Credit Intermediation and Related Activities (U.S., NAICS 522)

A Histometrics rollup primer for public- and private-market investors. NAICS (North American Industry Classification System) 2022 code 522 is a subsector — a three-digit code — that gathers everything American finance does around lending: taking deposits and making loans, lending without deposits, and running the fee-and-plumbing services that sit around credit and payments. This page synthesizes across its three child industry groups (5221, 5222, 5223); it does not repeat the leaf detail. Scope: United States, current through mid-2026. Figures are cited facts; statements about the future are labeled "judgment."

1. Overview

NAICS 522 is, in plain terms, the credit and payments machine of the U.S. economy — the subsector that holds the nation's deposits, writes most of its loans, and moves and settles most of its money. It sits inside sector 52 (Finance and Insurance) and is defined by a single job: intermediating credit — standing between savers and borrowers — plus the services that make that possible. What it deliberately leaves out is the rest of finance: securities and investment (subsector 523), insurance (524), and funds and trusts (525) live next door. 522 is credit.

The distinctive thing about this level is not any one child but the clean three-way split in how the job gets done, organized by a single question — who takes the credit risk, and how is it funded?

• 5221 Depository Credit Intermediation — lenders that take the risk and fund it with insured deposits (banks, credit unions, thrifts).
• 5222 Nondepository Credit Intermediation — lenders that take the risk but fund it wholesale, without deposits (card issuers, captive auto/equipment lenders, nonbank mortgage, the mortgage secondary market).
• 5223 Activities Related to Credit Intermediation — firms that take little or no credit risk and earn fees on the flow (payment networks and processors, loan brokers, loan servicers, money transmitters).

Two of the three are lenders funded two different ways; the third is the fee-earning plumbing around them. That framing — deposit-funded lending / wholesale-funded lending / fee services — is the spine of the whole subsector, and it drives everything downstream: the economics, the owners, the regulators, and the way an investor gets in. Section 2 leads with the comparison; the rest of the page treats 522 as a whole.

2. What's inside — the three children and how they differ

NAICS 522 contains exactly three industry groups (four-digit codes). Together they cover the entire U.S. credit-and-payments system. They are near-equal on some measures and wildly unequal on others, and the pattern of where they match and diverge is the whole point.

GSE = government-sponsored enterprise (Fannie Mae, Freddie Mac); ABS = asset-backed security; MBS = mortgage-backed security. Receipts and firm counts are the 2022 Economic Census; jobs are 2023 County Business Patterns [1][2].

Four contrasts to carry through the rest of this primer:

1. The two lenders are a near-tie on receipts — but not on the balance sheet. Depository (5221) and nondepository (5222) each book about 44–45% of the subsector's income, within a rounding error of each other. That parity is misleading. On the balance sheet, which is how you actually size a lender, depositories dwarf nondepositories — U.S. banks, credit unions, and thrifts hold roughly $29 trillion in assets [4][7][8]. Nondepositories earn nearly as much income off a far smaller book because they lend at higher rates (card APRs, subprime auto), and because the biggest bank-owned card and auto portfolios are classified inside 5221's Commercial Banking, not 5222 [4][5]. Read the receipts tie as "two big income pools," never as "two equal-sized lenders."

2. Dollars and firms are inverted. Rank the children by revenue and 5221 is first, 5222 a close second, 5223 a distant third. Rank them by number of firms and the order flips: 5223 has the most companies (16,131), 5222 next (13,616), and 5221 the fewest (9,076) — because the fee tier is a long tail of tiny loan-broker shops while depository banking is a smaller number of much larger institutions. Receipts per firm runs about $80 million in depository, $53 million in nondepository, and just $11 million in the fee tier [1][2]. Big-and-few versus small-and-many is the ownership story in one line.

3. Ownership could hardly differ more — and one whole child cannot be bought. Depository value splits between listed bank holding companies and member-owned credit unions that have no stock at all [4]. Nondepository is a patchwork of public card pure-plays, wholly-owned manufacturer captives, and government-controlled GSEs under conservatorship [5]. Related activities pairs a handful of world-class listed payment networks with member-owned clearing utilities and thousands of private brokers [6]. There is no single "owner profile" for 522 — which is exactly why you invest in the child, not the code.

4. Interest rates move all three — in different directions. A falling-rate wave is a tailwind for a mortgage broker (refinancings) and for banks with cheap deposits, but a headwind for the same period's loan servicers (their servicing rights lose value) and a mixed bag for wholesale-funded lenders. The rate cycle is the one variable that touches every child, and it does not push them the same way — so "rates are falling" is not a 522-wide thesis.

The distinctions that split investors within each child — bank charter type, captive versus independent, network versus processor — live in the leaf primers, not here.

3. Size — this level's rollup figures

These are our ingested federal ground-truth figures for NAICS 522 specifically. The three children sum to the subsector almost exactly, which is worth showing because it validates every share in Section 2:

At roughly $1.62 trillion in receipts, 177,000 locations, and 2.95 million workers, 522 is one of the largest slices of the U.S. economy — and average pay of about $104,000 [1] marks it as a high-skill, high-wage sector, though the average hides a real spread (retail bank branches at the low end, payment-network engineers at the high end).

The measurement caveat — receipts are income, not the size of the credit market. For the two lending children the Census "receipts" line is roughly interest plus fee income, not balances or credit outstanding. A lender's economic weight sits on its balance sheet: depository assets alone are about $29 trillion [4][7][8] — nearly 18 times the $1.62 trillion receipts figure — and the loans the nondepository firms originate and securitize (U.S. mortgage debt ~$21 trillion, auto ~$1.67 trillion, cards ~$1.28 trillion) each dwarf the relevant child's revenue [5]. These receipts must never be added to household- or mortgage-debt totals; they measure the operating income of the businesses, not the flow of credit they intermediate. When you size any lending part of 522, use assets, deposits, and loan balances — never receipts.

Undercount caveats — heaviest exactly where small and individual ownership dominates. Three gaps matter, and each falls on a different child:

• Employer-only frame misses the long tail. County Business Patterns and the Economic Census count only businesses with paid (W-2) staff. That undercounts the smallest, individually run operators in every child: volunteer-run credit unions with no payroll (5221), sole-proprietor mortgage shops, single-branch payday and pawn storefronts (5222), and 1099 commission-only loan originators plus tiny cash-access shops (5223) — where licensing data recorded roughly 221,000 active mortgage loan originators against ~55,000 counted broker employees [4][5][6]. The true long tail is larger than 176,966 establishments and 38,314 firms.
• Government "plumbing" is excluded. The Census counts private and shareholder-owned businesses. It leaves out the government corporations that do this exact work: the Federal Reserve's own clearing (FedACH, Fedwire, FedNow), Ginnie Mae (over $2.8 trillion of guaranteed MBS), the Farm Credit System (~$438 billion in loans), and the Export-Import Bank [5][6]. Fannie Mae and Freddie Mac are counted (they are shareholder-owned), which is why 5222's receipts are so large — but much of the surrounding federal credit apparatus is not.
• Mixed vintages, and a small firm-count gap. Receipts, firms, and concentration are 2022; establishments, employment, and payroll are 2023 — treat the two snapshots as adjacent, not simultaneous. The children's firm counts sum to ~38,823 against a level total of 38,314 because a company operating in more than one industry group is counted once at the subsector; treat the 38,314 subsector figure as authoritative.

Concentration — unconcentrated, and more so than any child. For the subsector, the top 4 firms earned 20.7% of receipts, the top 8 30.6%, the top 20 43.3%, and the top 50 56.4%, with a Herfindahl-Hirschman Index (HHI — a standard concentration score, summing every firm's squared market share, where anything under 1,500 is "unconcentrated") of just 162.8 [2]. That 162.8 is lower than any individual child (depository ~164, related activities ~388, nondepository ~516) — pooling three fragmented industries with different leaders and almost no firm spanning them produces an even more dispersed whole. Two familiar warnings apply: this is a national receipts view, whereas banking and payments competition is regulated at the local or product-market level and looks far tighter there; and measured by assets rather than receipts, the four largest banking companies control a much larger share of the system than the 20.7% CR4 suggests [4].

4. Investable universe — where value concentrates

Public-market value in 522 is not spread evenly. It concentrates in three pockets, each in a different child, and skips large stretches of the subsector entirely (credit unions, captives, brokers, government plumbing). Named rosters and multiples live in the leaf primers; the map is:

• Bank equity (in 5221) — the broadest set. The deepest opportunity is depository banking: the "universal" money-center banks, super-regionals, and dozens of listed regionals, plus a long private tail of community banks, reachable through individual stocks, bank ETFs, and bank preferreds/subordinated debt [4]. But credit unions contribute no listed equity — the largest are enormous and member-owned, and you cannot buy any of them; the only public proxy is the technology-vendor ecosystem that also serves banks [4].
• Payment networks and processors (in 5223) — the highest-quality economics in the subsector. The two global card networks and the scaled processors and merchant acquirers are capital-light, fee-earning toll-takers with network effects and, at the network end, operating margins above 60% — among the best economics in all of finance [6]. Much of the most systemically important infrastructure here, though, is not listed: member-owned clearing utilities and marquee private fintech. Loan brokers, at the other end of this child, have almost no pure public play — the real ownership opportunity is private [6].
• Card issuers and nonbank lenders (in 5222) — a concentrated set of public pure-plays, plus bonds. Listed card issuers, near-pure-play consumer and auto lenders, nonbank mortgage originator-servicers, and mortgage REITs give direct exposure; the manufacturer captives are reachable only through the parent company's stock; and the GSE commons trade over-the-counter as speculative bets on the terms of an eventual government exit, not ordinary stocks [5]. The dominant institutional instruments in this child are bonds, not equity — card, auto, equipment, and consumer-loan ABS, and agency MBS that already sit inside nearly every bond fund.

The synthesis for an allocator. Across all three children, public equity value clusters in bank franchises, payment toll-takers, and a handful of card/nonbank-lender pure-plays — while huge stretches of the subsector (credit unions, captives, GSEs, clearing utilities, the broker tail) are unbuyable in the stock market and reachable only through debt, the parent company, or private ownership. In private and institutional markets the dominant 522 exposure is securitized debt and servicing — ABS, MBS, whole loans, and mortgage servicing rights — plus private-equity ownership of community banks, captive and mortgage platforms, and late-stage fintech. A single discipline runs through all of it: never equate a firm's payment volume, origination volume, or loan balance with its revenue or with Census receipts — those are always far larger than the fee-and-interest income this subsector actually measures.

5. How the money works

522 runs three different economic engines, and telling them apart is most of the analysis:

• Depository (5221) — net interest margin on a deposit-funded book. Interest earned on loans and securities minus interest paid on deposits and borrowings equals net interest income; against average earning assets that spread is the net interest margin (NIM), running around 3.3–3.4% across the deposit system in recent data [4]. Cheap, sticky, insured deposits are the raw material and the moat; roughly 10x leverage turns a ~1% return on assets into a double-digit return on equity. Credit unions run the same engine tax-free and answer to members, not shareholders [4].
• Nondepository (5222) — spread-and-fee on wholesale funding, often sold not held. With no deposits, these lenders borrow wholesale (warehouse lines, ABS, MBS, repo) and many originate to distribute, selling or securitizing loans rather than holding them. The formula is interest and fees, plus gains on selling loans, minus funding cost, minus credit losses [5]. Because funding is market-based, profitability is unusually sensitive to two external variables at once — the cost/availability of wholesale funding and the credit-loss rate — and both turn against owners in the same downturn. The number that matters is the spread after credit losses, not the headline yield.
• Related activities (5223) — fees on flows, low credit risk. These firms monetize volume and services, not balances, so bank-style NIM does not apply. Networks and processors keep a take rate in basis points on payment volume; servicers earn a servicing fee (~0.25% a year of unpaid principal) whose core asset, the mortgage servicing right, gains value when rates rise; brokers earn a commission per closed loan with almost no balance-sheet risk and high operating leverage [6]. The through-line is network effects and operating leverage — the rails are built, so incremental volume drops to profit at very high rates.

The unifying idea. Move left to right across 522 and you trade balance-sheet risk for fee income: depositories carry the most credit and duration risk cushioned by insured funding; nondepositories carry credit risk on riskier wholesale funding; related activities shed most credit risk to earn a toll on the flow. Different master metric for each — NIM (5221), spread after credit losses (5222), take rate / fee per transaction (5223).

6. Demand drivers

Because all three children ultimately ride the credit cycle, the same macro forces move the whole subsector — but, as Section 2 warned, not always in the same direction:

• Interest rates and the yield curve — the master variable. The Federal Reserve cut three times in late 2025 to a roughly 3.50%–3.75% federal funds rate [16]. A steeper curve widens the borrow-short/lend-long spread that feeds depository NIM; rates set funding cost, affordability, and credit for nondepository lenders; and they move brokering and servicing in opposite directions inside the fee tier (a rate drop lifts refinancings but cuts servicing values).
• Consumer spending, employment, and the credit cycle. Loan demand, card volume, and delinquencies all track jobs and incomes; a healthy consumer feeds every child, a weakening one lifts charge-offs first in cards and subprime.
• Housing, vehicle, and equipment sales. No purchase, no purchase-tied loan; the housing/mortgage cycle drives the thrift book, the nonbank-mortgage machine, and the servicing/broker fee tiers together.
• The cash-to-digital shift. The multi-decade migration from cash to cards and digital payments is a durable structural grower for the fee tier — the Federal Reserve counted 236.6 billion U.S. noncash payments worth $140.0 trillion in 2024 [12].
• Bank-to-nonbank migration. Lending keeps shifting from deposit-funded banks (5221) to wholesale-funded nonbanks (5222), which now originate most U.S. home loans and a rising share of consumer credit — a slow, persistent transfer of share within the subsector.

7. Regulation

Credit intermediation is among the most heavily supervised activities in the country, but there is no single regulator for 522 — oversight follows the activity and the charter, and that fragmentation is itself a cost and a barrier to entry:

• Depository (5221) — the full prudential overlay. The Office of the Comptroller of the Currency (OCC) charters national banks; the Federal Reserve supervises holding companies and sets monetary policy; the FDIC insures deposits up to $250,000 and resolves failures; the National Credit Union Administration (NCUA) charters and insures credit unions through a parallel $250,000 fund; and state regulators oversee state charters [4]. Frameworks include Dodd-Frank, Basel III capital and liquidity standards, and the Community Reinvestment Act; the unfinished "Basel III endgame" capital rules were re-proposed in March 2026 in a lighter form [4].
• Nondepository (5222) — patchy and uneven. The Consumer Financial Protection Bureau (CFPB) administers the consumer rulebook (Truth in Lending, the CARD Act, RESPA, Ability-to-Repay); roughly 20 states cap small-loan APRs near 36%; but no federal safety-and-soundness regulator oversees nonbank mortgage firms — a gap the Financial Stability Oversight Council (FSOC) and GAO flagged in 2024–25 [5][14]. The Federal Housing Finance Agency (FHFA) runs Fannie Mae and Freddie Mac as conservator [5][9].
• Related activities (5223) — activity-licensed. Money transmitters register with the Financial Crimes Enforcement Network (FinCEN) and run Bank Secrecy Act / anti-money-laundering programs under state Money Transmitter Licenses; mortgage servicers fall under CFPB and GSE rules; loan brokers under the SAFE Act licensing regime; and payments faces Federal Reserve interchange rules, the 2025 GENIUS Act stablecoin framework, and a live DOJ antitrust suit against Visa [6].

The common threads. Two regimes reach across all three children: the CFPB's consumer-lending rulebook (any consumer-facing lender or servicer, whatever its charter) and the Bank Secrecy Act / AML regime. The federal posture swung deregulatory in 2025 (CFPB leadership change, deprioritized enforcement), even as the durable, hard-to-reverse constraints stayed at the state level [5][6].

8. Consolidation

The defining structural fact across all of 522 is that the number of firms falls while the survivors get larger — driven by the rising fixed cost of technology, compliance, and scale. The logic differs by child, but 2025 was a landmark year, and the marquee deals cut across more than one code at once:

• Depository (5221) consolidates to spread technology and compliance costs — U.S. commercial banks fell from over 8,000 around 2000 to roughly 3,900 today, thrifts are in secular decline, and cash-rich, tax-advantaged credit unions have become aggressive buyers of small community banks [4].
• Nondepository (5222) consolidates toward scale in cards and mortgage, with the biggest single question being not a merger but the possible partial privatization of the GSEs, which would redraw ownership of the subsector's largest single revenue pool [5].
• Related activities (5223) consolidates toward network scale in payments and toward larger servicing books, while the loan-broker tier stays structurally fragmented [6].

Two deals that reshaped three codes at once. Capital One's ~$35 billion acquisition of Discover (closed May 2025) touches all three children — it made Capital One the largest U.S. card issuer (5222), acquired Discover's payment network (5223), and did so as a bank holding company (5221), nudging the Visa/Mastercard duopoly toward a "triopoly." Rocket's $14.2 billion acquisition of Mr. Cooper (closed October 2025), the largest independent-mortgage deal ever, joined the biggest nonbank mortgage originator (5222) to the biggest servicer (5223) in one book [5][6][10][11]. The durable advantages that decide these contests are the same across the subsector: cheap and reliable funding, underwriting-data quality, servicing and compliance infrastructure, and charters and licenses that are hard to replicate.

9. Risks

The hazards share a common spine across 522 — credit losses, funding, interest rates, concentration, and policy — with a different sharpest edge for each child:

• Credit cycle (dominant for the two lenders). Rising unemployment lifts delinquencies and charge-offs; cards and subprime consumer/auto books and office-heavy commercial real estate are hit first, and losses lag the loans that caused them [5]. A credit union can only rebuild capital from retained earnings, making it the least able to absorb a shock [4].
• Funding and liquidity (sharpest for 5222). Depositories can suffer deposit runs — now measured in hours given digital banking, as Silicon Valley Bank showed in 2023 — but at least hold insured funding and Fed access; nondepositories have neither deposits nor a lender of last resort, so when ABS, warehouse, and repo markets seize, they cannot fund or offload loans, and servicers face margin calls exactly when cash is scarce [5][14].
• Interest-rate and duration risk. The "borrow short, lend long" mismatch and unrealized securities losses weigh on depositories; prepayment and servicing-value swings whipsaw the fee tier; the direction that helps one child hurts another [4][6].
• Fee/take-rate compression and disintermediation (sharpest for 5223). Interchange politics, merchant litigation, digital money-transfer competition, and real-time account-to-account rails and stablecoins could route volume around today's fee earners over time [6].
• Concentration and systemic importance. Two GSEs stand behind trillions in mortgages and two networks behind most card volume — a stumble at any of them is a national event [5][6].
• Policy risk. A serious move against the credit-union tax exemption, a heavier Basel endgame, a GSE conservatorship exit, state 36% rate caps, and interchange legislation all sit outside management control and could reprice a whole child overnight [4][5][6].
• Model, operational, cyber, AML, and private-market opacity. These firms are critical infrastructure holding sensitive data; a clearinghouse outage or AML failure is severe; and private lenders, funds, and securitization vehicles report infrequently and are hard to value — the Federal Reserve has flagged rising bank interconnection to private-credit funds and BDCs (business development companies) [15].

10. How to invest, and the outlook

There is no single "buy 522" trade — the subsector is three engines, and the route depends entirely on which one you want:

• Deposit franchises (5221): individual bank stocks (money-centers for scale and diversified fees, regionals for leverage to the local credit cycle), bank preferreds and subordinated debt for income, and bank ETFs; privately, community-bank equity. Credit unions are not investable — a member's return is priced into the product, and the only public proxy is the technology-vendor ecosystem [4].
• Wholesale-funded lenders (5222): listed card issuers and consumer/auto lenders directly; manufacturer captives only through the parent's stock; nonbank mortgage originator-servicers and mortgage REITs for origination and MSR cyclicality; the GSE commons as a speculative, binary policy bet. The dominant institutional route is debt — card, auto, equipment, and consumer ABS, and agency MBS already inside most bond funds [5].
• Fee toll-takers (5223): payment-network compounders (premium multiples, toll-taker economics), cheaper and more cyclical scaled processors, higher-volatility fintech platforms, and value/yield money-transfer names; servicers as a rate/MSR bet; and thematic payments/fintech ETFs for diversified exposure. Loan brokers have no clean public play — the real ownership is private [6].

Private and institutional routes across the whole subsector are dominated by debt and servicing, not equity — ABS, agency and Ginnie Mae MBS, whole loans, and mortgage servicing rights — plus private-equity ownership of community banks, captive and mortgage platforms, and late-stage fintech (Stripe the marquee name). Diligence is loan-level: performance by vintage and borrower segment, advance rates and covenants, servicing oversight, and the legal status of any bank-partnership or securitization structure.

Outlook (judgment). Entering 2026 the subsector is broadly constructive but uneven, and the three engines diverge. Depository banks carry margins near multi-year highs and a softer capital re-proposal that could fund buybacks, offset by commercial-real-estate losses concentrated in regionals; credit unions look like more of the same steady growth, with the tax exemption the one Washington wildcard [4]. Nondepository lending keeps taking share from banks and turns on the credit cycle, with GSE privatization the single biggest repricing event on the horizon [5]. Related activities splits into a durable payments grower (watched for interchange and disruption), a steadier servicing book, and a cyclical broker recovery riding lower rates [6]. The shared swing factor is the rate path — but remember it does not push the three children the same way. The common thread across all of 522: the best positions are franchises with a structural funding or network advantage — sticky low-cost deposits, an entrenched payment rail, a scaled servicing book — not merely high-yielding financial stocks. Size expectations to the specific engine, and to the credit cycle, not to a straight line.

For the complete, worked treatment of each child — company rosters, deal history, and full economics — see the three industry-group primers: 5221 Depository Credit Intermediation, 5222 Nondepository Credit Intermediation, and 5223 Activities Related to Credit Intermediation.

---

Sources

(Synthesized from the three child primers for NAICS 5221, 5222, and 5223, plus our ingested federal ground-truth statistics file for NAICS 522.)

1. U.S. Census Bureau, County Business Patterns: 2023, NAICS 522 and children 5221 / 5222 / 5223 (establishments, employment, annual and Q1 payroll). Histometrics ingested federal ground truth for NAICS 522. https://www.census.gov/programs-surveys/cbp.html
2. U.S. Census Bureau, 2022 Economic Census — Concentration of Largest Firms for the United States, NAICS 522 and children (firms, receipts, CR4/CR8/CR20/CR50, HHI). Histometrics ingested ground truth. https://data.census.gov/table/ECNSIZE2022.EC2200SIZECONCEN
3. U.S. Census Bureau, 2022 NAICS Manual — subsector 522 structure and industry-group 5221 / 5222 / 5223 definitions and exclusions. https://www.census.gov/naics/reference_files_tools/2022_NAICS_Manual.pdf
4. Histometrics primer — NAICS 5221, Depository Credit Intermediation (child figures; FDIC / NCUA data; ~$29T depository assets; NIM ~3.3–3.4%; credit-union tax exemption; Basel III endgame; consolidation).
5. Histometrics primer — NAICS 5222, Nondepository Credit Intermediation (child figures; card / captive / GSE detail; wholesale funding; originate-to-distribute; FSOC nonbank-servicer risk; household-debt scale).
6. Histometrics primer — NAICS 5223, Activities Related to Credit Intermediation (child figures; payment networks / processors / utilities; servicing and MSR economics; loan-broker fragmentation; take rates).
7. FDIC, Quarterly Banking Profile — Fourth Quarter 2025 and Q1 2026 (commercial-bank and system assets ≈ $25.3T, NIM, net income). https://www.fdic.gov/quarterly-banking-profile
8. National Credit Union Administration, Quarterly Credit Union Data Summary (credit-union assets ≈ $2.43T, ~145M members, tax exemption). https://ncua.gov/analysis/credit-union-corporate-call-report-data/quarterly-data-summary-reports
9. Federal Housing Finance Agency, Conservatorship of Fannie Mae and Freddie Mac (GSE oversight, guarantee-fee and capital framework). https://www.fhfa.gov/conservatorship
10. Capital One Financial Corp., Form 8-K, Completion of Acquisition of Discover (closed May 2025; ~$35B; issuer + network). https://www.sec.gov/Archives/edgar/data/927628/000119312525122059/d934475d8k.htm
11. Rocket Companies, Rocket Closes $14.2 Billion Acquisition of Mr. Cooper (closed October 2025; largest independent-mortgage deal; originator + servicer). https://www.rocketcompanies.com/press-release/rocket-companies-closes-14-2-billion-acquisition-of-mr-cooper/
12. Federal Reserve Board, 2025 Triennial Payments Study — Initial Findings (236.6 billion U.S. noncash payments worth $140.0 trillion in 2024). https://www.federalreserve.gov/newsevents/pressreleases/other20260701a.htm
13. Consumer Financial Protection Bureau, Regulations and Nonbank Supervision (TILA/Reg Z, CARD Act, RESPA, Ability-to-Repay; SAFE Act licensing; servicing rules). https://www.consumerfinance.gov/rules-policy/regulations/
14. U.S. Financial Stability Oversight Council, Report on Nonbank Mortgage Servicing (2024); U.S. Government Accountability Office, Nonbank Mortgage Companies (GAO-25-107862, 2025). https://home.treasury.gov/system/files/261/FSOC-2024-Nonbank-Mortgage-Servicing-Report.pdf
15. Board of Governors of the Federal Reserve System, Financial Stability Report, May 2026 (consumer delinquencies; bank interconnection to private credit and BDCs). https://www.federalreserve.gov/publications/files/financial-stability-report-20260508.pdf
16. Congressional Research Service, Federal Reserve Cuts Interest Rates in Late 2025 (federal funds rate 3.50%–3.75%). https://www.congress.gov/crs-product/IN12635