The private credit boom of 2018–2025 was not built solely by Apollo, Ares, Blackstone, and Blue Owl. It was built in partnership with several hundred regional and mid-size banks that originated the loans, provided the warehouse financing, and fed the pipeline — in exchange for fee income, relationship deposits, and the ability to grow loan volumes without holding the credit risk. The arrangement was mutually beneficial while it worked. The PE platforms got origination scale. The banks got revenue without capital consumption. And the market got $1.7 trillion in private credit AUM in less than a decade.

The arrangement is structurally dependent on a permanent, growing PE buyer. It has never been stress-tested. And it is now approaching a stress test.

This memorandum documents the full O-to-D model architecture, quantifies the fee income dependency across the banking sector, traces the four mechanisms by which PE platform pullback reaches bank balance sheets, identifies the warehouse line problem as the most immediate balance sheet risk, examines the real-economy transmission through small business credit tightening, draws the 2007 mortgage O-to-D collapse as the operative historical analogue, and frames specific positioning across bank equity, credit, and the broader middle-market lending market.

Originate-to-distribute (O-to-D) is not a new concept in banking — it is the same structural model that drove the mortgage securitization boom of the 2000s. What is new is its application to middle-market and direct lending, and the specific form it takes in partnership with PE credit platforms.

The model is attractive because the first three revenue streams require almost no capital — origination, servicing, and relationship banking fees are essentially pure fee income with minimal balance sheet consumption. The bank grows its loan volume and fee base without growing its risk-weighted assets. Return on equity improves. Regulatory capital ratios improve. The model looks, from the outside, like a productivity improvement — a bank becoming more efficient. It is, structurally, a model that works only as long as the PE buyer on the other side of every loan sale never stops buying.

The aggregate figures above are distributed unevenly. The risk is concentrated in three categories of banks:

Note: Exposure classifications are based on publicly available call report data, SEC filings, and investor presentations. Risk levels reflect structural O-to-D dependency, not overall institutional health.

The financial profile of a bank with material O-to-D operations looks superficially excellent in a growth market: loan volumes are high, fee income is growing, net interest margin is supplemented by non-interest income, and return on equity is above peer. The same profile, stress-tested, reveals a revenue structure that is fragile at the precise points where it appears strongest.

Sector-wide estimates. Individual bank impact varies significantly based on concentration in O-to-D business.

Originate-to-distribute banks classify loans as "held for sale" (HFS) on their balance sheet from the moment of origination until the PE platform purchases them. Under US GAAP (ASC 948), held-for-sale loans must be carried at the lower of cost or fair value — meaning that if the fair value of the loan falls below its origination cost (because private credit spreads widen, because the borrower's credit quality deteriorates, or because the market for that type of credit dries up), the bank must immediately recognize the loss in earnings through a valuation allowance. It cannot amortize the loss over time.

When the PE platform stops buying, the bank faces a choice: continue holding loans in the HFS category (carrying them at lower of cost or fair value, recognizing mark-to-market losses every quarter until a buyer is found) or reclassify them to held-for-investment (HFI). Reclassification to HFI brings its own accounting consequences:

• At the point of reclassification: The loan must be transferred at its current fair value, not its amortized cost. If the loan was originated at par and the market has moved 200 bps wider, the bank crystallizes the mark-to-market loss at the moment of reclassification — it is recognized in earnings immediately, not deferred.
• After reclassification: The loan becomes subject to CECL (Current Expected Credit Loss) provisioning — the bank must estimate and reserve for lifetime expected credit losses on the now-held loan, which may be a multiple of the prior period's credit loss provision because the loan was originally underwritten for distribution, not for portfolio holding.
• Capital impact: Risk-weighted assets increase by the full face value of the reclassified loans, consuming regulatory capital at exactly the moment when earnings are being impaired by the mark-to-market losses and increased provisions.

Warehouse lines are the least-discussed and most immediately dangerous component of bank exposure to private credit stress. Unlike O-to-D loan sales — where the credit exposure transfers to the PE platform upon sale — warehouse lines represent a direct, on-balance-sheet credit exposure to the PE credit platform itself. The bank is effectively lending to the PE platform at short duration, secured by a pledged pool of loans that the PE platform has not yet been able to sell or CLO-format.

The normal warehouse line lifecycle is: PE platform draws on the line → uses funds to buy/originate loans → packages loans into a CLO or sells to institutional investors → repays the warehouse line with CLO proceeds. The cycle typically takes 3–12 months. In a functioning CLO market with strong institutional demand, warehouse lines are self-liquidating.

A warehouse line becomes stranded when the CLO market freezes — when institutional investor demand for CLO tranches dries up, when rating agency CLO criteria tighten, or when the underlying collateral quality deteriorates to the point where CLO formation is uneconomical. In that scenario:

• The PE platform cannot form a CLO and cannot repay the warehouse line on its normal schedule.
• The bank extends the facility maturity — either contractually (if the agreement allows) or through a forbearance/modification.
• The bank is now holding a term loan to a PE credit platform secured by a pool of middle-market loans — an asset it never intended to hold, with risk characteristics it did not underwrite for portfolio holding.
• If the PE platform itself faces financial stress (from management fee revenue decline, key-man events, or its own borrowing cost increases), the warehouse line becomes a credit exposure to a potentially impaired counterparty.

Warehouse lines are disclosed in bank call reports as unfunded commitments or funded loans depending on their drawn status. Most banks do not separately identify warehouse lines to PE credit platforms in public disclosures — they are aggregated into "C&I loans" or "other commercial loans." The only way to identify warehouse line exposure is through:

• Management discussion and analysis sections of 10-K/10-Q filings (where some banks describe warehouse activity)
• Investor day presentations where bank management discusses specialty lending segments
• Call report Schedule RC-C (loan detail by purpose category) — which provides category-level data but not counterparty identification
• FDIC summary statistics cross-referenced against bank balance sheet growth trends in specific C&I sub-categories

The opacity means that aggregate warehouse line exposure to PE credit platforms is not measurable from public data alone. It can only be estimated — and the estimates, ranging from $100B to $300B in total outstanding commitments across the banking system, carry wide confidence intervals.

Most O-to-D loan sale agreements contain representations and warranties (R&W) provisions requiring the originating bank to repurchase loans from the PE platform buyer if certain conditions are subsequently found to have been breached at origination. These are standard provisions in loan sale documentation — the PE platform needs assurance that the loans it is buying meet the underwriting standards represented at origination.

Repurchase demands are not triggered by credit performance deterioration alone — that risk transferred with the loan. They are triggered by documentation failures or misrepresentation at origination:

• Underwriting standard breaches: If a loan was represented as meeting specific debt service coverage, loan-to-value, or leverage ratio thresholds and subsequent review reveals the original underwriting was incorrect or the data was manipulated, the PE platform can demand repurchase.
• Documentation deficiencies: Missing collateral filings, unrecorded liens, incorrect borrower identity documentation, or defective security interests discovered post-sale.
• Fraud or misrepresentation: Any fraudulent misrepresentation in the loan file — either by the borrower or by the originating bank's own staff — that was not disclosed at sale.
• Early payment default (EPD) clauses: Many loan sale agreements contain provisions requiring the seller to repurchase loans that default within 90–180 days of origination — regardless of cause — because early defaults are presumed to indicate origination failure.

The most relevant historical parallel is the post-2008 mortgage repurchase cycle. Bank of America, JPMorgan, Wells Fargo, and Citigroup collectively paid over $60 billion in mortgage repurchase settlements between 2010 and 2015, primarily to Fannie Mae, Freddie Mac, and the Federal Housing Finance Agency. These were not losses from loans the banks held — they were losses from loans the banks had sold, which were returned through the repurchase mechanism when the GSEs determined origination standards had been breached. Private credit repurchase losses are unlikely to reach mortgage-cycle scale — but the mechanism is identical, and the documentation standards in middle-market direct lending have been notably lax during the 2020–2025 boom period, when deal volume and competition pressured underwriting rigor.

The preceding sections document bank-level financial stress. This section documents the transmission of that stress into the real economy — specifically, into the small business and middle-market borrower base that depends on regional and community bank lending for its capital access. This transmission channel is the least discussed in private credit research because it does not show up in fund NAVs, CLO data, or FABN spreads. It shows up in the NFIB Small Business Optimism Survey, the Fed's Senior Loan Officer Opinion Survey (SLOOS), and in the employment data of the 15 million US small businesses that have no alternative to their local bank.

Small businesses (under 500 employees) account for approximately 44% of US private sector payroll and 43% of GDP. They are almost entirely dependent on regional and community banks for credit — the largest PE credit platforms do not make $500K or $5M loans. When regional bank O-to-D capacity contracts, the small business credit market has no institutional alternative. The PE credit platforms moving down-market (unitranche loans as small as $25–50M) are not reaching the sub-$10M borrower. The credit tightening in this segment is effectively permanent until the O-to-D model re-establishes itself — a multi-year process that requires PE platforms to resume buying, CLO markets to re-open, and banks to rebuild confidence in their distribution pipelines.

Regional banks do not enter a private credit stress scenario with clean balance sheets. They enter it already under regulatory scrutiny for elevated commercial real estate (CRE) concentrations — a separate, pre-existing stress that regulators have been flagging since 2023. The interaction between CRE stress and O-to-D disruption is not additive. It is multiplicative.

As of year-end 2025, the FDIC and OCC have identified over 200 US banks with CRE loan concentrations exceeding the interagency guidance thresholds: 300% of total risk-based capital for total CRE, or 100% of total risk-based capital for construction and land development loans. These banks are already under enhanced supervisory scrutiny, required to maintain higher capital buffers, and restricted in dividend and buyback activity. Many of these same banks are among those most active in O-to-D partnership origination — because CRE-concentrated banks need fee-based income diversification to improve their non-interest income profile.

A subset of regional banks most exposed to O-to-D disruption also have elevated office CRE concentrations — loans to office property owners and developers whose collateral values have declined 30–50% in major markets since 2022. These banks are simultaneously: managing CRE workout situations that consume management bandwidth and capital; facing O-to-D fee income declines that reduce their earnings cushion; absorbing mark-to-market losses on HFS loan reclassifications; and receiving increased attention from bank examiners. The probability of a regional bank failure or forced sale is highest in this intersection — the banks that are both CRE-stressed and O-to-D dependent are the most vulnerable to a confidence-driven deposit run if any of the above stresses becomes public.

The 2007–2008 collapse of the mortgage originate-to-distribute model is the only precise historical parallel to what is building in the private credit O-to-D system. The structural similarities are detailed enough to be instructive without being deterministic — the scale, speed, and regulatory context differ, but the core mechanism is identical.

Several factors differentiate the current O-to-D risk from the 2007 collapse — some mitigating, some aggravating:

• Mitigating — No consumer exposure: Middle-market and direct lending O-to-D involves commercial borrowers, not homeowners. Commercial defaults are faster to resolve (collateral can be liquidated more cleanly), and there is no political/social dimension to commercial foreclosure that delays resolution as it did with residential mortgages.
• Mitigating — Smaller absolute scale: The private credit O-to-D market, while large, is smaller than the 2007 RMBS/CDO complex in nominal terms — approximately $1.7T in private credit AUM vs. $11T in outstanding mortgage debt at peak.
• Aggravating — Less regulatory infrastructure: Mortgage O-to-D was overseen by the OCC, OTS, and state banking regulators — with well-established examination frameworks for loan quality. Private credit O-to-D operates in a less-examined regulatory space, with less standardized underwriting documentation and fewer supervisory checkpoints.
• Aggravating — Simultaneous stresses: Banks in 2007 were not simultaneously managing CRE concentration concerns, PE-owned insurer FABN stress, and retail non-traded BDC exposure. Regional banks in 2026 face all of these concurrently — the O-to-D risk is one component of a multi-vector stress, not an isolated event.

Bank regulators — the OCC, FDIC, Federal Reserve, and state banking departments — examine banks using frameworks designed for traditional banking risks: credit risk, interest rate risk, liquidity risk, and operational risk. The O-to-D model introduces a category of risk that falls between the existing examination categories: pipeline dependency risk — the risk that a bank's business model is structurally dependent on a non-regulated counterparty (the PE credit platform) continuing to perform a function (buying loans) that it has no obligation to continue.

The Basel III Endgame rules finalized by US regulators in 2024 (with phased implementation through 2028) increase risk-based capital requirements for large banks (Category I–III). For regional banks in Categories III and IV ($100B–$700B in assets), the rules require enhanced operational risk capital and certain market risk adjustments. However, the rules do not address O-to-D pipeline risk, warehouse line risk, or PE platform counterparty concentration at Category IV banks ($100B–$250B). The community and mid-size regional banks ($5B–$100B) where O-to-D dependency is most concentrated are largely exempt from the enhanced Basel III framework — they operate under the simpler, pre-Basel III capital rules that provide no explicit buffer against pipeline dependency risk.

The O-to-D risk does not inevitably produce a banking crisis. The following conditions, sustained simultaneously, would allow the system to navigate the private credit cycle without triggering the cascade described above:

• Private credit defaults remain below 2.5% through 2027 — keeping PE platform NAVs intact, CLO collateral quality within model parameters, and PE buying appetite sustained. This is the most load-bearing condition. If defaults remain low, the entire cascade fails to initiate.
• CLO formation remains above $10B/month — allowing warehouse lines to continue cycling normally and preventing the warehouse extension problem from materializing. CLO demand is itself dependent on the default condition above.
• Regional banks proactively reduce O-to-D concentration before the cycle turns — rebuilding hold-to-maturity loan portfolios, diversifying fee income, and reducing warehouse line commitments. There is no evidence this is happening at scale; the O-to-D model remains more profitable than the alternative in a benign market.
• Regulators implement O-to-D examination guidance — identifying at-risk banks before the crisis and requiring them to hold capital buffers against pipeline disruption. The OCC has signaled awareness of PE-bank partnership risks but has not issued formal guidance as of early 2026.
• PE platforms develop alternative distribution channels — reducing their dependence on CLO formation by expanding insurance company, pension fund, and sovereign wealth fund direct investment into private credit. This is occurring at the margin (Athene, Global Atlantic, and F&G are already doing this through FABN issuance) but adds a new vulnerability (the FABN maturity wall) rather than eliminating the existing one.