Mortgage Pipelines and Warehouse Lines

Table of Contents

    Mortgage pipelines and warehouse lines connect a lender’s promise to make a loan with the cash needed to fund it and the eventual sale to an investor. A lender can have a large book of rate locks before it has advanced that principal. After closing, it owns funded loans that consume financing capacity until the sale settles.

    The practical challenge is to manage price risk, uncertain closings and cash liquidity together. Hedging a loan’s market value does not provide its closing funds, and an expected sale does not make a warehouse borrowing disappear. This guide focuses on U.S. residential mortgages originated for sale, using agency execution rules checked in November 2026 and separately dated company and facility examples.

    What Is a Mortgage Pipeline?

    A mortgage pipeline is the set of prospective loans moving through application, processing, underwriting and closing. Some applications have floating pricing; others have an agreed rate lock. For clarity, this article calls loans that have closed and funded warehouse inventory, separately from the unfunded pipeline. The OCC makes this distinction when discussing originations intended for sale. [1]

    A Commitment Is Different From a Cash Advance

    A $400,000 rate lock refers to the principal of a prospective mortgage. It does not mean the lender has already borrowed $400,000 on its warehouse line. The commitment can nevertheless have economic value and interest rate risk before any mortgage principal is funded.

    Follow one loan through the process. The borrower applies, may lock pricing, qualifies for closing, and receives a funded mortgage. The originator then delivers the loan to a whole-loan buyer or uses an eligible securitization route. The sale’s cash settlement allows the originator to repay its interim financier. A loan intended for the lender’s permanent portfolio follows a different funding strategy.

    The loan changes state before the warehouse financing is repaid
    Stage What the originator has Main question
    Application and underwriting A prospective loan, with floating or locked terms. Will the applicant qualify and proceed?
    Rate lock before funding A conditional commitment to specified pricing for a period. How much will close, and how will market prices move?
    Closing and funding A funded mortgage asset, supported by own cash and any warehouse advance. Are the loan and documents eligible for financing and sale?
    Delivery and sale settlement Sale proceeds, subject to the transaction’s terms and settlement process. Has the financier been paid and its interest in the loan been released?

    Three prices run through this chain: the borrower’s mortgage rate, the investor’s purchase price, and the originator’s funding cost. Fixing one does not necessarily fix the others. The resulting securities and their investor cash flows are covered in the separate guide to mortgage-backed securities.

    Rate Locks, Pull-Through and Fallout

    A rate lock protects agreed borrower pricing for a specified period, subject to its conditions. It does not bypass underwriting. The CFPB explains that expiration or changes to the application, such as loan amount, credit or verified income, can affect locked terms. Extension and any ability to obtain a lower rate depend on the lender’s policy and agreement. [2]

    Pull-through is the proportion of a defined commitment cohort that reaches closing; fallout is the part that does not. A purchase can fail, an applicant can be declined, or a borrower can withdraw. Falling market rates can encourage borrowers to seek better terms, while rising rates make an existing below-market lock more attractive. These are influences on behavior, not guarantees about every borrower. [1]

    For forecasting principal, estimate each unfunded locked loan’s conditional closing probability and weight it by that loan’s amount:

    Expected Funded Principal From Unfunded Locks
    Expected principal = Σi (Locked loan amounti × Pull-through probabilityi)
    A probability-weighted principal forecast, not a warehouse borrowing balance or a complete hedge-sizing model.

    Suppose an originator has $100 million of unfunded rate locks and a principal-weighted pull-through estimate of 75%. Separately, it has $12 million of already funded loans on a hypothetical line advancing 98%. The two populations do not overlap.

    Hypothetical pipeline forecast versus actual warehouse principal
    Measure Calculation Amount
    Unfunded locked principal Prospective loan amounts before closing $100 million
    Expected closings from those locks $100 million × 75% $75 million
    Already funded warehouse loans Separate closed-loan inventory $12 million
    Warehouse funding outstanding $12 million × 98% $11.76 million

    The forecast $75 million has not been borrowed. The funded inventory requires $11.76 million of financing principal and $12 million − $11.76 million = $240,000 of the originator’s cash, before other costs. If the original locked cohort ultimately closes at 60%, it produces $100 million × 60% = $60 million, which is $15 million below the forecast.

    Define the cohort, observation period and weighting consistently. A count-based closing rate can differ from a principal-weighted rate when loan sizes vary. Update estimates for application stage, product, elapsed time and market conditions instead of applying one permanent percentage to every lock.

    How Mortgage Pipeline Hedging Works

    When market mortgage yields rise, the sale value of a previously locked lower-rate loan can fall. An originator can use a forward sale to offset that exposure: selling now for future delivery can gain value when comparable market prices fall. The hedge must cover the loans likely to fund and the funded inventory still awaiting sale, while recognizing any sale commitments already in place. [1]

    The TBA market provides forward agency-MBS exposure used in this process. A TBA hedge does not itself advance the mortgage’s closing cash, and an individual mortgage is not automatically deliverable as a TBA security. Loan eligibility, pooling, coupon and settlement requirements still matter.

    Uncertain volume complicates the hedge. If rates fall and more borrowers withdraw, a lender can be left with a larger forward sale than its actual loan production supports. Buying back or pairing off the excess can require cash. If rates rise and more locks close than forecast, some loans may be insufficiently hedged. In the earlier example, $75 million of forecast closings becomes $60 million of actual production; the $15 million difference illustrates volume uncertainty, rather than prescribing a $75 million TBA position.

    Actual sizing also depends on price sensitivity, product mix, coupon, loan characteristics and timing. Loans and hedges can move differently, creating basis risk. Options can address some nonlinear exposure at a premium; swaptions are one separate interest rate option tool, rather than a universal solution for every mortgage pipeline. The OCC’s hedging discussion emphasizes changing fallout estimates and the remaining risks of different instruments. [1]

    A Price Hedge Does Not Guarantee Cash Liquidity

    A hedge can help offset an asset’s economic loss while still requiring collateral or settlement cash at an inconvenient time. Conversely, an apparent hedge gain may not be available in the account or at the moment needed to meet a warehouse demand. Evaluate the cash schedule alongside the net market exposure. Rocket identifies margin-call exposure on certain funding facilities and origination hedges in its 2025 filing. [3]

    How a Mortgage Warehouse Line Works

    A mortgage warehouse line supplies short-term funding to an originator for eligible mortgage loans awaiting sale. It finances the originator’s inventory; it is different from the homeowner’s long-term mortgage. The originator generally contributes its own cash and repays the warehouse financing from the takeout sale.

    The legal structure matters. Funding can use a secured lending arrangement or a master repurchase agreement, under which loans are transferred and later repurchased under the financing terms. Calling both “warehouse funding” describes their business role without making their ownership, collateral or insolvency treatment identical.

    Rocket provides a dated operating example. Its 2025 Form 10-K describes primarily master-repurchase funding, generally financing about 97%–98% of loan principal, with some loan types financed at lower percentages. It says loans generally remained in funding facilities for less than 45 days before pooling and sale, and that sale proceeds repaid the facilities. These are Rocket’s reported practices, rather than universal advance rates or promised sale times. [3]

    In a simplified facility with no other constraints:

    Advance and Originator Cash
    Advance = Eligible funding value × Advance rate
    Funding gap = Loan funding amount − Advance
    The agreement defines eligible value, applicable rates, exclusions and any additional requirements.

    A 98% advance leaves a 2% haircut against that specified value. The haircut helps determine the originator’s financing gap; it is not the homebuyer’s down payment or the mortgage’s property loan-to-value ratio. The haircut amount equals the financing gap when the eligible funding value equals the loan amount being funded and no other adjustments apply. Some contracts use principal, cost, market value or a constrained combination to determine the base. Revaluation, ineligibility or other limits can reduce the amount actually available.

    Distinguish the advertised facility limit from the available amount and from funding already outstanding. Eligible collateral, product sublimits, required liquidity or net worth, and committed versus discretionary funding provisions all affect access. A nominal unused limit is not unconditional cash.

    Closing logistics also matter. Pulte Mortgage’s August 11, 2026 Truist agreement expressly permits qualifying wet-funding transactions, where closing funds are sent before the financier or custodian receives the required documents. The agreement imposes conditions, sublimits and a permitted period for that exposure. Thus “warehouse” does not mean every funding dollar waits until a complete final file is in custody. [4]

    Warehouse Funding and Loan-Sale Example

    Consider a hypothetical $400,000 mortgage funded at par. Assume its eligible funding value is $400,000, the line advances 98%, and the originator supplies the rest. The warehouse rate is a fixed illustrative 6% annually on an actual/360 basis. Assume no fees, interim principal changes or earlier interest payments.

    Funding at Closing

    Warehouse advance = $400,000 × 98% = $392,000.

    Originator contribution = $400,000 − $392,000 = $8,000.

    Now assume the takeout buyer accepts the eligible loan and actually settles on day 30 at a clean price of 101% of the unchanged principal. The clean price excludes accrued borrower interest. Under this example’s settlement convention, all warehouse interest is paid with the principal at sale.

    Hypothetical day-30 sale and warehouse repayment
    Cash item Calculation Amount
    Sale proceeds $400,000 × 101% $404,000
    Warehouse interest $392,000 × 6% × 30/360 $1,960
    Principal and interest paid to financier $392,000 + $1,960 $393,960
    Cash left for originator $404,000 − $393,960 $10,040

    The $10,040 includes the return of the original $8,000 contribution. Subtracting it leaves $10,040 − $8,000 = $2,040, equal to the $4,000 sale premium minus $1,960 warehouse interest. This narrow amount excludes borrower interest, servicing value, hedging results, origination and operating costs, fees and taxes. It is neither total profit nor an accounting gain-on-sale calculation.

    The investor’s purchase repays the originator’s interim funding; it does not pay off the homeowner’s mortgage. The borrower continues to owe the mortgage under its terms. In practice, financing documents determine payment dates and payoff mechanics, so warehouse interest may have been paid before the sale.

    Dwell Time, Aging and Margin Calls

    Dwell time is how long a funded loan remains in warehouse inventory before takeout. Delayed documents, underwriting questions, investor rejection or a market disruption can lengthen it. Longer dwell occupies capacity, ties up the originator’s contribution and can add funding, hedging or extension costs. Aging reports help identify why loans remain unsold. [1]

    To isolate funding cost, keep the earlier $392,000 advance and 6% actual/360 rate constant and vary only days outstanding. Ignore fees, marks, interim payments and principal changes. This calculation assumes neither a right to retain an actual advance for these periods nor an unchanged takeout price.

    Hypothetical financing-only cost of longer dwell
    Days outstanding Calculation Warehouse interest
    30 days $392,000 × 6% × 30/360 $1,960
    45 days $392,000 × 6% × 45/360 $2,940
    60 days $392,000 × 6% × 60/360 $3,920

    An extra 30 days costs $3,920 − $1,960 = $1,960 in this simplified case. Actual agreements can instead require repayment, apply a different rate or stop treating an aged loan as eligible. Borrower interest may offset some carry costs, but it does not remove the funding and sale risks.

    A collateral mark can require cash before sale

    Reset to the loan’s initial $400,000 eligible value and $392,000 outstanding advance. In a separate hypothetical stress scenario, the eligible value falls to $380,000. Assume the same 98% advance rate and a contract requiring immediate cash paydown without a minimum call threshold; ignore accrued interest.

    The Haircut Stays the Same, but Available Funding Falls

    New permitted advance = $380,000 × 98% = $372,400.

    Cash principal paydown = $392,000 − $372,400 = $19,600.

    The $19,600 is a cash requirement. It differs from the $20,000 decline in the loan’s valuation, and any hedge offset needs separate analysis. This scenario does not assume the successful $404,000 sale in the previous example.

    Pulte’s filed agreement shows why the actual terms matter. Its margin provisions use a contractually defined market-value measure, include a minimum aggregate threshold in the absence of specified defaults, permit acceptable additional conforming loans as an alternative to cash, and set payment deadlines. The simplified no-threshold example above does not describe that facility’s actual call. [4]

    Mandatory vs Best-Efforts Takeout Commitments

    A takeout is the intended investor purchase that converts the funded loan into sale proceeds. Its commitment type helps determine which party bears the risk that production will not arrive. Fannie Mae’s November 4, 2026 execution topic distinguishes the following whole-loan choices. [5]

    Fannie Mae whole-loan execution choices
    Feature Mandatory Best efforts
    What is committed A specified dollar amount of qualifying loans, within stated tolerances, by a date. A specific loan by a date if it closes.
    If expected loans do not close The delivery obligation remains; a shortfall can require pairing off under the agreement. Genuine qualifying fallout generally avoids a pair-off fee.
    After the specific loan closes Deliver according to the commitment’s requirements. Delivery is required; it is not a free choice to sell the closed committed loan elsewhere.
    Conditions that remain Eligibility, pricing, documents, delivery, timing and other contract terms. Eligibility, accurate committed data, delivery, timing and other contract terms.

    Fannie Mae’s detailed best-efforts topic distinguishes applicant withdrawal or lender declination from failure to deliver a closed committed loan. It also provides for duplicate-commitment price adjustments and extension fees. Thus “best efforts” should not be read as cost-free cancellation in every circumstance. Other investors have their own contracts. [6]

    A commitment is also different from cash already received. Fannie Mae requires good delivery, including loans meeting underwriting and legal criteria, the original commitment, and acceptable documents and data. A loan can have a planned buyer and still fail to settle on schedule. [7] For loans outside the agency channel, examine the actual private investor’s requirements rather than importing agency execution terms; the resulting securities are discussed separately in non-agency MBS.

    How Loan Sales Repay the Warehouse Line

    At settlement, the sale and financing-release instructions must work together. The takeout investor pays as instructed, the warehouse financier receives the required payoff, and its interest in the loan is released through the applicable custody and transfer arrangements. Only the residual cash is freely available to the originator, subject to any other obligations.

    Fannie Mae’s bailee-letter rules illustrate this control: the financier’s interest is released upon the required proceeds reaching it, while documents are held for it pending payment or return. The seller must make its instructions match those of the warehouse lender; mismatches can cause trading or settlement delays. [8]

    A securitization route adds another distinction. Receiving an MBS in exchange for loans is not the same as receiving cash from selling that MBS. Financing remains to be addressed until the required payoff occurs. Revolving capacity can then support new originations, provided the facility continues to permit advances.

    Repayment of the warehouse financing does not necessarily end every sale obligation. Seller representations and warranties can create later exposure, explained in mortgage putbacks and repurchase risk. That post-sale remedy is different from both a pre-sale margin call and the routine “repurchase” that unwinds a repo financing.

    How to Evaluate a Mortgage Originator’s Funding Risk

    Analyze the operating cash cycle alongside the earnings forecast. This is a specialized application of working capital management: more originations can require more cash before the related sale proceeds arrive.

    1. Separate the populations. Reconcile applications, unfunded rate locks, expected closings, funded warehouse inventory and completed sales. Avoid counting a funded loan twice.
    2. Test the closing forecast. Examine principal-weighted pull-through by stage and product, and what happens when rates or borrower behavior change.
    3. Read the facility. Compare total limits, committed capacity, eligible funding value, actual advances, haircuts, covenants, sublimits and maturity or termination provisions.
    4. Check the takeout. Identify the buyer, commitment type, price adjustments, eligibility, document status, delivery deadline and settlement instructions.
    5. Stress cash timing. Combine slower sales, collateral marks, possible hedge margin, interest and reduced funding availability. Determine which cash or acceptable collateral is actually accessible when needed.
    6. Monitor aged exceptions. Investigate unsold loans individually rather than assuming every delay will clear at the expected price.

    Common Mistakes

    • Calling pipeline notional funded debt. A rate-lock amount, expected closing volume, commitment fair value and warehouse advance measure different things.
    • Using pull-through as the entire hedge model. Principal forecasts also need price sensitivity, product and timing analysis.
    • Treating the facility limit as guaranteed availability. Eligibility, contractual conditions and the amount actually committed can constrain access.
    • Confusing the warehouse haircut with homeowner equity. The haircut concerns originator financing against the contract’s funding base.
    • Assuming a planned buyer guarantees settlement. Good delivery, acceptable collateral and correctly matched instructions still matter.
    • Counting all residual sale cash as profit. It includes recovery of the originator’s cash contribution, and other revenues and costs remain.
    • Assuming sale eliminates all recourse. Distinguish paid-off financing from continuing contractual seller obligations.

    Limitations of Pipeline and Warehouse Analysis

    Pipeline forecasts depend on changing borrower behavior and timely loan-level data. A single average can hide different products, stages, sale routes and problem loans. A period-end balance can also conceal large funding needs earlier in the month or quarter.

    Facility and takeout agreements vary by legal entity and transaction. An originator can face both asset-value losses and a cash shortage, even when its long-run loan economics appear acceptable. The examples here hold many inputs constant to isolate the mechanics; they do not estimate an actual lender’s profitability, capital requirement or stress survival.

    Frequently Asked Questions

    No. The lock sets conditional pricing for a period before closing. The prospective loan can still fail to fund. Its reference principal and economic value are different from the cash advanced when the mortgage is funded.

    The mortgage originator or designated financing entity obtains interim funding for eligible loans. The homeowner owes a separate mortgage. Some warehouse arrangements use a legal sale-and-repurchase structure rather than a conventional secured borrowing.

    A TBA hedge manages agency-MBS price exposure. It does not itself provide the cash needed to fund the homeowner’s loan. The originator still needs its own resources or financing, and the hedge can have its own collateral and settlement obligations.

    The originator must address the loan and its financing under the applicable agreements. It may need to cure delivery problems, find an eligible alternative buyer, provide more cash or repay the advance. An investor’s refusal does not automatically cancel the originator’s funding obligation.

    No. In Fannie Mae’s described execution, genuine qualifying fallout generally avoids a pair-off fee, but the specific loan must be delivered if it closes. Eligibility, accurate data, deadlines and other commitment requirements remain. Other buyers’ terms must be read separately.

    Sources

    1. OCC: Comptroller’s Handbook, Mortgage Banking, February 2014, current hosted copy with March 2025 cover update. Printed pages 42–45 and appendix B: pipeline, warehouse, fallout and hedge mechanics.
    2. CFPB: What’s a lock-in or a rate lock on a mortgage?, reviewed May 2, 2023. Conditional pricing, application changes and extension questions.
    3. Rocket Companies: 2025 Form 10-K, filed March 2, 2026. Liquidity and Capital Resources, pages 74–75; funding, collateral marks, sale proceeds and margin exposure.
    4. Pulte Mortgage / Truist: Master Repurchase Agreement, dated August 11, 2026, SEC Exhibit 10.1. Sections 4, 6 and 7: closing funding, repayment, pricing and margin conditions.
    5. Fannie Mae: C1-1-01, Execution Options, November 4, 2026. Whole-loan mandatory/best-efforts commitments and MBS execution.
    6. Fannie Mae: C2-1.2-02, Best Efforts Commitment Pricing, Periods, and Fees, March 3, 2021 topic. Fallout, closed-loan delivery and fees.
    7. Fannie Mae: C2-2-01, General Requirements for Good Delivery of Whole Loans, August 7, 2018 topic. Purchase conditions and loan/document data.
    8. Fannie Mae: C1-2-05, Bailee Letters, April 6, 2022 topic. Financier interests, proceeds and matching delivery instructions.

    Sources accessed November 14, 2026. Facility and company examples describe their identified dates; the numerical scenarios are hypothetical and do not quote an available loan or financing offer.

    Disclaimer

    This article is for educational and informational purposes only and does not constitute investment, lending, accounting or legal advice. Mortgage commitments, warehouse financing, collateral calls and loan sales depend on the applicable contracts, facts and law. Consult qualified professionals about a specific transaction or financing decision.