet data drawn from its 10-K. For an industrial company that ratio would signal distress. For a bank it is ordinary, because the assets are loans and securities rather than machinery, and the model only works at scale. Understanding this is the first step, because it tells you to judge a bank on the return it earns on a thin slice of equity, not on how much it owes.

The consequence is that the tools you would reach for first with most companies, especially the price-to-earnings ratio, tend to mislead here. Bank earnings are volatile in a specific way, and the balance sheet, not the income statement, is where the truth usually sits. The rest of this guide walks the measures that work.

Why the P/E ratio misleads

The price-to-earnings ratio is unreliable for banks because bank earnings swing with the credit cycle, and they swing at the worst possible times for a naive reader. In good years, loan losses are low and profits look large, so the P/E looks cheap. Then a downturn arrives, losses mount, the bank sets aside reserves, and earnings collapse, so the P/E looks expensive just as the stock may be at its most attractive. The ratio sends the wrong signal at both turns. The general limits of the multiple are covered in the price-to-earnings ratio, and banks are where those limits bite hardest.

There is a second reason. A bank's assets are overwhelmingly financial, loans and securities carried at or near their market value, so its book value means something concrete in a way a factory's does not. That makes price-to-tangible-book value a more honest gauge of what you are paying, which we come back to below.

The metrics you need to analyze a bank

Five measures carry most of the weight when you analyze a bank: return on equity, return on assets, the efficiency ratio, net interest margin, and the cost of credit. Together they describe how well a bank turns its borrowed money into profit, and how much risk it runs to do so. None stands alone, because leverage links them, and a strong reading on one can hide a weak one on another.

Return on equity and return on assets are the headline quality gauges. Return on equity shows profit against the owners' capital, and a sustained figure near 15 percent has long marked a strong bank. Return on assets shows profit against the whole balance sheet, and near 1 percent is the traditional benchmark. In fiscal 2025 JPMorgan earned a return on equity of about 15.7 percent and a return on assets of about 1.3 percent, per Tenet data, both healthy. Because leverage links the two, read them together, using the guides to return on equity and return on assets.

The efficiency ratio is the bank's version of an operating-margin measure.

Efficiency ratio = noninterest expense / (net interest income + noninterest income)

It shows how many cents a bank spends to generate a dollar of revenue, so a lower number is better. The low 50s is excellent; above 65 percent is a sign of a bloated cost base. A large, well-run bank like JPMorgan typically operates with an overhead ratio in the low-to-mid 50s.

Net interest margin captures the core lending spread: interest earned on loans and securities, minus interest paid on funding, measured against interest-earning assets. In fiscal 2025 JPMorgan reported net interest income of about $95.4 billion against total interest income of roughly $193.3 billion, per its 10-K, meaning it paid out close to half of what it earned in interest to fund itself. A wider and steadier margin points to cheap, sticky deposits and disciplined lending.

The source of that funding is itself a quality signal worth checking. A bank funded largely by ordinary checking and saving deposits, which pay little and rarely leave, has a durable cost advantage over one that leans on wholesale borrowing that reprices fast and can flee in a panic. This deposit franchise does not appear as a single ratio, but it shows up in a low, stable cost of funds and a net interest margin that holds when rates move. It is often the most valuable and least visible asset a bank owns, and it is a large part of why a scaled deposit-taker like JPMorgan can sustain the returns above.

Credit costs: what actually sinks banks

Banks are rarely destroyed by weak revenue; they are destroyed by bad loans. The most important risk in any bank is credit quality, the chance that borrowers do not repay, and it is why a single strong year of earnings tells you far less than the quality of the loan book behind it. A bank can post record profits for years by lending aggressively, then lose it all in one cycle when those loans go bad.

The figures to watch are the provision for credit losses on the income statement, the allowance for loan losses on the balance sheet, and the share of loans that are non-performing. Rising provisions mean the bank expects more borrowers to default. A thin allowance relative to the loan book means a smaller cushion when trouble comes. It also helps to know what the bank lends against, because a book concentrated in one risky area, commercial property or a single industry, can sour all at once, while a diversified book spreads the danger across borrowers whose fortunes do not all move together. Because these costs are partly a management judgment about the future, they can be understated in good times, which is exactly when caution is cheapest and most tempting to skip.

This is the deeper reason the price-to-earnings ratio misleads. Reported earnings depend on a reserve estimate that management sets, so two banks with identical loan books can show different profits simply because one is more conservative. The disciplined approach is to assume losses are understated at the top of a cycle and to prize banks that reserve steadily through it. The broader idea of judging a business by what can go wrong is developed in understanding business risks.

Reading price-to-tangible-book value

Price-to-tangible-book value is usually the most useful valuation gauge for a bank, because it compares the market price with the hard equity actually backing the business, stripped of goodwill and other intangibles. Since a bank's assets are mostly financial and marked near fair value, tangible book value is a reasonably concrete floor of worth, which makes the ratio more honest than price-to-earnings across a cycle.

Price to tangible book = share price / tangible book value per share

A quality bank compounding a 15 percent return on equity will usually trade above one times tangible book, because it grows its own equity faster than a weaker rival, adding to the value backing each share year after year. In July 2026 JPMorgan's shares near $338 sat at roughly three times its fiscal 2025 tangible book value per share of about $107, per Tenet data, a premium the market grants for consistently high returns. A struggling bank, by contrast, often trades below tangible book, a warning that investors doubt the stated value of its loans. The general logic of the measure is set out in the price-to-book ratio.

The table below summarizes what good and weak look like across the core measures. Figures are illustrative benchmarks, not any single bank.

Where to go from here

Analyzing a bank means reading the balance sheet before the income statement: how much it earns on its equity, what it spends to run, the spread it captures on lending, and above all the quality of its loans. Because leverage is structural, lean on return on equity and price-to-tangible-book rather than raw debt or a single year's P/E. For the ratios themselves, read the price-to-book ratio and return on equity above, then see the same balance-sheet lens applied to related businesses in how to analyze an insurance company and how to analyze a REIT. To dig into a real name, open the JPMorgan financials on Tenet.