← Research

The banking sector, priced twice

What the market says a bank costs, against hope-free tangible value

Published 2026-08-14 · 227 banks · 161 investable · median gap -36.8%

Every bank in the universe is priced twice: once by the market, and once by a model that strips out growth expectations and values only what is tangibly there. Across 227 US banks, the market's distribution sits about 37% above the model's — the growth premium rendered as a lateral shift. The interesting question is not the average gap but the exceptions: banks trading at or below hope-free value, and banks whose tangible value rose while their price fell. Both turn out to be rarer, and more structural, than the raw screen suggests.

-36.8%
Median hope-free gap to market price
24
Banks at or below tangible value, before filtering
5
Survive as investable and normally structured
5 / 210
Show value rising while price fell

The two distributions

The same banks, priced both ways, on one scale. The model's distribution is not just lower but tighter — a bounded dial cannot run away the way an unbounded market multiple can.

25th pct
$23.46
$16.81
-28%
Median
$38.03
$24.26
-36%
Mean
$60.89
$38.33
-37%
75th pct
$64.08
$37.39
-42%

Right column is the model's discount to market at that point of the distribution. The maximum pair is omitted from the plot — at $2,218.49 market against $1,812.61 model it would flatten every other row on a shared linear scale. Per-share dollar levels are partly just slice count, so shapes matter more than any single bank's position.

Market closing price Model value, growth stripped out

The franchise ranking

The premium a bank trades at over its own tangible value is the market's franchise ranking, stated as a number.

BAC 1.35× Stick moved ~9% in a year the price swung ~40%
WFC 1.4× Franchise premium in line with BAC
JPM 2.6× The market's top franchise ranking in the sector

Below the stick is mostly structural

The raw below-value screen looks like 24 opportunities. Most of it is two known artifacts: microcaps, and mutual-holding-company thrifts whose public minority trades at a permanent discount because no acquirer can ever buy the whole company.

  1. 01 Trade at or below the stick 24 banks
  2. 02 Microcap or MHC thrift — structural discount 19 banks
  3. 03 Investable and normally structured 5 banks

Closest to hope-free value

The investable, normally-structured pool of 161 banks, ranked by distance to tangible value. A positive "below by" means the market is paying less than the model says the bank is tangibly worth.

TickerCloseModelBelow byvs sectorMcapInsiderFlags
WSBF$20.84$23.1711.2%48%358M18%
FFIC$15.47$17.1210.7%47.5%524M7%UNRES-PREF
RBB$27.15$29.197.5%44.3%465M7%
KRNY$9.52$9.833.3%40.1%602M14%
PKBK$34.39$35.272.6%39.4%410M13%UNRES-PREF
EGBN$27.71$27.56-0.5%36.3%847M8%
PCB$27.90$27.06-3%33.8%397M26%
OCFC$19.42$18.42-5.2%31.7%1364M4%
ALLY$43.70$40.81-6.6%30.2%13592M0%
HOPE$13.99$12.43-11.1%25.7%1795M5%

Value rising, price falling

The sharpest disagreement the instrument can show. Of 210 banks measured, price fell more than 5% for six, and of those the tangible value rose more than 3% for five. Two honest readings: the market is walking away from improving fundamentals, or it is pricing a deterioration the trailing filings cannot see yet.

TickerPrice 1yModel 1yDivergenceCloseModelvs sector
FINW-20.1%+18.4%38.5%$14.27$13.9334.4%
PNBK-24.5%+8.2%32.7%$1.17$0.63-9.6%
CLBK-24.5%+6.3%30.9%$10.85$11.0638.8%
UNB-13.6%+9.4%23%$23.63$19.4619.2%
BANF-9.2%+11%20.2%$113.06$48.70-20.1%

What the sector looks like

  1. The market's price distribution sits about 37% above the model's — the growth premium rendered as a lateral shift. The stick's distribution is tighter, because a bounded dial cannot run away the way an unbounded market multiple can.
  2. A stick is a staircase, not a line. It steps at filing dates and crawls in between: BAC's moved roughly 9% in a year its price swung about 40%.
  3. Below-stick is mostly structural, not a mispricing. Of 24 banks at or below hope-free value, 19 are microcaps or mutual-holding-company thrifts. Only 5 survive in the investable, normally-structured pool.
  4. Price falling while the stick rises is rare — five cases in a rally year — and heterogeneous. One is compelling, one is an expensive bank simply converging down toward value. The instrument makes the disagreement measurable; it cannot arbitrate it.

Limitations

Where the instrument is weakest. Nothing here has demonstrated predictive skill, and the watchlist names are candidate journal entries rather than recommendations.

  • The stick takes the book's marks at roughly face value. When the market prices a bank near hope-free value it is often doubting exactly those marks — near-stick pricing on a real bank frequently means credit fear, not a bargain.
  • Vintage approximation. Tangible floor and pillar 2 are genuinely recomputed per quarter, but pillars 3 and 4 are held at today's values, because peer percentiles per vintage need backtest infrastructure and float has no history at all.
  • Per-share dollar levels are arbitrary. A bank cut into fewer slices shows a higher price per slice, so price-level rankings measure slice size, not company value. The gap and vs-sector columns are the fair comparisons.
  • Nothing here has demonstrated predictive skill. The watchlist names are candidate journal entries, not recommendations.
  • Merger vintages breathe. Depressed merger-quarter bases inflate the following year's momentum and then unwind; the floor stays honest through it, but the momentum pillar does not.

Universe of 227 banks pulled 2026-08-14; 223 priced on both measures. Audit queue (absolute upside above 75%): BYFC, TBBK.