What changes is the second layer: banks answer for loans and credit losses; payments and fintech answer for fee-revenue recognition and non-GAAP.
Five things to take away
- One industry on paper, two on the staff’s desk. 234 of 370 letters went to depository banks on credit and loans; the other 136 went to payments, fintech, and consumer finance on fee revenue and non-GAAP.
- MD&A driver quantification is the universal spine, present in 35% of all letters and ranked first in every tier. The staff keeps rejecting “volume,” “mix,” and “TPV-driven” explanations without quantification.
- The signature objection is tangible-common-equity-ex-AOCI, which the staff treats as an individually-tailored non-GAAP measure under Reg G, now policed down at community banks, not just large ones.
- Revenue recognition is the fintech mirror of the banks’ loan question, 24% of letters, with principal-vs-agent and disaggregation depth rising with company size (BILL, PayPal, Discover, FICO).
- Segment reporting (ASU 2023-07) is the clear 2026 riser, a top-three theme only from 2025; fair value (ASC 820) has faded out of the top ten.
Two industries, one rulebook
The same Division of Corporation Finance reviews both halves of this sector, but the questions split cleanly by business model. Banks get asked about credit; payments and fintech get asked about revenue. Only one theme sits on both desks: MD&A.
| Track | Population | The review centers on |
|---|---|---|
| Bank track | 234 letters | The loan book: allowance roll-forwards, charge-offs, CRE concentration, CECL forecast assumptions, interest-rate sensitivity (Item 305), and the staff’s signature objection to TCE measures that strip out unrealized securities losses. |
| Payments & fintech | 136 letters | The income statement: how fee revenue is recognized under ASC 606 (gross vs net, principal vs agent), disaggregated, and reconciled to promoted metrics; non-GAAP prominence; crypto-asset disclosure for exchanges and custodians. |
What does the SEC ask financial-services companies about most?
Eight themes account for nearly all substantive comments. Three are bank questions, three are payments-and-fintech questions, and MD&A sits across both:
Share of the 370 financial-services letters raising each theme.
| Theme | What the staff asks for | Standard | Example |
|---|---|---|---|
| MD&A drivers | Quantify each revenue and expense driver and the offsetting factors within a line item; do not attribute change to “volume” or “TPV.” | Reg S-K Item 303 | PayPal, SoFi |
| Revenue recognition | Disaggregate fee revenue; document the principal-vs-agent (control) analysis; separate subscription from transaction. | ASC 606-10-55 | BILL, FICO |
| Loans & credit losses | Full allowance roll-forward, charge-off and nonaccrual detail, CRE concentration with LTV, CECL forecast assumptions. | ASC 326 / 310 | Old National, Triumph |
| Non-GAAP | Remove individually-tailored capital measures; give GAAP equal prominence; justify adjusted-EBITDA add-backs. | Reg G / Item 10(e) | Repay, Atlanticus |
The MD&A spine: explain the drivers
One theme outranks every other at every tier. The staff’s instruction is always the same, applied to spread businesses and fee businesses alike: quantify what moved the number, and the factors that offset it, rather than naming a metric and stopping.
“Do not attribute a revenue change to ‘volume,’ ‘mix,’ or total payment volume without quantifying the drivers and the offsetting factors within the line item.”
The recurring Item 303 instruction, paraphrased across 130 letters in Finrep’s population.
PayPal “typically attribute[s] changes in revenues to changes in TPV” even though it states TPV “does not directly relate to transaction revenues.” Advise or revise the MD&A.
Explain why “shifting to a variable-priced product resulted in an increase in revenue,” and whether subscription and transaction revenue should be separately disclosed.
Quantify period-over-period drivers, disclose offsetting factors within a line item, and add granularity by loan type for origination, sales, and securitization income.
Quantify how much of the change in the provision for credit losses came from higher GMV versus tightened consumer underwriting.
The lesson is portable across the sector: every promoted metric (TPV, GMV, ACV, take-rate) must reconcile to the GAAP line it claims to explain. Where it does not drive the line, the staff asks the company to stop attributing the line to it.
Trends by company size tier
| Tier | Population | Top themes | Named examples |
|---|---|---|---|
| Micro-cap under $300M | 31 filers · 69 letters | MD&A 38% · Loans 28% · RevRec 25% · Non-GAAP 20% | Blue Ridge Bankshares (going-concern + BaaS wind-down) · Repay (adjusted-EBITDA add-backs) |
| Small-cap $300M–$2B | 59 filers · 133 letters | MD&A 38% · RevRec 25% · Loans 20% · Non-GAAP 17% | Triumph Financial (factored receivable, no allowance) · Flywire (non-GAAP prominence) |
| Mid-cap $2B–$10B | 47 filers · 101 letters | MD&A 32% · Loans 23% · RevRec 22% · Non-GAAP 11% | Old National (TCE-ex-AOCI removal) · BILL (subscription vs transaction) · Synovus (Item 305) |
| Large-cap over $10B | 24 filers · 60 letters | MD&A 37% · RevRec 28% · Non-GAAP 15% · Segment/goodwill 10% each | PayPal (TPV driver dispute) · American Express (rate sensitivity) · Coinbase (Circle agreement exhibit) |
Why company type maps onto size: micro and small tiers are dominated by community banks and emerging fintech, so loan-table and revenue-policy questions cluster there. The mid tier is regional banks plus scaled fintech, where credit-loss disclosure climbs to near-parity with revenue recognition. Large-cap is networks, money-center banks, and big processors, the most technically demanding reviews.
Three years of drift: what’s rising, what’s fading
2024 is the most complete year and should anchor any read. MD&A and loan/credit comments rose with the share of banks under review; segment reporting is the new arrival in 2025.
- ▲ Rising: Segment reporting (ASU 2023-07) surfaced as a top-three theme only from 2025, as first-year significant-expense disclosures cleared review. Crypto fair value (ASU 2023-08) persists for exchanges and custodians.
- ▼ Fading: Fair-value measurement (ASC 820), long the most-commented topic overall, dropped out of the 2025 top ten. Going-concern comments stayed confined to micro-cap distress and did not broaden.
2026 letter counts are partial: EDGAR publishes staff letters on a roughly 20-business-day lag, so recent months are undercounted.
The signature objection: TCE-ex-AOCI
If one comment defines SEC review of this industry, it is the rejection of bank capital measures that exclude accumulated other comprehensive income and unrealized securities losses. The staff treats them as individually-tailored non-GAAP measures.
The staff’s position: a tangible-common-equity figure that strips out unrealized securities losses is an individually-tailored recognition and measurement method, prohibited under Reg G.
Citing C&DI Question 100.04 and Rule 100(b) of Regulation G, the staff has repeatedly asked banks to remove these measures from future filings, or be ready to defend why they are not tailored. The objection began at larger banks and is now policed routinely at community banks: Old National, Virginia National, Embassy Bancorp, Patriot National, Capitol Federal, Princeton Bancorp.
What to do before filing
Remove tangible-common-equity and capital measures that exclude AOCI or unrealized securities losses, or document the defense in advance. Confirm every non-GAAP measure gives the GAAP figure equal or greater prominence, with a clear reconciliation. (See the SEC’s non-GAAP C&DIs.)
Methodology & population
Every EDGAR “UPLOAD” letter (the staff’s outgoing comment letters) filed June 2023 to June 2026, filtered to payments and financial-services filers by SIC code and name, limited to 10-K and 10-Q reviews. Registration and offering reviews were excluded by design. From ≈36,000 UPLOAD letters scanned, the financial-services filter yields 370 periodic-review letters across 164 filers, 346 on Form 10-K, 24 on 10-Q. By year: 84 (2023) / 235 (2024) / 51 (2025–26 partial).
- In scope: commercial banks and thrifts (SIC 6020–6036), consumer and business credit (6141–6172), money transmission (6099), finance services and fintech (6199), insurance brokers (6411), and named payments/fintech filers in software SIC codes (Block, PayPal, Visa, FIS, Marqeta, BILL, Sezzle, Coinbase, and peers).
- Out of scope: SPAC and blank-check shells, asset-backed and structured issuers, crypto miners and crypto-treasury vehicles, generic enterprise software, property-casualty and life underwriters, and all registration and offering reviews.
- Themes are keyword-tagged and confirmed against letters read in full.
Frequently asked questions
How many SEC comment letters did financial-services companies receive from 2023 to 2026?
370 comment letters across 164 distinct filers, per Finrep’s analysis of EDGAR UPLOAD letters tied to 10-K and 10-Q reviews between June 2023 and June 2026. 234 went to depository banks and 136 to payments, fintech, and consumer-finance companies.
What is the most common SEC comment topic for banks?
After MD&A driver quality (first everywhere), banks answer most for loans and receivables (18% of letters): allowance roll-forwards, charge-off detail, CRE concentration with LTV, and CECL forecast assumptions, plus the recurring Reg G objection to TCE measures that exclude AOCI.
What is the most common SEC comment topic for fintech and payments companies?
Revenue recognition under ASC 606 (24% of letters): gross-versus-net and principal-versus-agent analyses, disaggregation of fee revenue, and separating subscription from transaction revenue. Non-GAAP prominence (15%) is the companion question.
Is TCE excluding AOCI a permitted non-GAAP measure?
The SEC staff’s position, citing C&DI 100.04 and Rule 100(b) of Regulation G, is that tangible-common-equity measures stripping out unrealized securities losses are individually-tailored and prohibited, and it now raises this at community banks, not just large ones.
