How to Value a Bank: FIG Valuation Explained

Introduction

"How do you value a bank?" is one of the fastest ways an interviewer separates candidates who memorized a DCF from candidates who actually understand what they are valuing. Reach for enterprise value, EBITDA, or unlevered free cash flow and you have already failed, because none of those concepts mean anything for a financial institution. The entire standard toolkit quietly assumes that debt is how a company funds itself. For a bank, debt is the product.

The timing makes this a live topic, not a dusty textbook exercise. Large US banks are posting some of the strongest results in their history: JPMorgan reported a 23% return on tangible common equity in the second quarter of 2026 excluding one-time items, and bank valuations have risen with the results, with the median US bank trading around 1.46 times adjusted tangible book value this spring, per S&P Global Market Intelligence's analysis of US bank stocks. Yet banks trade at wildly different valuations. JPMorgan changes hands at close to three times tangible book value while plenty of slower-growth regionals still sit near or even below book, and the gap is explained almost entirely by one number: return on equity. This guide walks through why the standard playbook breaks, which multiples actually matter, how the P/B versus ROE relationship works, and how to talk through it all under interview pressure.

Every valuation you learn first is built around enterprise value: you value the whole business independent of how it is financed, then subtract net debt to get to equity. That logic silently assumes debt sits on the right-hand side of the balance sheet as a funding choice. Banks blow up that assumption, so the enterprise-value bridge you rely on for an industrial or a software company simply does not apply.

Debt Is Raw Material, Not Financing

An industrial company borrows to build a factory. A bank borrows so it can lend that money back out at a higher rate. Deposits, wholesale funding, and issued debt are the raw material a bank buys, marks up, and resells. You cannot strip out financing to isolate an "unlevered" business, because the leverage *is* the business. A bank running with 10 times more liabilities than equity is not over-levered in the industrial sense; that ratio is the normal operating shape of the model.

Because of this, the clean separation between operating decisions and financing decisions collapses. There is no version of the bank that exists without its funding, so there is no enterprise value to compute in the first place. If you understand nothing else about FIG valuation, understand this: the enterprise-value framework is not just harder to apply to banks, it is conceptually meaningless. For a refresher on the standard bridge that does not survive here, our guide to enterprise value versus equity value lays out the machinery you are setting aside.

Interest Is Operating, Not Below the Line

For most companies, interest expense sits below operating income; it is a financing cost you add back to reach EBIT and EBITDA. For a bank, interest income and interest expense are the top and second lines of the income statement. Net interest income, the spread between what a bank earns on assets and pays on funding, is the core of revenue. Moving interest "below the line" would delete most of the bank's actual profit engine, which is why analysts never do it.

No Clean Unlevered Free Cash Flow

Unlevered free cash flow depends on being able to model the business before financing effects. Since a bank's cash flows *are* financing flows, there is no clean unlevered number to discount. Worse, the standard free cash flow build (EBIT, taxes, add back depreciation, subtract capex and working capital) has almost no analog: banks have negligible capex, working capital is a meaningless concept when your inventory is loans, and reinvestment shows up as regulatory capital that must be held against new assets rather than as plant and equipment.

Everything Runs on Equity Value

Put the three points together and the conclusion is forced: bank valuation is done entirely on an equity basis. You value the equity directly, using equity multiples and an equity-based discount model, and you never take a detour through enterprise value. The comparison below summarizes how each pillar of standard valuation gets replaced when you move to a bank or another FIG business.

Once you accept that everything is equity-based, the multiples fall out naturally. All of them start from equity value (market capitalization or share price) rather than enterprise value. Three do the heavy lifting: price to earnings, price to book, and price to tangible book. Our broader primer on common valuation multiples covers the enterprise-value cousins; here the focus is strictly on the equity multiples that banks actually trade on.

Price to Earnings for Banks

P/E works for banks the same way it works elsewhere: share price divided by earnings per share, or equity value divided by net income. Because net income already sits after interest and after taxes, it is a genuinely equity-level number, which makes P/E legitimate for a bank in a way EV/EBITDA never is. Bank P/E multiples tend to sit lower than the market average, often in the high single digits to low teens, reflecting cyclicality, sensitivity to credit losses, and heavy regulation.

The weakness of P/E for banks is that reported earnings can swing hard with the credit cycle. Provisions for loan losses can turn a strong year into a weak one and vice versa, so a single year's P/E can mislead. That volatility is exactly why balance-sheet multiples carry more weight in FIG than in most sectors.

Price to Book and Price to Tangible Book

Because a bank is essentially a portfolio of financial assets and liabilities carried at or near fair value, its book value of equity is far more economically meaningful than book value is for an industrial company. That makes P/B, market value of equity divided by book value of equity, the signature bank multiple. A bank trading at 1.0x book is priced at exactly the accounting value of its net assets; above 1.0x the market expects it to earn more than its cost of equity, and below 1.0x the market expects it to destroy value.

Analysts usually refine P/B into P/TBV, price to tangible book value, which strips goodwill and other intangibles out of book equity. The logic is that goodwill from past acquisitions cannot absorb loan losses or be lent out, so it should not count toward the capital base that actually backs the business.

Tangible Book Value

Tangible book value is total shareholders' equity minus goodwill and other intangible assets (and minus preferred equity when calculating tangible book value available to common shareholders). It represents the hard, loss-absorbing net worth of a bank. Because intangibles cannot cover credit losses or be redeployed into lending, analysts value banks on price to tangible book value rather than raw book value.

Why Tangible Book Value Is the Anchor

Tangible book value matters because it is the number regulators, investors, and acquirers all watch. It approximates the capital genuinely available to absorb losses, and it is the base against which return on tangible common equity is measured. In the second quarter of 2026, Bank of America reported tangible book value per share of roughly $29.37, up about 7% year over year, and management teams routinely frame growth in tangible book value per share plus dividends as the cleanest measure of value creation for shareholders.

If P/TBV is the multiple, return on equity is the reason a bank earns whatever multiple it trades at. Two banks with identical tangible book value can trade at very different prices, and the difference is almost always the return each one generates on that book. Understanding the mechanics of return metrics, which our ROIC, ROE, and DuPont analysis guide breaks down in detail, is the single most useful preparation for a FIG interview.

Return on Equity and Return on Tangible Common Equity

ROE is net income divided by average shareholders' equity. Banks increasingly headline ROTCE, return on tangible common equity, which divides net income available to common shareholders by average tangible common equity. ROTCE is structurally higher than ROE because the denominator excludes goodwill and intangibles, and it is the truest measure of how hard the loss-absorbing capital is working. JPMorgan's 23% ROTCE in the second quarter of 2026 excluding significant items, disclosed in its quarterly earnings filing, is exceptional; a mid-teens ROTCE is considered strong for a large bank, and single-digit returns signal a bank struggling to clear its cost of capital.

The Justified P/B Formula

The link between the multiple and the return is not hand-waving; it is a formula derived directly from the dividend discount model. The justified price-to-book multiple is:

Here $ROE$ is return on equity, $g$ is the sustainable long-term growth rate of book value, and $K_e$ is the cost of equity. The formula makes the earlier threshold precise. When $ROE = K_e$, the numerator and denominator are equal and P/B collapses to exactly 1.0x. When ROE exceeds the cost of equity, P/B rises above 1.0x; when ROE falls short, P/B drops below 1.0x. The whole valuation of a bank compresses into the spread between what it earns on equity and what its shareholders demand.

Reading the P/B Versus ROE Relationship

This is why JPMorgan can trade near three times tangible book while a sleepy regional trades below it: a 23% return against a cost of equity around 10% justifies a large premium, while a bank earning 8% against the same cost of equity justifies a discount. It also explains why bank stocks re-rate so sharply on rate moves and credit outlooks, both of which feed directly into ROE. To sanity-check the cost-of-equity input in that formula, our walk-through of how to calculate WACC covers the CAPM build for the cost of equity, which is the only discount rate that matters for a bank.

Bank valuation trips up more candidates than any other technical topic: Work through P/TBV, ROE, and dividend discount model questions with full worked answers on the practice platform, start practicing interview questions for free and find your weak spots before an interviewer does.

You cannot value a bank without understanding the leash it operates on. Regulators require banks to hold a minimum cushion of high-quality capital against their assets, and that requirement, not management ambition, ultimately governs how much a bank can grow, lend, and pay out. Capital is the binding constraint that debt capacity is for a leveraged buyout, and it feeds straight into the dividend discount model that follows.

CET1, Risk-Weighted Assets, and Buffers

The headline regulatory ratio is the CET1 ratio: common equity tier 1 capital divided by risk-weighted assets. The denominator matters as much as the numerator. Rather than weighting every asset equally, regulators assign risk weights, so a Treasury bond carries a near-zero weight while an unsecured corporate loan carries a full one. A bank can therefore grow its balance sheet without raising its risk-weighted assets one-for-one, depending on what it holds.

CET1 Ratio

The CET1 (common equity tier 1) ratio is a bank's highest-quality capital, mainly common stock and retained earnings, divided by its risk-weighted assets. It is the primary measure regulators use to judge whether a bank holds enough loss-absorbing capital. For large US banks the required level combines a 4.5% minimum, a stress capital buffer of at least 2.5%, and a surcharge for globally systemic banks, per the Federal Reserve's large-bank capital requirements.

The requirement is not a single number. The Federal Reserve sets it as a minimum of 4.5% plus a stress capital buffer of at least 2.5% plus, for the largest institutions, a globally systemic bank surcharge of 1.0% or more. That is why the biggest US banks run reported CET1 ratios well above the floor: JPMorgan reported a 14.1% standardized CET1 ratio in the second quarter of 2026, a comfortable margin over its requirement.

Why Excess Capital Matters in Valuation

Capital held above the requirement is money the bank can return to shareholders through dividends and buybacks or deploy into new lending. That excess is directly value-relevant. A bank sitting on capital well above its minimum has a store of value that either boosts distributions (raising the dividend discount model output) or funds growth (raising future earnings). Conversely, a bank near its minimum has no room to maneuver and may even need to raise dilutive equity, which the market punishes.

This is the bridge from regulation to valuation. The dividend discount model you are about to build is capped at the top by exactly this dynamic: a bank can only pay out what it does not need to hold as regulatory capital against the assets it wants to keep.

Since there is no unlevered free cash flow to discount, the DCF is replaced by the dividend discount model (DDM), which values equity directly as the present value of expected dividends. This is not a downgrade from a "real" DCF; it is the correct model for an entity whose cash flows are inseparable from its financing. The mechanics rhyme with the DCF you already know, so the intuition from our walk me through a DCF guide transfers, with two key swaps.

Why Dividends Proxy Distributable Earnings

The first swap is the cash flow. In a DCF you discount free cash flow; in a DDM you discount dividends, because dividends are the only cash a shareholder actually receives and, for a bank, they proxy distributable earnings, the profit left over after retaining enough capital to satisfy regulators and fund growth. A bank earning strong ROE but forced to retain most of it to support a growing balance sheet distributes little, and the DDM correctly captures that its value to a shareholder is lower than raw earnings suggest.

Discounting at the Cost of Equity

The second swap is the discount rate. Because you are valuing equity directly, you discount at the cost of equity $K_e$, never at WACC. A blended WACC would double-count the funding that is already embedded in the bank's net interest income. The simplest form is the Gordon growth version, which mirrors the Gordon growth terminal value you use elsewhere:

where $D_1$ is next year's dividend, $K_e$ is the cost of equity, and $g$ is the long-run growth rate. In practice analysts build a multi-stage model: explicit dividend forecasts driven by projected ROE and payout ratios for several years, then a terminal value using this Gordon growth formula.

Dividend Discount Model

The dividend discount model values a company's equity as the present value of all its expected future dividends, discounted at the cost of equity. For banks it replaces the standard discounted cash flow model because a bank has no meaningful unlevered free cash flow, and its distributable cash is constrained by regulatory capital requirements. Dividends serve as a proxy for the earnings a bank can actually pay out after holding required capital.

The Residual Income and Excess Returns Approach

A close relative of the DDM, and a favorite of the interviewer who wants to see depth, is the residual income or excess returns model. It values equity as current book value plus the present value of future excess returns, where an excess return is the profit a bank earns above its cost of equity on its capital:

The elegance is that it says the same thing as the justified P/B formula in a different form. If a bank earns exactly its cost of equity forever, every excess-return term is zero and value equals book, so P/B is 1.0x. Every point of ROE above the cost of equity adds value on top of book. It is a clean way to explain, in an interview, precisely why premium banks trade above tangible book and troubled ones trade below it.

Get the complete technical framework: Download the comprehensive 160-page PDF, access the IB Interview Guide covering bank valuation, the dividend discount model, and every core methodology interviewers test.

Numbers make the framework stick, so here is a compact example you could reproduce on paper in an interview. Suppose a bank generates a sustainable ROTCE of 15%, its cost of equity is 10%, and its tangible book value can grow at 4% per year over the long run. Plug those into the justified multiple:

So this bank should trade at roughly 1.83 times tangible book value. If its tangible book value per share is $50, the justified share price is about $92. Now change one input to see the sensitivity: hold everything else constant but drop ROTCE to 10%, exactly the cost of equity, and the multiple collapses to 1.0x, giving a $50 share price. Raise ROTCE to 20% instead and the multiple jumps to about 2.67x, or roughly $133 per share on the same book.

That sensitivity is the entire story of bank valuation in one table of outputs. The tangible book value barely moved; the value did, and it moved because ROE moved relative to the cost of equity. This is why FIG analysts obsess over net interest margin, credit costs, and efficiency: every one of them ultimately shows up in ROE.

Comparable company analysis for banks looks different from the EV/EBITDA grid you would build for an industrial. The columns are equity multiples and equity returns, and the operating metrics are bank-specific. Building a clean comps set is a core FIG analyst task, and it is worth knowing exactly which fields sit on the page.

Which Metrics Sit on the Comps Page

A bank comps page typically carries these columns rather than the enterprise-value metrics you see elsewhere:

Net Interest Margin

Net interest margin (NIM) is net interest income divided by average interest-earning assets. It measures the spread a bank earns between the yield on its loans and securities and the rate it pays on deposits and other funding. A wider NIM feeds directly into higher return on equity, which is why NIM sits on every bank comps page and moves sharply when interest rates change.

Reading a Bank Comps Table

The read on a bank comps table is a coherence check between multiple and return. A bank trading at a high P/TBV should also show a high ROTCE; if it does not, either the market expects returns to improve or the stock is expensive. A bank at a low P/TBV with a strong efficiency ratio and rising ROTCE can be a value opportunity or a value trap, and distinguishing the two is the analytical work. Pair the multiple with the return every time, because in FIG the two are mechanically linked in a way they are not for most sectors. This same balance-sheet-first lens runs through our overview of the financial institutions group and the deals it advises.

FIG is broader than commercial banks, and the valuation logic flexes across sub-sectors while keeping the same equity-based DNA. The common thread is that these are all balance-sheet businesses where debt-like liabilities are the raw material, so equity multiples and returns dominate.

Insurance: Book Value and Embedded Value

Insurers are valued much like banks, on P/B and P/E, because they too hold large investment portfolios funded by policyholder liabilities rather than by operating the standard EBITDA model. Property and casualty insurers lean on book value multiples and return on equity, with the combined ratio (losses plus expenses divided by premiums) as the key operating metric. Life insurers add a specialized measure called embedded value, which estimates the present value of the in-force book of policies plus adjusted net worth, capturing the long-tailed economics that a single year's earnings miss.

Asset Managers and Other Financials

Not every FIG business is a balance-sheet business, and the valuation follows the economics. Asset managers, exchanges, and advisory firms are capital-light and fee-driven, so they are valued more like normal companies, often on P/E and sometimes on EV/EBITDA because their earnings are not funded by a book of loans or policies. The lesson for an interview is to diagnose the model first: if the balance sheet is the engine, use equity multiples and returns; if fees are the engine, the standard toolkit comes back into play.

Interviewers know exactly where candidates stumble on FIG valuation, and a handful of traps recur so often that avoiding them signals real understanding. Each one flows from forgetting that a bank is an equity-and-capital story, not an enterprise-value story.

Using EV/EBITDA or a Standard DCF

The cardinal error is applying enterprise-value logic. Quoting EV/EBITDA, computing net debt, or building an unlevered DCF for a bank all reveal that a candidate has not internalized why the framework does not apply. If you catch yourself reaching for enterprise value, stop and reframe on equity: P/TBV, P/E, and a dividend discount model at the cost of equity.

Forgetting Provisions and the Credit Cycle

A subtler trap is ignoring credit costs. Provisions for loan losses can swing a bank from a strong year to a weak one, so valuing a bank on a single year's earnings without considering where it sits in the credit cycle produces misleading multiples. Strong candidates normalize for the cycle and treat a trough-year P/E with suspicion, the same way they would question a peak-margin EBITDA elsewhere.

Levered Versus Unlevered Confusion

The last trap is mixing levered and unlevered concepts. Because a bank's model is inherently levered, there is no unlevered version to isolate, and any attempt to compute unlevered free cash flow or an unlevered beta adjustment for the operating business is a category error. Everything about a bank is levered by design, so you discount equity cash flows at the cost of equity and never blend in a debt cost through WACC.

Valuing a bank is less about learning a new set of formulas and more about noticing that the assumptions behind your standard toolkit quietly break. Once you see that debt is the product rather than the financing, everything else follows: you value equity directly, you anchor on tangible book, you let ROE relative to the cost of equity set the multiple, and you swap the DCF for a dividend discount model constrained by regulatory capital. Get those connections straight and the dreaded "how do you value a bank" question becomes one of the easiest ways to demonstrate genuine understanding.

The 2026 backdrop makes the exercise concrete. With large banks posting record returns and trading across an enormous range of P/TBV multiples, the difference between a JPMorgan at three times book and a struggling regional below book is not mystery, it is math: return on equity measured against the cost of equity, filtered through the capital a regulator lets each bank deploy. Master that relationship and you will not only answer the interview question, you will actually understand why bank stocks move the way they do. For the wider context on what FIG bankers do with these valuations day to day, the financial institutions group overview is the natural next read.