WATCHconviction 2published 2026-09-11

WSBF — Waterstone Financial, Inc. · 2026-09-11 · Verdict: WATCH · Conviction 2

Price $21.36 (screen row, v2 run dated 2026-09-11; no live quote) · Mkt cap $384M · P/E 12.8x TTM · P/TBV 1.09x · ROTE 8.6% TTM · Div yield 3.2% · ADV $1.6M Financial lane: EV/EBIT is not used for a bank, per DEEPDIVE_PROMPT.md. Sources: 10-K 2026-02-26 (Items 1, 1A, 7), 10-Q 2026-08-06, DEF 14A 2026-04-09, 8-Ks 2026-03-17, 04-27, 05-19, 06-23, 07-23, Form 4s (12m). No transcript in the bundle.

Desk stats - Revenue trend. FY2025 net interest income $56.7M, up 22.9%, noninterest income down 4.6% to $85.2M, total revenue up 4.7% to $141.9M (10-K 2025, Item 7). Q2 2026 net interest income $16.0M, up 16.8%, noninterest income $24.2M, down 0.4% (Q2 8-K). Price, not volume: average loans grew 0.6% while margin went 2.60% to 3.03%. - Normalised after-tax profit. TTM pre-tax $38.1M (quarterly sums, Q2 8-K). One adjustment: remove the $327K net reserve release over four quarters (releases of $269K and $558K in Q3/Q4 2025, provisions of $264K and $236K in Q1/Q2 2026). No impairments, discontinued operations or litigation items ("There were no Significant Items for the years ended December 31, 2025 and 2024", 10-K 2025, Item 7). Normalised pre-tax $37.8M: $28.3M ($1.64/share) at the desk's 25% convention, $30.1M at the actual ~20.4% rate (Wisconsin exempts commercial and agricultural loan income, 10-K 2025, Item 7). - Price / normalised after-tax profit. $384M / $28.3M = 13.6x; 12.8x on GAAP TTM net income of $30.1M. - Bank metrics. P/E 12.8x TTM, P/TBV 1.09x (book $19.53 at 30 Jun 2026, no goodwill), ROTE 8.6% TTM and 9.65% annualised in Q2, tangible BVPS +8.6% in FY2025 ($17.53 to $19.03), equity/assets 15.6%. - Sustainable? Partly. Organic and cash-backed, but a funding-cost story with a finite tail: $864.9M of certificates mature within a year at a 3.57% retail cost, down from 4.11% (10-Q 2026-08-06). Management guides nothing. - What the screen got wrong. It paired FY2025 net income ($26.4M) with the June 2026 balance sheet, so its 14.5x P/E, 7.5% ROTE and 2.80% yield are three quarters stale (TTM: 12.8x, 8.6%, 3.2%) and its 41.3% growth is FY25 versus FY24. The lane has no revenue column, so it cannot see that 59% of TTM revenue is mortgage gain-on-sale income from a separate, barely profitable business.

1. What the business does

WaterStone Bank is a 1921-vintage Wisconsin thrift, 14 Milwaukee-area branches, 1.85% deposit share, ninth of 44 institutions in its metro (10-K 2025, Item 1). Its $1.68B loan book is 46% multi-family ($781.9M), 27% one-to-four family and 19% commercial real estate (Q2 8-K). Bolted on is Waterstone Mortgage Corporation, a national originator with 55 offices in 24 states that originated $2.05B of loans held for sale in 2025 for sale into the secondary market.

2. Why it might be mispriced, and the honest answer

The candidate edge is a conglomerate discount: a 48%-efficiency-ratio bank stapled to a mortgage originator running at 93% to 100% (Q2 8-K segment tables), so the consolidated 77.4% ratio makes the whole look mediocre. On TTM segment figures the bank earned $27.1M and the mortgage arm $2.8M.

That edge is largely spent. The stock is up 22.6% in six months and sits 0.7% below its 52-week high (universe_v2.csv). At 1.09x tangible book this is not priced as a broken thrift, and there is no forced seller, spin-off or index event.

3. Credit quality, margin and deposit mix

Credit is clean: nonperforming assets 0.30% of assets at 30 Jun 2026, nonaccrual loans 0.39% of loans, allowance 1.06% of loans and 283% of non-performing loans, net recoveries of $122K in FY2025 (Q2 8-K; 10-K 2025, Item 7). Concentration is the offset: multi-family, commercial real estate and construction total $1,174M against roughly $340M of bank Tier 1 capital, about 3.4x by my calculation from the disclosed balances and the 15.43% Tier 1 to average assets ratio, above the 300% supervisory guidance threshold, and the company intends to keep growing these books (10-K 2025, Item 1A).

Time deposits are $895M of $1,421M, and the 10-K carries the risk factor "We rely heavily on certificates of deposit, which has increased our cost of funds" (Item 1A). That is why margin fell to 2.17% in 2024 and why it has since added 43 basis points: retail CD cost fell 4.11% to 3.57% while loan yields rose 5.73% to 5.93% (Q2 8-K).

4. Balance sheet and capital allocation

Borrowings of $412M are almost all FHLB advances, $190M long-term maturing 2027 to 2030 (10-K 2025, Item 7). Capital return is the clearest positive: the dividend rose 13.3% to $0.17 in March (8-K 2026-03-17), April's authorisation of 2,000,000 more shares took the remaining total to 2,148,285, about 11.9% of the company (8-K 2026-04-27), and share count fell from 18,360K to 17,975K in six months. Insiders point the other way: 213,115 shares sold for $3.68M over twelve months, zero open-market purchases, with director Michael Hansen selling his entire 81,413-share position at $15.16 in October 2025 and director Patrick Lawton exiting all 49,255 shares near $18.60 in February 2026 (Form 4s).

5. Management: said versus did

CEO William Bruss called a "robust dividend payout ratio" a priority (8-K 2026-03-17), and the dividend rose and stock was repurchased in every quarter reviewed. On Q2 he was candid about the other half: mortgage "achieved solid growth in loan origination volumes ... but this growth was offset by softer sales margins" (Q2 8-K). Accurate: originations rose 5.6% to $621.8M while gross margin fell 3.84% to 3.60% and segment pre-tax income fell 20.2% (Q2 8-K segment table).

6. Valuation

Reverse DCF: at $21.36 against $19.53 of tangible book, with a 10% cost of equity and 2% growth, the current price implies a sustainable return on tangible equity of about 10.7%, above the 8.6% delivered over twelve months and the 9.65% of its best quarter.

Probability weighted, about $23.40, roughly 10% above the price. Fair value, not a bargain.

7. Catalysts and timeline

Q3 2026 results in late October: the margin print and whether retail CD cost breaks below 3.50%. Execution of the 11.9% authorisation. Federal Reserve cuts through 2027. A sale or separation of Waterstone Mortgage, nowhere signalled in the filings but the cleanest way to close the discount.

8. Pre-registered kill criteria

  1. Net interest margin below 2.90% for two consecutive quarters.
  2. Mortgage banking segment pre-tax income negative in two of any rolling four quarters.
  3. Nonaccrual loans above 0.75% of total loans, or any single-quarter multi-family charge-off above $3M.
  4. Fewer than 150,000 shares repurchased in a quarter while the stock trades below 1.2x tangible book.

9. Verdict and summary

WATCH, conviction 2. Waterstone is a genuinely improved bank: net interest margin went from 2.17% in 2024 to 3.03% in the June 2026 quarter as an $865M certificate book reprices down faster than its multi-family loans roll, credit is near pristine at 0.30% nonperforming assets with net recoveries, and capital return is aggressive, with the dividend up 13.3% and an authorisation covering 11.9% of the shares. The screen understates this by pairing FY2025 earnings with a June 2026 balance sheet, so the real multiple is 12.8x trailing, not 14.5x. But the market has noticed: the stock is 0.7% off its 52-week high at 1.09x tangible book, which implies a sustainable 10.7% return on tangible equity against the 8.6% delivered, and 59% of revenue still comes from a mortgage originator running at a 93% efficiency ratio that earned $22K pre-tax in March 2026. Add two directors who sold out entirely and no insider purchase in twelve months, and roughly 10% probability-weighted upside does not pay for a 46% multi-family concentration. Worth owning at 0.9x book, not 1.09x.

Research for discussion, not investment advice. Positions and sizing are the reader's decision.

Source markdown: 2026-09-11_WSBF.md · how these notes are built · every verdict tracked since publication.

Research and education only. Nothing here is investment advice or a recommendation to buy or sell any security. No price targets are recommendations; positions and sizing are the reader's decision. Past performance does not predict future results.