Price-to-Book vs ROE: How Investors Should Value Bank Stocks in 2026

Key Takeaway

Every few quarters, a familiar screen lights up across retail terminals and professional blotters alike: large U.S. banks trading near or below one times book value. As of early November 2026, that screen still looked seductive. An invest-like.com snapshot dated 3 November 2026 put Citigroup near 0.98 times book with a multi-year average return on equity around 6.2 percent, while JPMorgan Chase sat near 2.49 times book with average ROE near 15.2 percent. A naive sort that crowns the lowest price-to-book name as the “cheapest” bank will keep selecting the low-ROE franchise and skipping the high-ROE compounder. That inversion is the entire point of pairing price-to-book with return on equity.

Bank valuation is where popular non-financial tools quietly fail. Enterprise value over EBITDA and free-cash-flow discounted-cash-flow models treat debt as a financing choice; for banks, interest-bearing liabilities are raw material. Equity residual logic—price over book or tangible book, judged against ROE or return on tangible common equity—is the cleaner frame. Sweetvaluelab’s Bank Health Board, with prices as of 2 November 2026 and trailing twelve-month figures through 30 September 2026, shows JPMorgan at about 3.19 times tangible book with ROTCE near 23.0 percent, while Citigroup sat near 1.52 times tangible book with ROTCE near 13.0 percent. The justified multiple from a simple Gordon residual identity makes the gap intuitive: when ROE equals the cost of equity, fair price-to-book collapses toward 1.0 times; when ROE sits well above the cost of capital, a premium book multiple can still be the cheaper claim on residual earnings.

This guide walks through what price-to-book measures, why ROE (and ROTCE) is the other half of the bargain test, how the justified P/B formula works with a numeric example, when tangible book is the better denominator for acquisitive banks, how early-November 2026 peer prints for JPM, BAC, WFC, C, GS, and MS line up, and which credit and capital falsifiers turn a low book multiple into a value trap. Educational note only: nothing here is investment advice. Multiples and capital ratios move with prices and filings; verify live figures before you act.

Why Bank Screens Mislead on Cheapness

Debt is the product, not a side lever

In industrial and most technology equities, investors routinely start with EV/EBITDA or an unlevered free-cash-flow model because enterprise value strips out capital structure noise. Banks invert that logic. Customer deposits, wholesale funding, and other interest-bearing liabilities are the inventory through which the franchise earns net interest income and fees. Subtracting those liabilities to invent an “enterprise” picture often destroys more information than it clarifies. That is why equity-side multiples—price-to-book, price-to-tangible-book, and earnings multiples on net income—remain the workhorse language for deposit banks and broker-dealers, even when the broader market conversation is obsessed with EV ratios.

Low book multiples can price a franchise that earns below its cost of capital

A stock at 0.7 times book is not automatically a 30 percent discount to intrinsic value. If the bank’s sustainable ROE sits below its cost of equity, the market is often telling you that a dollar of book equity is worth less than a dollar in present value. Screens that sort only on price-to-book systematically overweight those franchises in credit stress, restructuring, or chronically weak fee platforms. The complementary column—multi-year average ROE or trailing ROTCE—forces the disagreement into the open before position size enters the conversation.

Peer bands beat absolute “cheap” labels

Bank price-to-book also embeds expectations about the rate cycle, loan growth, fee intensity, and regulatory capital. A 1.2 times book print that looks rich versus a recession trough can look cheap versus a mid-cycle peer that earns mid-teens ROE with a fortress CET1 ratio. Absolute thresholds without peer context and without a return on equity overlay are how “value” banks quietly become permanent capital sinks. The rest of this article builds that dual-column habit with early-November 2026 data rather than slogans.

Price-to-Book: What the Multiple Actually Captures

Market value over accounting equity

Price-to-book divides equity market capitalization by book value of equity, or share price by book value per share. Book equity is an accounting construct: retained earnings, paid-in capital, AOCI, and related adjustments under the applicable GAAP or IFRS rules. For banks, that book figure is tightly linked to regulatory capital conversations, but it is not identical to CET1, Tier 1, or tangible common equity. Investors who treat “book” as a cash pile waiting to be distributed are already one step into the trap. Book is the residual claim after assets are marked under accounting conventions and liabilities are recognized—useful as a scale measure, dangerous as a liquidation fantasy.

Why banks still live on the book multiple

Banks are balance-sheet businesses. Net interest income, credit costs, and fee income all scale with the equity base the firm is allowed and willing to run. That makes price-to-book a natural way to compare how much the market pays for each dollar of accounting equity across peers with different share counts and absolute market caps. It also makes the multiple hypersensitive to write-downs, AOCI swings on securities portfolios, and goodwill from acquisitions. When those items move, price-to-book can jump without any change in the quality of the deposit franchise—another reason the raw screen needs a return overlay and, often, a tangible-book refinement.

Limits that show up before you even open a peer table

Price-to-book says nothing about earnings power by itself. Two banks at 1.2 times book can embed completely different implied ROE paths. It also struggles when book equity is distorted by large intangibles, deferred tax assets of uncertain realizability, or aggressive mark assumptions. Those limits are not reasons to abandon the multiple for banks; they are reasons to insist on ROE, tangible book, and capital quality as co-equal columns.

ROE and the Justified P/B Formula

Return on equity as the earnings claim on book

Return on equity divides net income by average book equity over a period. Multi-year averages matter more than a single quarter because bank earnings are cyclical through credit costs and rate betas. The invest-like.com 3 November 2026 snapshot pairs each large bank’s price-to-book with a multi-year average ROE precisely so investors do not confuse a trough or peak print with a sustainable franchise return. JPMorgan’s roughly 15.2 percent average ROE against a 2.49 times book multiple tells a different residual story than Citigroup’s roughly 6.2 percent average ROE against a 0.98 times book multiple, even before you layer capital ratios or fee mix.

Justified P/B from the Gordon residual identity

Under a simple constant-growth equity residual model, justified price-to-book equals (ROE − g) ÷ (r − g), where ROE is expected return on equity, r is the cost of equity, and g is the perpetual growth rate of residual earnings (with ROE > g and r > g). Worked example with ROE of 12 percent, cost of equity of 10 percent, and growth of 3 percent: (0.12 − 0.03) ÷ (0.10 − 0.03) = 1.29 times. When ROE equals the cost of equity, the numerator and the capital charge cancel in a way that collapses justified P/B toward 1.0 times. Drop ROE to 8 percent with the same r and g and the justified multiple falls to roughly 0.71 times. That arithmetic is why a high-ROE bank at 1.4 times book can be “cheaper” than a low-ROE bank at 0.6 times book when both are judged against their justified multiples rather than against a flat screen threshold.

What the formula does not forgive

The identity is only as honest as its inputs. Optimistic ROE that assumes a permanent credit utopia, a cost of equity that ignores bank beta in stress, or a growth rate that exceeds long-run nominal GDP without a shrinking payout will invent fake cheapness. Use the formula as a discipline device—map today’s price-to-book to the ROE the market is implying, then ask whether credit, capital, and cycle evidence can support that ROE—not as a one-cell spreadsheet oracle.

P/TBV, ROTCE, and Acquisitive Banks

Tangible book strips goodwill and related intangibles

Price-to-tangible-book divides market value by tangible common equity—book equity minus goodwill and certain other intangibles. For banks that grow through acquisitions, reported book can be inflated by purchase accounting in ways that do not represent deployable capital. Tangible book is the cleaner scale for those names, which is why professional bank comps lean on P/TBV alongside or instead of headline P/B. Sweetvaluelab’s 2 November 2026 board is explicit on this point: JPMorgan near 3.19 times tangible book, Bank of America near 1.79 times, Wells Fargo near 1.96 times, and Citigroup near 1.52 times, with prices of roughly $340.58, $49.57, $85.89, and $123.10 respectively.

ROTCE as the matching return metric

Return on tangible common equity pairs naturally with tangible book. Trailing twelve-month ROTCE through 30 September 2026 on that same board printed about 23.0 percent for JPMorgan, 15.8 percent for Bank of America, 17.2 percent for Wells Fargo, and 13.0 percent for Citigroup. Reading P/TBV without ROTCE repeats the original screen error with a better denominator. A 1.5 times tangible-book name earning low-teens ROTCE is not automatically more attractive than a 3 times tangible-book name earning low-twenties ROTCE; the justified-multiple logic still applies, only now on tangible capital.

CET1 keeps the capital constraint visible

Tangible book and ROTCE still omit the regulatory binding constraint. Common Equity Tier 1 ratios on the same early-November board sat near 14.1 percent for JPMorgan, 11.2 percent for Bank of America, 10.3 percent for Wells Fargo, and 12.8 percent for Citigroup. A bank can look “cheap” on book while running thinner capital buffers, which limits buybacks, raises the odds of equity issuance in stress, and changes the distribution of residual earnings to common holders. Treat CET1 as a falsifier column next to P/TBV and ROTCE, not as a footnote.

Large US Banks Side-by-Side in Early November 2026

Headline P/B and multi-year average ROE

Using the invest-like.com snapshot dated 3 November 2026, the large-bank cluster sorted roughly as follows on price-to-book: Morgan Stanley near 2.54 times with average ROE near 12.0 percent and P/E near 17.2; JPMorgan near 2.49 times with average ROE near 15.2 percent and P/E near 15.9; Goldman Sachs near 2.05 times with average ROE near 10.6 percent and P/E near 15.2; Wells Fargo near 1.39 times with average ROE near 10.2 percent and P/E near 12.1; Bank of America near 1.18 times with average ROE near 9.6 percent and P/E near 11.2; and Citigroup near 0.98 times with average ROE near 6.2 percent and P/E near 14.2. Notice that Citigroup’s earnings multiple is not the cheapest even though its book multiple is the lowest in the set—another reminder that “cheap on book” and “cheap on earnings” disagree when ROE is weak.

Tangible-book lens on the same week

Sweetvaluelab’s prices as of 2 November 2026 and TTM metrics through 30 September 2026 reorder the conversation toward tangible capital. JPMorgan’s roughly 3.19 times P/TBV against about 23.0 percent ROTCE is a premium multiple on a premium return. Bank of America’s roughly 1.79 times P/TBV against about 15.8 percent ROTCE and Wells Fargo’s roughly 1.96 times against about 17.2 percent ROTCE sit in a middle band where the justified-multiple question is whether mid-teens tangible returns can persist through the credit cycle. Citigroup’s roughly 1.52 times P/TBV against about 13.0 percent ROTCE looks cheaper on the raw tangible multiple and less cheap once you ask what ROE the market would need to justify a re-rating toward peer premiums.

Corroboration without pretending sources agree to the decimal

Stockanalysis.com recently showed JPMorgan near about 2.57 times book, about 3.05 times tangible book, and ROE near roughly 18 percent—close enough in direction to the early-November tables above to corroborate that JPM trades at a clear book premium tied to superior returns, even when exact prints differ by source methodology and as-of dates. Prefer the November 2–3 tables as the primary peer set for this article, and treat third-party corroboration as a check on direction rather than a license to average incompatible definitions.

When you want that dual-column comparison to stay live rather than freeze inside a static blog table, Alphio’s Conversational Trading workflow is built for asking follow-ups on bank multiples, ROE paths, and peer bands in plain language—then refining the watchlist without rebuilding the spreadsheet from scratch each Sunday night.

Pair that research loop with Scheduled Task so the P/B-versus-ROE and P/TBV-versus-ROTCE refresh runs on a cadence you choose instead of depending on memory after every Fed week.

When Low P/B Is a Credit Trap

Nonperforming loans and coverage before the “bargain” label

A depressed book multiple often arrives with rising nonperforming loans, thinner coverage ratios, or both. Credit costs flow through earnings first and through capital second; by the time price-to-book looks “statistically cheap,” the equity account may already be absorbing the stress the market fears. Investors who buy the low P/B print without reading NPL formation, net charge-offs, and allowance coverage are underwriting a credit call they never stated. The falsifier is simple: if asset quality is still deteriorating, the ROE you plugged into the justified P/B formula is already obsolete.

CET1 and distribution capacity

Capital is the binding constraint on buybacks, dividends, and balance-sheet growth. A bank at 0.9 times book with a CET1 ratio pressed against its internal target or regulatory expectation may be cheap on paper and expensive in optionality. Compare the early-November CET1 cluster—JPMorgan near 14.1 percent versus Wells Fargo near 10.3 percent on the Sweetvaluelab board—not as a beauty contest, but as a map of who has room to return capital if earnings hold and who may be forced to conserve. Low P/B plus thin capital is a different animal from low P/B plus surplus capital and a credible turnaround ROE path.

Cycle timing and AOCI

Rate shocks can swing accumulated other comprehensive income and, with it, reported book value, without an immediate change in the deposit franchise’s long-run earning power. Cycle timing also matters for when loan losses are recognized versus when spreads peak. Buying “cheap book” at the wrong point in that sequence is how value investors become involuntary long-term holders of restructuring stories. Demand a thesis for where you are in the credit and rate cycle before you treat any sub-1.0 times book print as mean-reverting arithmetic.

Building a Dual-Column Watchlist Process

Columns that force the disagreement into the open

A practical bank watchlist does not need forty fields. It needs price-to-book and price-to-tangible-book, multi-year average ROE and trailing ROTCE, CET1, a simple credit snapshot (NPLs, net charge-offs, coverage), and a one-line justified P/B estimate under explicit r and g assumptions. When those columns disagree—low P/B with low ROE, or high P/TBV with high ROTCE—the disagreement is the research agenda, not a bug to average away. Add a peer-band column so “cheap” always means cheap versus a defined group, not versus an absolute folklore threshold.

Rules before screens become position size

Encode a few non-negotiables before automation or discretionary sizing kicks in. Examples of process rules, not advice: refuse to label any bank a bargain on P/B alone; require ROTCE or multi-year ROE next to every tangible or headline book multiple; escalate review when CET1 is falling while the book multiple compresses; and re-run justified P/B when consensus ROE changes by a material amount after earnings. Those rules keep the Gordon identity honest and keep credit falsifiers from being optional.

From research loop to execution discipline

Once the dual-column process is stable, the operational risk shifts from “wrong multiple” to “inconsistent follow-through.” Alphio’s Copy Trade feature is useful when you want to mirror a disciplined bank or financials specialist whose published process already pairs book multiples with returns rather than chasing sub-1.0 times headlines. For investors who prefer rules engines over discretionary copy, Autonomous Trading can sit downstream of the same watchlist logic so entries and exits inherit the P/B-versus-ROE constraints you already wrote down.

Conclusion

Price-to-book remains a core bank multiple precisely because EV/EBITDA and unlevered free-cash-flow frames mis-specify the role of debt in a deposit franchise. The early-November 2026 evidence is unambiguous on the pairing problem: Citigroup near 0.98 times book with roughly 6.2 percent average ROE is not “cheaper” in residual terms than JPMorgan near 2.49 times book with roughly 15.2 percent average ROE, and the tangible-book board through 30 September 2026 tells the same story with P/TBV and ROTCE. Justified P/B = (ROE − g) ÷ (r − g) is the compact reason a high-ROE bank at a premium multiple can still be the better claim, while a low-ROE bank at a discount multiple can be fully priced—or worse. Credit quality, CET1, and cycle timing remain the falsifiers that turn neat spreadsheet discounts into traps.

If you want that dual-column discipline to stay live after the next earnings week, open a research thread in Alphio Conversational Trading, schedule peer refreshes with Scheduled Task, and—when you are ready to act inside explicit constraints—use Copy Trade or Autonomous Trading so the process you just read does not die in a forgotten spreadsheet. Educational content only; verify live prices, filings, and capital ratios before any decision.