Most advisory work in specialty finance is sale work. This is not. The practice advises owners on the machine rather than the exit: the warehouse facility, the forward flow agreement, the borrowing base, the advance rate, the eligibility schedule, the covenant and cure package, the capital structure behind a card or lending program, and the credit evidence supporting all of it.
Engagement profile
| Focus | Funding, balance sheet and program economics |
|---|---|
| Enterprise value | $5M to $300M |
| Assets | Consumer, commercial, equipment, receivables, card |
| Geography | United States and Canada |
| Engagement | Advisory, not a sale mandate |
| Timeline | Six weeks to nine months |
| Advisor | Senior advisor led |
What this practice covers
The companies are consumer and commercial lenders, equipment finance and leasing companies, factors, merchant cash advance providers, asset-backed originators, card issuers and program managers.
The problem is usually the same. A funding structure assembled one facility at a time now caps growth in a way origination volume cannot lift.
Advance rate is the growth constraint
Advance rate decides how much of the owner’s own capital each dollar of originations consumes. Same product, same borrower, a materially different advance rate, and the equity required per dollar originated can double while the growth rate halves. Owners negotiate hardest on the coupon and take the advance rate and eligibility schedule as drafted.
Windsor Drake does not publish advance rate benchmarks. Ranges circulated in vendor material are not built on a disclosed sample or methodology, and a number that cannot be sourced is not worth quoting to a lender. What moves an advance rate is the company’s own documented loss performance, not a market average.
Who this is for
Lenders, specialty finance companies and payments businesses in the United States and Canada with enterprise values of $5 million to $300 million. Owners funding on a warehouse, a forward flow, a securitization, a bank line or their own balance sheet. Companies at a renewal, an upsize, a lender replacement or a first institutional facility. Card issuers and program managers whose economics depend on sponsor bank terms, interchange and funding cost. Owners deciding whether to hold what they originate or sell it forward.
What a warehouse facility actually tests
A warehouse is not a loan against a company. It is a loan against a pool, redetermined continuously: eligible receivables, less ineligibles, less concentration excesses, multiplied by the advance rate, recalculated daily or weekly rather than at quarter end.
Eligibility criteria remove assets from the base silently
Minimum score thresholds, first payment received, a complete document package, a perfected lien, no delinquency beyond a stated ageing, approved jurisdictions. Failing assets are not rejected. They are funded with equity.
Concentration limits do the same at pool level
Single obligor caps, state and geographic limits, sector caps. A book can be fully performing and still starved of availability.
Pricing is the smallest lever on the page
Drawn margin, unused fee and upfront fee together matter far less than the base rate environment and the advance rate. With the effective federal funds rate at 3.63% on 11 August 2026, the base is doing more work than the spread.
Triggers and cures decide who runs the company under stress
Delinquency and cumulative loss triggers, minimum tangible net worth, minimum liquidity and excess spread tests are standard, and the consequences escalate: enhanced reporting, then modified funding mechanics, then a change in cash flow priority that turns a revolving facility into an amortizing one. Spend the negotiating capital on cure rights: notice period, cure by equity contribution, cure by repurchase of the offending assets, and the number of cures permitted in twelve months. Headroom that exists only at current performance is not headroom.
Forward flow is a sale, and it prices like one
A warehouse is committed capacity to fund assets you keep. A forward flow is contracted demand to buy them. A buyer agrees to purchase assets meeting an agreed schedule at an agreed price over a period, removing the funding requirement for the flowed portion and capping its yield at the same time. The economics turn on purchase price as a premium or discount to par, allocation targets and minimum allocation percentages, servicing consistent with the assets the originator retains, and soft stop and hard stop origination events that pause or permanently end purchases.
Volume in this market is institutional. Upstart and Fortress Investment Group announced a $1.25 billion forward flow agreement on 29 April 2026, following a $1.2 billion agreement between the same parties in May 2025.
Repurchase obligations are the term owners underprice
Loan-level representations, first payment default provisions and eligibility breach mechanics create a contingent liability that sits off the balance sheet until it does not. Size it against the company’s own historical breach and early default rates, reserve for it, and negotiate a stated cap, a defined cure window, a substitution right where the asset can be replaced rather than bought back, and a sunset after which representations expire.
Program economics
A lending program is priced per funded dollar per year: yield, less cost of funds, less expected loss, less servicing, less acquisition cost. Cut by vintage, channel and credit band it shows which cohort subsidises which, and it is the only version a warehouse lender or flow buyer accepts.
Card and payments programs turn on a different variable. Interchange is set by network, product and issuer status, and issuer status is structural rather than commercial. Federal Reserve data published 19 December 2025 show that in 2024, covered debit issuers averaged $0.23 of interchange per transaction, or 0.47 percent of value, across all networks, against $0.51 and 1.21 percent for exempt issuers. The Board’s companion biennial report, covering 2023, put the average debit transaction at $46.26, covered issuer authorization, clearing and settlement costs at $0.041 per transaction, and covered issuer fraud losses at 17.6 basis points of transaction value.
That gap is why sponsor bank selection is an economic decision before it is a compliance one, and why single-sponsor dependency prices as concentration risk. See bank sponsorship advisory. The rule is also unsettled: the United States District Court for the District of North Dakota vacated Regulation II in its entirety on 6 August 2025 and stayed its own order pending appeal, the Eighth Circuit heard argument on 13 May 2026, and as at 13 August 2026 no decision has issued. Program models should be tested against more than one interchange outcome.
Balance sheet strategy
A balance sheet lender consumes equity with every origination and earns the full spread. An originate-and-sell platform consumes almost none and earns a fee. Most companies of scale run both, and the work is deciding the split deliberately: which vintages to retain, which to flow, what retained interest to hold, and how much unencumbered liquidity to keep against a trigger event.
Static pools are the currency
No institutional counterparty accepts a blended portfolio loss rate. They rebuild the book from the loan tape, and their answer becomes the number in the term sheet.
Item 1105 of Regulation AB, 17 CFR 229.1105, is the benchmark to build against even where nothing is being registered. It requires static pool disclosure of delinquencies, cumulative losses and prepayments, presented by prior securitized pool of the sponsor for that asset type, or by vintage origination year where the sponsor has less than three years of experience securitizing that asset type, covering five years or the sponsor’s full history with the asset type if shorter, in each case to the extent material. Revolving asset master trusts are subject to a separate regime.
Cut losses by cohort and track them to maturity, and document every change to charge-off and re-ageing policy, because an undisclosed change discredits the dataset. An owner arriving with eight years of static pools argues about the discount rate. An owner arriving with a portfolio average argues about the data.
What has moved in 2026
Bank appetite for nonbank lending is strong in volume and tight in terms. Lending to nondepository financial institutions accounted for roughly 40 percent of all US bank loan growth in 2025 on Federal Reserve data analysed by American Banker, and the Federal Reserve’s July 2026 Senior Loan Officer Opinion Survey, released 3 August 2026, reported significant net shares of banks describing standards for every category of nondepository borrower as at the tighter ends of their historical ranges since 2011. Capacity is available. Terms are not.
Private credit has expanded into the gap. The Alternative Credit Council and AIMA put the global private credit market at US$3.5 trillion at the end of 2025 across corporate, asset backed, real estate and infrastructure lending. Securitization is open: KBRA forecast approximately $385.2 billion of US ABS issuance in 2026, another post-crisis record following a record 2025, in a forecast published 20 November 2025.
Consumer credit is stable but split. New York Federal Reserve data for the second quarter of 2026, released 11 August 2026, show the annualized flow into serious delinquency at 6.97 percent of balances for credit cards and 3.00 percent for auto loans. TransUnion research published 30 April 2026 describes a K shaped market in which super prime grew from 36.9 percent of consumers at the end of 2019 to 40.7 percent at the end of 2025. Lenders whose static pools separate those populations are financed on their own numbers. Lenders whose data cannot are priced as though the whole book is the weaker group.
Illustrative example, not a specific transaction
A consumer installment lender funded on a single warehouse finds that a meaningful share of its pool sits outside the borrowing base, driven by a first payment seasoning requirement and a state concentration cap set three years earlier. Headline cost of funds looks competitive. Actual equity consumption per dollar originated is far higher than the advance rate alone implies, and growth has stalled.
The work is evidence, not a financing. Several years of originations are rebuilt into static pools by cohort and score band, and the loan tape is reconciled to the servicing system and the ledger. On that record the eligibility schedule is renegotiated at renewal, the state cap widened, a second lender added, and a forward flow signed for the thinnest margin cohort with a capped repurchase obligation. The coupon barely moves. The equity released is the outcome.
This example is provided for illustration. Specific details, parties and outcomes have been omitted or generalized. It does not represent a specific Windsor Drake engagement.
Who does this work
This practice is led by Bruce Goldstein, Senior Advisor. He is a registered Investment Banking Representative of Independent Investment Bankers, Corp., a FINRA member firm, CRD 2288224, registered since 1992, with no disclosures on the public record. His registration history includes Sandler O’Neill & Partners from 1992 to 1995, Keefe, Bruyette & Woods from 1997 to 2001, and Milestone Advisors from 2002 to 2005. He was also a founding partner of Middlemarch Partners. On the operating side he was a founding partner in a MasterCard issuing business and the founder of an online consumer lending company, so he has negotiated bank sponsorship and raised the funding behind a loan book rather than only read the agreements. His record is verifiable through FINRA BrokerCheck.
Regulatory note
Windsor Drake is not a broker-dealer. Where an engagement involves the offer or sale of securities, including certain asset-backed and capital-raising transactions, those services are provided through Independent Investment Bankers, Corp., a FINRA member firm, under a separate engagement. Nothing on this page constitutes an offer to buy or sell any security or a solicitation of any such offer. All engagements are confidential.
Specialty finance and payments advisory questions
How is this different from selling the company?
It is the opposite end of the same expertise. A sale process presents the funding structure to buyers. This practice fixes it. The work is the warehouse facility, the borrowing base, the advance rate, the eligibility schedule, the covenant and cure package, forward flow terms, and the credit evidence supporting all of them. Owners engage it because growth has stalled, a renewal is coming, a single lender has become a single point of failure, or the equity consumed per dollar originated is higher than it needs to be. No sale is contemplated or required.
How is a warehouse advance rate set, and can it be renegotiated?
It is set first by asset class and then by evidence. Where a company sits inside its lender’s band is decided by documented loss performance, data quality and the lender’s confidence in the servicing, not by a published benchmark. Advance rates are renegotiated at renewal, at upsize and when a second lender is added, and the argument that moves them is static pool evidence rather than a request.
What is a borrowing base, and why do performing assets fall outside it?
The borrowing base is eligible receivables, less ineligibles, less concentration excesses, multiplied by the advance rate, recalculated daily or weekly rather than at quarter end. Assets fall outside it because of eligibility criteria that operate silently: minimum score thresholds, first payment received, a complete document package, a perfected lien, no delinquency beyond a stated ageing, and approved jurisdictions. Concentration limits do the same at pool level through single obligor, geographic and sector caps. A performing asset that fails any of those tests is funded with equity.
How does a forward flow agreement differ from a warehouse facility?
A warehouse is committed capacity to fund assets the company keeps. A forward flow is contracted demand to buy them. The buyer agrees to purchase assets meeting an agreed eligibility schedule at an agreed price over a period, which removes the funding requirement for the flowed portion and caps its yield at the same time. Terms that decide the economics include purchase price as a premium or discount to par, allocation targets and minimum allocation percentages, servicing consistent with the assets the originator retains, and soft stop and hard stop origination events. Volume is institutional: Upstart and Fortress Investment Group announced a $1.25 billion forward flow agreement on 29 April 2026, following a $1.2 billion agreement between the same parties in May 2025.
How should a repurchase obligation be sized and negotiated?
As a contingent liability with a measurable history, not as boilerplate. Loan-level representations, first payment default provisions, fraud provisions and eligibility breach mechanics all create repurchase exposure. Size it by measuring actual breach and early default rates in the company’s own book over several years, reserve against that number, and negotiate a stated cap, a defined cure window, a substitution right where the asset can be replaced rather than bought back, and a sunset after which representations expire. An uncapped repurchase obligation on a book the originator no longer controls is the term most likely to be underpriced at signing.
What credit data does a warehouse lender or forward flow buyer require?
Static pools, not portfolio averages. Item 1105 of Regulation AB, 17 CFR 229.1105, sets the public benchmark and is worth building against even where nothing is being registered: delinquencies, cumulative losses and prepayments, presented by prior securitized pool of the sponsor for that asset type, or by vintage origination year where the sponsor has less than three years of experience securitizing that asset type, over five years or the sponsor’s full history with the asset type if shorter, in each case to the extent material. In practice a counterparty also wants vintage curves by product, channel and credit band, charge-off and re-ageing policy with every historic change documented, recovery rates, roll and cure rates, and a loan tape that reconciles to the servicing system and the general ledger. Blended averages understate losses in a growing book because immature vintages dilute the mean.
What decides the economics of a card program?
Issuer status, sponsor bank terms, funding cost and fraud, in roughly that order. Interchange is set by network, product and whether the issuing bank is covered by the Durbin Amendment. Federal Reserve data published 19 December 2025 show that in 2024 covered debit issuers averaged $0.23 of interchange per transaction, or 0.47 percent of value, across all networks, against $0.51 and 1.21 percent for exempt issuers, and the Board’s companion biennial report covering 2023 put the average debit transaction at $46.26, covered issuer authorization, clearing and settlement costs at $0.041 per transaction and covered issuer fraud losses at 17.6 basis points of transaction value. That gap makes sponsor bank selection an economic decision before it is a compliance one. The rule is also unsettled, so program models should be tested against more than one interchange outcome.
What are funding conditions for specialty finance companies in 2026?
Available but demanding. Lending to nondepository financial institutions accounted for roughly 40 percent of all US bank loan growth in 2025 on Federal Reserve data analysed by American Banker, while the Federal Reserve’s July 2026 Senior Loan Officer Opinion Survey, released 3 August 2026, reported significant net shares of banks describing standards for every category of nondepository borrower as at the tighter ends of their historical ranges since 2011. Private credit has expanded into that gap, with the Alternative Credit Council and AIMA putting the global private credit market at US$3.5 trillion at the end of 2025, and KBRA forecasting approximately $385.2 billion of US ABS issuance in 2026, another post-crisis record. Capacity exists at every layer. Terms are set by the quality of the credit evidence a company can produce.
Sources: Item 1105 of Regulation AB, 17 CFR 229.1105; Federal Reserve Regulation II interchange fee data published 19 December 2025 and the Board’s biennial report covering 2023; Federal Reserve July 2026 Senior Loan Officer Opinion Survey, released 3 August 2026; Federal Reserve Bank of New York Household Debt and Credit Report for the second quarter of 2026, released 11 August 2026; Alternative Credit Council and AIMA private credit market sizing, year end 2025; KBRA US ABS issuance forecast published 20 November 2025; TransUnion research published 30 April 2026; Upstart and Fortress Investment Group forward flow announcement, 29 April 2026. Registration detail per FINRA BrokerCheck.
Windsor Drake Research · Published 13 August 2026