The 1,000-origination threshold exempts most non-bank commercial finance and Farm Credit System gets a full pass. Federal relief means nothing if you still build the same data pipes for nine state regimes, and "reasonable methods" is the line that will bite at exam time.

The same-week comparison that makes the spine. Treasury Secretary Bessent announced an enhanced anti-discrimination eligibility framework for CDFI funding the prior week (April 28, 2026, covered in the prior Beyond Banks edition).⁹ The same federal apparatus that is loosening data-collection requirements on smaller commercial lenders is tightening eligibility for what Treasury characterized in its April 27 announcement as "predatory practices" by CDFI grant recipients. The two moves are not contradictory; they are aimed at different policy targets through different channels. The operator implication is that capital-source choice now carries asymmetric regulatory exposure: lenders sourcing from CDFI capital pools face tighter strings on use of funds, while lenders sourcing from depository or warehouse capital face lighter data-disclosure obligations on the lending side.

Federal exemption does not collapse your state-level build. Nine states have active commercial financing disclosure or registration regimes that operate independently of CFPB Section 1071 and that require their own data fields, retention periods, and filing cadence: California (CCFPL administered by DFPI, requires APR-equivalent disclosures and registration for commercial financing providers), New York (Commercial Financing Disclosure Law administered by DFS, registration plus disclosure), Virginia (Commercial Financing Disclosure Act effective 2023), Utah (registration regime), Georgia (commercial financing disclosure), Connecticut (commercial financing disclosure), Florida (commercial financing disclosure effective 2023), Kansas (registration), and Missouri (disclosure).^{10 11}

The build that matters is therefore not the federal data dictionary, but the union of all state-level dictionaries you operate in. CCFPL requires APR-equivalent calculation that the 2023 federal FIG did not require; the 2026 federal FIG removes the pricing data point entirely, which means tech teams that built APR fields under the 2023 rule still need to keep them for California. NY DFS requires named-recipient quarterly registration data that no other state requires. The Florida 2023 statute defines "covered transaction" differently from the federal Subpart B definition. The result is that a multi-state originator's compliance system has nine separate validators downstream of one origination event, and the federal rollback removes one validator (the most permissive one) without touching the other nine.

The strategic implication: any tech roadmap built on the assumption that the 2023 federal rule was the binding compliance constraint should be rescoped this quarter. The binding constraint is now whichever state regime in your footprint demands the most data, which for most multi-state MCA shops is California or New York. The federal rollback frees up budget at the federal-reporting layer; that budget is best deployed on state-by-state validator hardening, not on retiring the data pipes.

The Federal Register text permits institutions to estimate covered originations using "reasonable methods" when actual counts are not readily available, particularly for the look-back to 2025 origination data that was not collected against the new product definition.^{1 3} The phrase is doing significant load-bearing work and CFPB does not define it operationally in the rule text.

The exam-time risk runs in two directions. First, an under-estimate that brings an institution below the 1,000 threshold and out of coverage when actual originations would have crossed it exposes the institution to a coverage failure finding. CFPB and prudential regulators (OCC, FDIC, Federal Reserve, NCUA depending on charter) can require recreation of the data set going back to the relevant calendar year, with examination findings on the estimation methodology. Second, an over-estimate that pulls an institution into coverage when it should have been exempt costs the institution a compliance build it did not need, and there is no remedy for that beyond the regulatory cycle.

The defensible estimation methodology is the one that originator legal and compliance teams build with documented assumptions, applied consistently across estimation periods, with traceable source data. The undefensible one is a back-of-the-envelope number applied to make a coverage-call problem go away. Practical guidance for institutions in the 600-to-1,400-origination band on covered products: build the actual count, do not estimate. Institutions clearly above 2,000 or clearly below 500 can reasonably estimate from sampling. The middle band carries exam risk that is cheaper to retire by counting than by defending an estimation memo.

The MCA market is heterogeneous and the relevant question is not "is the market $10B" but "which sub-segments cross 1,000 originations per year on covered product types." Pure-play merchant cash advance originators (factoring rate plus daily ACH withdrawal) are exempt by product classification regardless of volume. The institutions that need to count are the hybrid platforms that originate both MCA-structured advances and term loans, lines of credit, or credit cards alongside.

Three segments cross the threshold materially. First, working capital lenders that originate term loans alongside MCA at scale (Credibly, OnDeck, Funding Circle US, Bluevine, Fundbox by product) easily clear 1,000 covered loan originations annually based on their disclosed origination volumes. Second, equipment finance shops that originate both equipment loans and equipment leases need to count the loan side (lease side is non-covered if structured as a true lease). Third, business credit card issuers including the fintech-issued cards (Brex, Ramp, Divvy, Mercury) clear the threshold on the card product alone.

The segments below the threshold and exempt: invoice factoring (factoring is non-covered per the rule), pure-play MCA without companion loan products, asset-based lending shops below 1,000 deals annually, equipment-only true-lease portfolios, and most regional commercial brokers and intermediaries who do not originate paper to balance sheet. The asymmetric outcome is that the largest specialty-finance shops (multi-product, multi-billion in originations) are in scope on the covered product slice; the niche shops (single-product or low volume) are out. The federal rollback narrows coverage from "essentially all small-business lending data" to "the largest multi-product specialty-finance platforms," which is closer to the original Dodd-Frank statutory intent than the 2023 rule was.

The structural test under Subpart B for what constitutes a covered "loan" or "line of credit" is whether the transaction creates a debtor-creditor relationship with a fixed-or-determinable repayment obligation. Merchant cash advance documentation that establishes a true purchase of future receivables (not a loan with repayment), with the operator's actual receivables as the only repayment source, with reconciliation rights tied to actual cash flow, and without absolute repayment obligation in the event of business failure, generally falls outside the loan definition. Revenue-based financing that uses fixed-percentage revenue capture without absolute repayment falls outside the line-of-credit definition for the same reason. Factoring of accounts receivable (true sale of invoices, not financing against invoices) falls outside both.

The contract language that holds up under both federal Subpart B and state-level true-lender or commercial-usury challenge typically includes: (1) explicit purchase-and-sale recital identifying the receivables sold, not borrowed against; (2) reconciliation provision that adjusts the daily or weekly remittance to actual cash collected, not a fixed amortization schedule; (3) bona-fide chargeback or non-recourse provision that allocates business-failure risk to the funder, not the operator; (4) absence of acceleration or default-judgment language tied to non-payment of a fixed principal balance.

The classification risk on hybrid products is highest where MCA documentation includes belt-and-suspenders provisions that look like debt remedies (personal guarantees with absolute repayment, COJ confessions of judgment, stipulated default amounts). California and New York courts have been the active recharacterization venues, and several MCA recharacterization rulings in 2024 and 2025 turned on exactly that kind of contract language. Originators in the 600-to-1,400-origination band on hybrid books should run a contract-language audit against the four-element test above before relying on the federal product exemption to keep them out of coverage. The federal exemption is only as good as the classification holding, and the classification is contract-language driven.

The required data fields under the 2026 rule are the statutory minimum (application date, application method as one of three remaining discretionary fields, action taken, action date, applicant census tract, applicant gross annual revenue, applicant ethnicity, applicant race, applicant sex, applicant business size, applicant principal owner identification, applicant North American Industry Classification System code, credit type, credit purpose, amount applied for, amount approved or originated, application date, action taken date) plus three remaining discretionary fields after the bureau dropped the rest.^{1 2}

Retention is three years from the date the institution submits the annual report to CFPB, with applicant identification information shielded from the underwriting decision-makers and held in a separate access-controlled environment per the Subpart B firewall requirement.¹ The exam-defensible format is structured CSV or fixed-width matching the FIG specification, with field-by-field validation against the published edits, and with version control on the FIG edition the institution built against (since the FIG is being reissued, this is non-trivial). Most institutions covered under the 2023 rule built against FIG 1.0; the 2026 reissue is expected to be FIG 2.0 with materially different field set. A defensible exam package shows the rule version, the FIG version, the validation rule set, the exception log for any field-level edit failures, and the final filed report. Originators that have been quietly building 2023 FIG 1.0 should plan to throw out the build that touches the deleted discretionary fields and rebuild to FIG 2.0 specification when CFPB releases it.

The two policy moves operate on different lender categories through different channels. The 1071 rollback reduces compliance cost on the lending side for non-bank specialty finance and community banks. The CDFI tightening adds anti-discrimination eligibility strings on the funding source for community-focused lenders that depend on Treasury-administered capital pools. A specialty-finance platform that funds itself via warehouse lines from money-center banks and securitization markets gets the deregulatory tail wind without the funding-side strings. A CDFI-funded community lender or non-bank that participates in Treasury programs gets both: lower lending-side disclosure obligations and tighter funding-side eligibility.

The capital-allocator question is not "is the regulatory environment friendly" because the answer depends entirely on which channel the capital flows through. The cleaner read is: depository capital and capital markets capital just got cheaper to deploy at the lending layer (less data overhead). Treasury-administered capital just got more conditional on the use of funds. Allocators with mandate flexibility should reweight toward the channels that benefit at both ends; allocators with CDFI or Treasury-program mandates need to budget for additional eligibility-monitoring overhead on the funding side. The Figure/Credibly announcement covered in this edition's headlines is the early live example of the benefit-at-both-ends configuration: an MCA and revenue-based-financing originator that sits outside the new federal data dictionary by product classification, and that just gained a parallel on-chain warehouse channel (Figure Connect) that may push warehouse spreads in originator-favorable directions if the Q3-Q4 2026 KBRA performance prints support it.

_{Sources
1 Federal Register: Small Business Lending Under the Equal Credit Opportunity Act (Regulation B), 2026-08494, May 1, 2026
2 Baker Donelson: CFPB Finalizes New 1071 Small Business Lending Rule, Key Takeaways
3 Steptoe: CFPB Substantially Scales Back Section 1071 Small Business Lending Rule and Resets Compliance Timeline
4 ABA Banking Journal: CFPB Finalizes Streamlined Small Business Lending Data Rule
5 JD Supra: CFPB Finalizes New 1071 Small Business Lending Rule
6 CU Times: CFPB Scales Back Small Business Lending Rule, Easing Burden on Credit Unions
7 CFPB: 1071 Rule Page
8 Consumer Federation of America: Statement on CFPB Revised Final Rule
9 Beyond Banks April 28, 2026: Treasury CDFI Predatory Review (companion edition)
10 California Department of Financial Protection and Innovation (DFPI), CCFPL Administering Agency
11 New York Department of Financial Services: Commercial Financing Disclosure Law Industry Letter
12 National Mortgage Professional: CFPB Scales Back Small Business Lending Rule
13 CFPB: Rules Under Development, Section 1071 Small Business Lending Data Collection
14 Auto Finance News: Auto Lenders Tighten Standards as Nonprime, Prime Delinquencies Up
15 Bloomberg Law: FAT Brands Creditors Challenge Bankruptcy Sales, Lender Takeover
16 PYMNTS: Credibly Taps Figure Marketplace to Expand Capital Access
17 Crowdfund Insider: Figure Teams Up With Credibly to Enhance SMB Financing with Blockchain Tech
19 GlobeNewswire (Figure / Credibly press release with KBRA securitization count and Q1 2026 $124M / $225M deal sizing)}

Neither catches a sitting court judgment.

A money judgment can land on the court docket months before it appears in any credit feed. By the time it surfaces, you've already wired.

Lenders working New York portfolios feel this most: the state's civil court system is split across five boroughs with separate dockets, making manual checks prohibitively slow.

Cobalt Intelligence's Court Cases API queries NY State and Miami-Dade County live, returning judgment details, parties, and dollar amounts before funding decisions are made.

Read more: Cobalt Intelligence Court Records Search - NY State Coverage

or

Schedule a FREE demo call

Auto Finance News reports subprime 60+ DPD at 6.65% (October 2025), the highest level since the early 1990s, with prime borrowers at 0.37 to 0.6%. The prime crossover is the relevant signal: when the top tier starts missing, the macro stress is no longer subprime-concentrated. Originations at lenders with subprime auto books fell in Q3 2025 as approval thresholds tightened.

For MCA, equipment finance, and working capital shops, the auto book is a leading indicator on the same SMB-owner population that pays the working-capital advance. If your borrowers are missing $500-to-$750 monthly car payments, the cash-flow stress shows up in the business repayment six to nine months later. Tighten payment-to-income thresholds, flag auto-stressed applicants in underwriting, and price the next vintage for the cycle, not the prior one.

According to the May 5, 2026 filing by the official committee of unsecured creditors in the Southern District of Texas Bankruptcy Court, the proposed $1B sale of FAT Brands operating units to a bondholder group is being challenged on the theory that approximately $195M in alleged manager advances from the FAT Brands parent entity to its securitization vehicles, growing from $11M in 2021 to $65M by Q3 2025 per the committee's filing, should receive payment priority over the bondholders' credit bid. According to court filings, 352 Capital (a $100M+ debt holder) separately filed suit contesting FAT Brands' proposed use of management fees and residual cash from the securitization vehicles to fund Chapter 11 operations.

The proposed sale is structured as $359.5M (Twin Peaks debt-to-equity), $595M (Fat Brands operating umbrella debt-to-equity), $8M cash to a Las Vegas buyer for Hot Dog on a Stick, $2.5M cash to a Kuwait buyer for Elevation Burger, with Smokey Bones closed. Wilmington Savings Fund Society serves as securitization noteholder trustee. Per court filings, founder Andy Wiederhorn took a leave of absence as a condition of bankruptcy financing. The structural-precedent question for alt-lenders: if SDTX accepts the manager-advance recharacterization theory, every alt-lender holding senior unsecured paper at a sponsor with off-balance-sheet securitization vehicles has a new recovery-claim due-diligence question. Tighten covenants on parent-to-SPV intercompany cash-flow disclosure quarterly, not annually, on borrowers with similar structures.

Figure Technology Solutions announced May 5, 2026 that Credibly (over $3B lifetime SMB origination, four KBRA-rated securitizations per Credibly's own announcement, including a Q1 2026 deal sized at $124M expandable to $225M per the same release¹⁸) will tokenize its SMB loan and revenue-based financing products onto Figure's three-layer stack effective Q2 2026: Democratized Prime (on-chain warehouse marketplace), Figure Connect (whole-loan tokenization for sales and securitization), and DART (blockchain-native eNote and lien registry replacing MERS-style legacy infrastructure for non-mortgage assets). Underlying chain is Provenance Blockchain, named in the Figure Markets corporate materials but not in the press release framing.

This is Figure's second Democratized Prime borrower partnership after auto fintech Agora Data (February 2026). For working capital lenders funded off bank warehouse lines, the question is whether Figure Connect produces measurable spread improvement on Credibly's KBRA execution; if it does, every alt-lender's warehouse comp gets renegotiated within two to three quarters. The first hard data point is the Q3-Q4 2026 KBRA reporting on Credibly's tokenized tranche performance, including whether DART recordation gets methodology recognition or a haircut in the rating action. State-level regulatory questions on tokenized SMB lending, true-lender treatment for non-bank token holders, and FinCEN considerations are unresolved; California DFPI and New York DFS are the most likely first-mover guidance issuers.