CASH — Pathward Financial, Inc. · 2026-09-10 · Verdict: WATCH · Conviction 3
Price $81.06 (screen close, 2026-09-09) · Mkt cap $1.70B (21,023,902 shares at 30 June 2026) · EV and EV/EBIT n/m for a bank · P/E 10.3x on FY2025 GAAP EPS · P/TBV 3.14x · ADV $16.5M Sources read: 10-K 2025-11-25 (Items 1, 1A, 7), 10-Q 2026-08-06, DEF 14A 2026-01-14, 8-Ks 2026-04-22, 2026-05-21, 2026-07-22, 2026-08-20, Form 4s (12m), Q3 FY2026 call transcript.
Desk stats - Revenue trend. Net interest income plus noninterest income was $839.9M in FY2025 (ended 30 September) vs $797.4M, up 5% (10-K Item 7). It has since turned down: nine months FY2026 $639.1M vs $653.1M, minus 2.1%; Q3 $189.6M vs $195.8M, minus 3.1% (8-K 2026-07-22). The cause is accounting, not demand: October 2025's consumer portfolio sale removed grossed-up interest income with offsets booked outside net interest income, cutting NIM from 7.43% to 6.59% (10-Q). Commercial finance interest income rose $6.1M and tax product revenue rose 13% to $107.7M over the same nine months. - Normalised after-tax profit. Start from FY2026 guidance of $7.80-$8.20 EPS (8-K 2026-07-22, slide 23). Add back the Q3 credit event: the CFO attributed roughly $28M-$30M of the quarter's $34M commercial finance provision to two specific reserves plus the related CECL build, "about $1 per share" (Q3 call). Subtract about $0.17 for secondary market revenue that ran $14.0M in Q3 against the company's own $5M-$7M quarterly normal. That is roughly $10.82 pre-tax at the guided 16-18% rate. Taxed at the desk's 25% rather than the solar-credit-inflated actual rate, normalised EPS is about $8.10, roughly $175M after tax, against FY2025 GAAP net income of $185.9M and $7.87 diluted. - EV / normalised after-tax profit. Not meaningful for a deposit-funded bank. The equivalent is price / normalised after-tax profit of 10.0x at 25% tax, 9.0x at the guided rate; on the FY2027 guide of $9.50-$10.00, 8.3x as guided and 8.9x re-taxed. - Financial-lane equivalents (no EV/EBIT). P/E 10.3x on FY2025 GAAP EPS; P/TBV 3.14x ($25.83 TBVPS at 30 June 2026); ROTE 21.61% in Q3 and 34.29% for nine months (Q3 call), from 30.65% in the September 2025 quarter; TBVPS up 16.9% year over year. Company CET1 11.51%, from 12.70%; Tier 1 leverage 9.66% (10-Q). - Growth sustainable? Organic and deposit-funded, but about half of guided FY2027 EPS growth is share count: at $9.75 on a plausible 19.5M shares, net income grows roughly 10% while EPS grows 22%. - What the screen got wrong. Its 9.1x P/E divides today's market cap, struck on 21.0M shares, by FY2025 net income earned on 23.6M average diluted shares; per share it is 10.3x. Its ROTE of 34.2% and net income growth of 1.4% are FY2025 figures, while nine-month FY2026 net income is down 6.9% to $137.0M and Q3 down 31%. It has no field for the material weakness in internal control disclosed at 30 September 2025, which forced restatement of FY2024, FY2023 and eleven interim periods (10-K Item 1A), so the XBRL history the screen ranks on was itself restated. Its TBVPS growth of 19.3% is flattered too: tangible common equity grew 7.1%, and the rest came from retiring 8.4% of the shares.
1. What the business actually does (from Item 1, in plain words)
Pathward is a $7.31B national bank in Sioux Falls that rents its charter to fintechs and keeps the deposits that come with it. Partner Solutions issues prepaid and debit cards, sponsors merchant acquiring and ATM networks, and runs tax-season refund advances and refund transfers for over 42,000 independent tax offices (10-K Item 1). Those deposits cost 0.01% in cash interest. The bank lends them through Commercial Finance: term loans, asset-based lending, factoring, leases and SBA/USDA guaranteed loans, $3.92B or 84% of a $4.67B book. It is a spread business bolted to a payments business, and the payments half is the franchise.
2. Why it is mispriced — the edge case
The candidate edge is a payments franchise being marked as a credit-impaired bank. The stock is 18.8% off its 52-week high after one quarter in which nonperforming assets went from 1.68% to 3.79% of assets, while the fee engine that produces most of the value never touched the bad loans. Management argues exactly this: "there's really no relationship between our credit book and our payment solutions partners" and "this is a reduction in net income, not an elimination of net income" (Q3 call, Pharr). The company bought 303,632 shares at $92.18 in the June quarter, 14% above today's price.
It is not enough. There is no forced seller: BlackRock, FMR, Vanguard, State Street and Dimensional hold about 44%, $16.5M trades daily, and there is no spin, index deletion or coverage gap. What there is instead is a company that will not size the problem. Asked for the balance and reserve on the workout loans, both CEO and CFO declined: "I don't think we're going to get into specific dollars on this" (Q3 call). That is a missing data point, not a misunderstood one, and it is what caps this at WATCH.
3. Credit quality, margin and deposit mix
Nonperforming loans went from $99.1M (2.05% of gross loans) at 30 September 2025 to $275.1M (5.28%) at 30 June 2026; classified assets from $258.6M to $391.5M; the allowance from $53.3M (1.14%) to $109.8M (2.15%), commercial finance from 1.18% to 1.71% (10-Q).
Almost all of it is one relationship. The increase "was related to certain renewable energy construction projects with a common developer" (10-Q), and the CFO put that developer at "at least 75% or 80%" of the sequential NPL rise (Q3 call). The disclosed ratios imply a roughly $5.2B book, so a sequential increase near $150M puts single-developer solar exposure on the order of $115M-$125M. That figure is derived, not disclosed. Context does not help: 60% of term lending exposure is solar and alternative energy, "most of which are construction projects" (10-K Item 1). Separately a working capital loan was written down for "a sophisticated fraud" and a legacy loan moved from workout to liquidation.
Coverage is thin. Of the allowance increase, $30.1M is the seasonal tax book, which reserves at 86.66% and unwinds; commercial finance carries roughly $68M against $255.4M of commercial nonaccruals, about 27%. Management's defence is that these are collateral-dependent and already written to expected net proceeds.
The franchise metric is stable. Reported NIM fell 84bp, but adjusted NIM, which charges the rate-linked card processing expense that 68% of deposits carry, was 5.27% against 5.33% (8-K 2026-07-22), and all-in deposit cost improved to 1.43% from 1.61%. New production yields 8.99% against a 7.42% book yield, with $150M of securities running off into loans over the next year (Q3 call).
4. Balance sheet and capital allocation
Holding-company debt is trivial: $19.9M of subordinated debentures and $13.7M of other borrowings against $149.4M of cash. Capital return is the story, with $923.8M of repurchases since 2Q19, 1,810,637 shares in nine months, and share count down from 22,953,608 to 21,023,902 year over year. The dividend is a token $0.05 a quarter (8-K 2026-08-20). The cost is CET1 down 119bp in nine months, and repurchases at 3.5x tangible book mechanically consume tangible book per share. FY2027 guidance cuts buybacks to 70-80% of net income; the CFO denies a credit motive ("no need to build an additional layer related to credit," Q3 call). Insiders own 1.43% as a group, with zero open-market purchases in twelve months against two director sales of 9,500 shares for $802K at $86.37 and $82.29.
5. Management: what they said vs what they did
Guidance history is honest. FY2026 was cut to $7.80-$8.20 explicitly because of the credit charges, and the CFO said so rather than adjusting them away. But the June 2025 call flagged one of these loans as shifting to nonperforming "with a path to work out," and a year later there was no path. The comparable claim now is "we are currently not aware of any other loans in our portfolio that have a similar unique fact pattern." Incentive design does not help: FY2025 bonus funded on net income (132% of target) and ROA (123%), long-term awards on three-year cumulative EPS and relative TSR (DEF 14A). No asset-quality metric appears, and EPS is the measure buybacks flatter most.
6. Valuation
Base (55%): FY2027 lands near the guide midpoint less $30M-$50M of further pre-tax solar charges, about $8.50 realised EPS at 10x, roughly $87. Bear (25%): a $120M exposure recovers half, credit spreads through the construction-stage energy book, and an unremediated material weakness draws a regulatory agreement restricting Partner Solutions growth and buybacks; $6.50 at 8x, roughly $52. Bull (20%): projects complete, no further reserves, top-of-guide $10.00, buybacks continue and a clean control opinion supports 12x, roughly $120. Probability-weighted $85 against $81.06, about 5% upside.
Reverse DCF: at $81.06 the market capitalises roughly $170M of sustainable after-tax earnings at about ten times, meaning it already grants management the FY2027 guide while pricing in no re-rating and about another year of elevated credit costs.
7. Catalysts and timeline
Q4 FY2026 results and the 10-K in late November 2026 carry three things at once: the first hard read on the solar workout, the conclusion on the material weakness, and confirmation or withdrawal of the FY2027 guide. Tax season builds through the December quarter and resolves in the March 2027 print. Liquidation proceeds on the two specific-reserve loans should surface over two to three quarters.
8. Risks and what would prove the thesis wrong (pre-registered kill criteria)
- Nonperforming loans stay above 5.0% of gross loans through the March 2027 quarter without the solar developer projects reaching completion.
- The FY2026 10-K reports the material weakness unremediated for a second consecutive year, or discloses a new one.
- Commercial finance net charge-offs exceed $25M in any single quarter, or the commercial finance allowance ratio exceeds 2.25%.
- Repurchases fall below 50% of net income for two consecutive quarters, or the company discloses any regulatory agreement restricting capital distributions.
9. Verdict and one-paragraph summary a smart friend could repeat
WATCH, conviction 3. Pathward earns 34% on tangible equity by renting its bank charter to fintechs for deposits that cost one basis point and lending the money out at 9%, and the screen's 9.1x P/E is right in spirit though really 10.3x once you stop crediting today's smaller share count with last year's net income. The payments half is fine: adjusted margin flat at 5.27%, tax revenue up 13%, new loans going on 157bp above the book yield. What broke is one corner of the loan book, where a single solar developer's half-built construction projects pushed nonperforming loans from 2.05% to 5.28% in nine months, and management will not say how large the exposure is or what it reserved against it. Backing into the disclosed ratios puts it near $115M-$125M against roughly $68M of commercial allowance, inside a term lending book that is 60% solar and alternative energy construction. Behind that sits an unremediated material weakness in internal control that already forced restatement of two fiscal years and eleven interim periods, which is how a credit event becomes a regulatory event. Probability-weighted value is $85 against $81.06, so you are not paid to underwrite an exposure the company itself declines to size, and the November 10-K settles both questions at once.
Research for discussion, not investment advice. Positions and sizing are the reader's decision.