You need to enable JavaScript to run this app.
By Travis Yule — CEO & Founder, Full Send Funding
At 7.5% prime a year of $200,000 costs $26,387 on a 7(a) loan and $48,000 on a 24% advance: the $21,613 gap is the price of speed, spent by a sixty-day wait.
In one sentence: A 7(a) loan is cheaper per dollar and slower per day, so the choice is one division — the margin a business forgoes each month while it waits, times the wait, against the rate difference in dollars — and on $200,000 at an assumed 7.5% prime that difference is $21,613, which a $10,000-a-month need spends in sixty days.
An SBA 7(a) loan is the cheapest working capital most small businesses will ever be offered, and it is slow. On a $200,000 loan priced at the regulatory ceiling for that size — 6.0 points over the base rate[2], or 13.5% at an assumed 7.5% prime — the first year's interest is $26,387. The same $200,000 from us at a 24% cost of capital on a twelve-month term costs $48,000. The rate difference is $21,613, and that is the number the whole decision turns on, because it is what you are paying for speed. If the money is not time-sensitive, pay nothing for speed and take the loan. If it is, the question is whether the sixty or ninety days a 7(a) loan takes will cost the business more than $21,613 in margin it cannot earn while it waits — and at $10,000 of monthly contribution foregone, a sixty-day wait costs $20,000, which is nearly the whole difference before a single day of slippage.
The usual framing — bank money is cheap and alternative money is expensive, so borrow from the bank if you can — is not wrong about price. It is wrong about the unit. A rate is a cost per dollar per year; a delay is a cost per day; and the second one is invisible on every term sheet, because no lender charges you for the weeks you spent assembling the file. Once the delay is priced in the same dollars as the rate, the answer stops being a matter of principle and becomes arithmetic, and the arithmetic goes both ways: there are businesses that should wait ninety days for the loan and take the cheaper capital, and there are businesses for which the loan is the expensive option even at a third of our rate.
What follows is the honest timeline for each route, the program's published ceilings and the credit standard that has applied since 2026, the break-even computed from stated inputs, and a worked example with named numbers. We fund working capital on the fast side of this comparison, and we say plainly in the section on the wrong tool where that is not the answer. One note on sourcing, stated here and not repeated: the SBA's lender manual, Standard Operating Procedure 50 10, governs every requirement in the program, and this article was written without a copy of the current edition to hand. Where the SOP is the authority — the eligibility screen in particular — we describe it in outline and cite only the regulation and the procedural notices that state its terms. The regulation, the SOP and the loan documents control; the lender you apply to will apply them, not this article.
A 7(a) loan is a bank loan with a federal guaranty on part of it. The lender lends its own money; the Small Business Administration promises to cover a share of the loss if the borrower defaults. Most 7(a) loans carry a maximum of $5,000,000, the SBA guarantees up to 85% of loans of $150,000 or less and up to 75% of loans above that, and lenders pay an upfront guaranty fee they are permitted to pass on to the borrower.[1] Since July 2026 an eligible borrower may combine a 7(a) loan with a 504 loan for up to $10,000,000 in SBA-backed financing, where the previous cumulative limit was $5,000,000.[7]
Three consequences follow from that structure, and they explain almost everything about the timeline.
The lender is still a lender. The guaranty reduces the bank's loss if you default; it does not change the bank's decision about whether you are likely to. The file is a bank file: tax returns, financial statements, a debt schedule, a personal financial statement, a business plan or projections where the use of proceeds calls for one, collateral where it exists. The SBA has its own eligibility screen on top of the bank's credit policy, and a loan has to clear both.
The guaranty is a federal program, so it is governed by regulation and procedure. The rate ceilings are in the Code of Federal Regulations[2], the SBSS pre-screen was retired by SBA notice[3][4], and the eligibility rules are in the SOP. None of it moves with the loan officer's mood, which is the program's great strength and the reason a file that misses a requirement cannot be talked through it.
There are faster variants, and they are smaller. SBA Express carries a maximum of $500,000[1] and a lower guaranty in exchange for letting the lender use its own processes. The CAPLines family includes a Working CAPLine — an asset-based revolving line for businesses unable to meet the credit standards for long-term credit, repaid by converting short-term assets into cash — and, with the exception of the Builders line, CAPLines carry a maximum maturity of ten years.[8] These are still bank loans on bank files. Express shortens the SBA's step; it does not shorten yours.
The cap on a variable-rate 7(a) loan is a spread over a base rate, and the spread falls as the loan grows: for loans of $50,000 and less the rate may not exceed 6.5 percentage points over the base rate, for loans up to $250,000 the ceiling is 6.0 points, up to $350,000 it is 4.5 points, and above $350,000 it is 3.0 points.[2] The base rate most lenders use is the prime rate, which is set by individual banks rather than by the Federal Reserve and is reported on the Fed's H.15 release as the rate posted by the majority of the largest twenty-five banks.[5] We have not reproduced a current prime rate here; the figure below assumes 7.5% and shows 8.5% beside it so a reader can see what a point of prime does, and the release is the place to look up today's number.
The smallest loans carry the widest spread: at an assumed 7.5% prime a loan of $50,000 or less may be priced up to 14.0%, and a loan above $350,000 up to 10.5%.
Maximum spreads over the base rate from 13 CFR §120.214 — 6.5 percentage points for loans of $50,000 and less, 6.0 to $250,000, 4.5 to $350,000, 3.0 above — added to an assumed prime rate of 7.5% and of 8.5%. The prime rate is assumed, not quoted; the current figure is on the Federal Reserve's H.15 release. Lenders may price below the ceiling; the upfront guaranty fee is not included.
Two things to read off that chart. First, the ceiling is a ceiling: a lender may price below it and a strong file at a competitive bank often is. Everything in this article prices the loan at the cap, which is the conservative assumption in the SBA's favour, because it makes the rate difference as small as it can honestly be made. Second, the smallest loans are the dearest, by design — a $40,000 7(a) loan at the cap is a fourteen-point loan at an assumed 7.5% prime, which is a long way from the single-digit rate people picture when they say "SBA".
What the ceiling does not include is the upfront guaranty fee, which the lender pays and may pass through to the borrower.[1] The fee schedule is published by annual notice and we have not reproduced it here, so the arithmetic below leaves it out. Including it moves every comparison in the same direction — it makes the loan dearer, not cheaper — so a reader who adds it is being fairer to us than the tables are.
The rule most articles still describe is no longer in force. The SBA historically pre-screened 7(a) Small Loan applications with a FICO SBSS score against a published minimum. Procedural Notice 5000-875701, effective January 16, 2026,[3] and the supplemental guidance that followed on March 1, 2026,[4] sunset that screen: the SBA no longer screens or assigns an SBSS score for 7(a) Small Loan applications, and lenders may use the same policies and processes they use to review their non-SBA loans of the same type and size.[3] What stands in its place is the lender's own conventional credit analysis — the policy it applies to its non-SBA loans of the same type and size[3] — and for a term loan that analysis turns on a debt service coverage test. For a 7(a) loan the floor is published: SOP 50 10 8 sets the minimum debt service coverage ratio at 1.0× (1:1) for 7(a) Small Loans — loans of $350,000 or less — and 1.15× for loans over $350,000.[9] A lender's own credit policy may ask for more than the floor, and the lender you apply to will apply its policy.
That is the requirement that, in our experience, most often stops a working-capital applicant — there is no public dataset of 7(a) decline reasons — and it is worth being precise about what it measures. Debt service coverage is cash available to service debt divided by scheduled debt payments. Whether the multiple is the SOP's floor or the lender's higher figure, the business has to show — from historical financial statements, not from projections alone — that it generates more cash than the principal and interest it will owe after the new loan is added, with a margin on top. A business that is profitable on paper but whose owner draws the profit out, a business whose last tax return shows a loss because it expensed a build-out, a business in its first eighteen months with no full-year statement to show: each of these can be a perfectly sound working-capital file and still fail a coverage test that is computed from documents it does not have or that do not yet say what the bank account says.
The remainder of the screen is eligibility rather than credit, and it is the part the SOP owns. In outline — the SOP's section numbers are not quoted here — it asks that the business be a for-profit operating company in the United States, within the SBA's size standard for its industry, with owners who have invested their own equity, that the type of business not be one the program excludes, that the owners not be delinquent on federal debt, and that the business be unable to obtain the credit on reasonable terms elsewhere. What matters for the timeline is that each of those conditions is documented, and every document is a day.
This is information, not legal advice; the regulation, the SOP and the loan documents control, and an attorney or your lender's SBA department should read yours.
There is no public dataset for the time a 7(a) loan takes from first conversation to funded. The SBA publishes its own review turnaround per program on its lender pages, and that step is the shortest one. The lender's stages before and after it — assembling the file, packaging, underwriting, committing, ordering third-party reports, closing — are not reported anywhere, and they are where the weeks go. What follows is our working model of a standard 7(a) working-capital loan at a bank, built from what borrowers who come to us after starting one describe, and stated as a range per stage so a reader can substitute their own lender's numbers. It is an estimate, not a survey.
Full Send Funding's working model of a standard 7(a) working-capital loan at a bank, stated as a low and high number of calendar days per stage and summed; it is an estimate built from what borrowers who began a 7(a) application describe, not a measured average, and no public dataset of application-to-funding time exists for the program. The low end assumes a complete file, a lender with delegated authority and no real estate or equipment to appraise. Our own timeline (decision within 24 business hours, funding within 24 hours of approval, one to three business days end to end) is stated in the prose rather than the table because it does not pass through these stages.
Read the total row and then the first row. Five to thirteen weeks is the honest range for a complete, eligible file; the low end assumes a borrower who has three years of clean returns and current financial statements ready on day one, a lender that holds delegated authority and can make the guaranty decision itself, and a loan with no real estate or equipment to appraise. The high end is a file that is missing one of those. And the first row is the one most borrowers underestimate: the loan is not slow because the bank is slow, it is slow because a bank file is a large document set and most operating businesses do not have it sitting in a drawer.
Our path is not in that table, because it does not have those stages, but for scale the honest version of it is also a range. The file is three months of bank statements and a ten-minute application; there is no separate packaging or guaranty step; the agreement is signed electronically. A complete file gets a decision within 24 business hours and funding within 24 hours of approval; end to end, one to three business days is typical, and a file submitted before 2pm ET with everything in it can fund the same day. The reason it is fast is not that we skip the analysis. It is that the evidence we read, the bank statement, already exists, is produced by a third party, and takes an underwriter an hour to read rather than a borrower three weeks to assemble. How working capital underwriting actually works covers what that hour looks at.
The comparison people make is rate against rate, and it is the wrong one, because the two products are measured in different units. Put both in dollars over the same period and the decision has three inputs.
The rate difference in dollars, D. The cost of the alternative over its term, less the cost of the loan over the same period. On our worked example that is $48,000 less $26,387, or $21,613. It is the price of speed, and it is fixed the day you sign. One thing that comparison quietly concedes to us, and a CFO will see it: the loan's $26,387 buys the use of nearly the whole $200,000 for the whole year, because only $10,159 of principal comes back in twelve months, while the advance's $48,000 buys a balance that starts at $200,000 and is returned weekly, so the average amount outstanding is a little over $100,000. Per dollar actually in hand per day the advance is dearer than the two totals suggest — the APR and factor-rate conversion shows how much — and it cuts the opposite way from the guaranty fee we left out. We keep the comparison in totals because the business is asking what a year of this capital costs it, and the totals answer that; a reader who converts both to an annualised rate should expect the gap to widen, not narrow.
The cost of waiting per day. The contribution margin, revenue, discount or penalty the business forgoes for each day the capital has not arrived. On an inventory purchase it is the gross margin on the sales you cannot make; on a payroll gap it is the crew you cannot keep; on an early-payment discount it is the discount; on a tax deadline it is the penalty. Call the monthly figure M, so the daily figure is M ÷ 30.
The wait, W, in days — the timeline above, or your lender's.
Waiting is right when M × W ÷ 30 is less than D, and wrong when it is more. The break-even monthly margin is D × 30 ÷ W: on a thirty-day wait, $21,613; on forty-five days, $14,409; on sixty, $10,807; on ninety, $7,204. A business that forgoes more than $7,204 a month of contribution by waiting ninety days for the loan has paid more for the loan than it would have paid us, and the rate on the term sheet never showed it.
At $10,000 a month foregone the cost of waiting crosses the $21,613 rate difference at about sixty days; at $20,000 a month every wait beyond a month is dearer than the faster capital.
Cost of waiting = monthly margin foregone × days waited ÷ 30. Rate difference D = $48,000 (a $200,000 advance at a 24% cost of capital on a twelve-month term) − $26,387 (first-year interest on a $200,000 7(a) loan at 13.5%, the ceiling of 6.0 points over an assumed 7.5% prime, on a ten-year amortization) = $21,613.
The chart makes the shape visible: the cost of waiting grows with the wait and with the margin at stake, and the rate difference is flat. At $5,000 a month foregone, every wait on the chart is cheaper than the rate difference — that business should take the loan. At $20,000 a month, every wait longer than about a month is dearer, and at ninety days the delay costs nearly three times the rate difference. At $10,000 a month the two cross at about sixty days, which is why the timeline table matters: for a business in that band, the difference between a five-week close and a thirteen-week close is the whole decision.
One refinement makes the model honest rather than merely tidy. The SBA route is uncertain: in the Federal Reserve's most recent Small Business Credit Survey, 42% of employer firms that applied for financing received the full amount they sought, 36% some or most of it, and 22% none.[6] Those figures are for all financing and all lenders, not for 7(a) loans, and we have not found a public approval rate specific to the program, so the table below treats the probability of a decline as an input to vary rather than a fact to cite. A decline after sixty days means the business has paid the sixty days of waiting and then pays the alternative's cost anyway, which is why the expected cost of the loan-first route rises quickly with the chance it fails.
Loan-first route = margin lost in the wait ($10,000 × 60 ÷ 30 = $20,000) + p × the advance's cost ($48,000, since a declined applicant takes the advance anyway) + (1 − p) × the loan's first-year interest ($26,387: $200,000 at 13.5% on a ten-year amortization). Advantage = $48,000 − loan-first cost. The decline probability is an input to vary, not a sourced program statistic; the Federal Reserve's Small Business Credit Survey figures for all financing (42% received the full amount, 36% some or most, 22% none) are quoted in the text for scale only.
At a zero chance of decline the loan-first route wins by $1,613 on these inputs — the $21,613 of rate difference less $20,000 of margin lost in the wait. That advantage is gone once the chance of a decline reaches $1,613 ÷ $21,613, about 7.5%, or one file in thirteen. At one chance in four it loses by $3,790, and at even odds by $9,194. The practical reading: if you would honestly put your own file's odds at a bank below three in four, on a need that costs $10,000 a month to leave unfunded, the loan-first route is the expensive one on expectation even though it is the cheap one on the term sheet.
A regional beverage distributor with $210,000 of average monthly deposits lands a new retail account. Stocking it needs $200,000 of inventory up front; the account turns that inventory roughly every sixty days at an 18% gross margin, which is $36,000 of gross margin per turn or about $18,000 a month once it is running. The account is contracted from a start date eight weeks out and the retailer will not hold the shelf space past it.
Route one: the 7(a) loan. The business is four years old, has three years of returns, and its last full-year statement shows coverage a bank will accept. Priced at the ceiling for a $200,000 loan — 6.0 points over an assumed 7.5% prime, 13.5% — on a ten-year amortization, the payment is $3,045.49 a month. In the first twelve months the business pays $36,546 in total, of which $26,387 is interest and $10,159 is principal, and it still owes $189,841 at the end of the year. Over the full ten years the interest is $165,458, but that is not the fair comparison, because the business keeps the money for ten years and the alternative gives it back in one; the like-for-like figure is what a year of the capital costs, and that is $26,387.
Route two: the advance. $200,000 at a 24% cost of capital on a twelve-month term — cost of capital here meaning total cost as a share of the amount funded, so the payback is $248,000 and the cost is $48,000 — collected as fifty-two weekly remittances of $4,769.23 against weekly deposits of about $48,461. That is 9.8% of receipts, which clears the business's worst normal week with room, which is the test we actually apply. The money is in the account inside the week.
Loan: $200,000 at 13.5% (6.0 points over an assumed 7.5% prime, the ceiling for a loan of $50,001 to $250,000), amortized over 120 months; interest and principal are summed over months 1–12 of that schedule. The upfront guaranty fee the lender may pass through is excluded. Advance: cost of capital means total cost as a share of the amount funded, so payback is $200,000 × 1.24 = $248,000, cost $48,000, and the weekly remittance is $248,000 ÷ 52.
The decision. D is $21,613. The account is worth about $18,000 a month; the wait for the loan, on a good file at a good lender, is five to thirteen weeks. At five weeks the wait costs $21,000 in margin — a dead heat with the rate difference, before the risk that the file takes longer or the shelf space goes. At eight weeks it costs $33,600 and the account may be gone entirely, which is not a cost of $33,600 but a cost of the whole account. This business should take the advance, buy the inventory, run the account, and — this is the part that gets left out — apply for the 7(a) loan anyway, with the new account's first three months of statements in the file, to refinance the position and carry the permanent working-capital base at the bank's rate. The advance was the right instrument for the eight-week deadline; the loan is the right instrument for the next ten years; and the mistake is treating them as competitors for one decision instead of as a sequence.
Now change one input. Same business, same $200,000 need, but the use is a warehouse refit that will lower operating costs by $3,000 a month once done, with no deadline. The cost of waiting is $3,000 a month against a rate difference of $21,613, and the break-even wait is more than seven months. That business should wait for the loan, and if we were asked we would say so.
The table below restates the break-even for a range of our own pricing, because 24% is one point in a range that runs from 4.5% to 45% cost of capital depending on qualifications and term length, and the answer changes with it.
Cost = $200,000 × the share in the first column. D = that cost − $26,387 (the loan's first-year interest at 13.5% on a ten-year amortization). Break-even monthly margin = D × 30 ÷ wait in days, rounded to the dollar; where D is negative the loan costs more for the year than the advance and there is no wait worth taking.
The first row is worth a sentence. A file that qualifies for the low end of our range on a twelve-month term is paying less for a year of the capital than the 7(a) ceiling charges, before the guaranty fee. That is not the usual case and we do not lead with it — the usual case is the 24% row — but it is a real one, and it is why "bank money is always cheaper" is a rule of thumb rather than a rule.
The two files are different documents answering the same question — will this business make the payments — from different evidence, and the difference is the source of both the price and the timeline.
The bank reads history. Two to three years of tax returns and financial statements, a debt schedule, a personal financial statement for each owner of 20% or more, a coverage ratio computed from the last full year, collateral where it exists, and since 2026 the bank's own credit policy — the one it applies to its non-SBA loans of the same size — in place of a score.[3] The evidence is retrospective and prepared: it shows what an accountant said the business did, annually. Its strength is depth. Its weakness is that it cannot see the last ninety days, and it is expensive to produce.
We read the bank account. Three months of statements, the deposit pattern month to month, the average daily balance, the negative days, and the existing positions already drawing on the same deposits. The evidence is contemporaneous and produced by a third party: it shows what the business actually did, day by day, through last Friday. Its strength is speed. Its weakness is that three months is a short window, which is why the payment is sized to the weakest normal month rather than the average one, and why deposit consistency matters more than deposit size. The credit pull is a soft credit pull only; roughly 90% of applicants are approved; and there is no collateral, no UCC-1 lien filed against the business, and no confession of judgment in the agreement. Ask any lender, a bank included, the same three questions.
The consequence for a borrower deciding between the two is a rule of thumb about documents, not about credit. If your last full-year financial statement already shows the coverage the bank needs and the returns are filed, the 7(a) file is mostly assembly and the low end of the timeline is realistic. If the business's best evidence is the last six months of deposits — because it is young, because it grew, because last year's statement carries a one-time cost — then the bank file will not show what the bank account shows, and the loan is not slow so much as unlikely, which is a different problem with the same cure: take the capital that reads the account, run the year, and apply for the loan with the year in hand.
We are the fast option in this comparison, so it is worth being unambiguous about the cases where the fast option is the wrong one. The section after this one covers the reverse case, and does not repeat these.
Two more things a business that waits should do while it waits. Keep every dollar of revenue running through one business account, because the bank will eventually want the statements too. And do not take a short-term position "to hold things over" while the loan is in process: a new daily debit appearing in the statements mid-underwriting changes the coverage ratio the bank is computing, and it changes it in the wrong direction. The stacking trap describes the mechanism; it applies to a bank file as much as to ours.
The 7(a) loan is the wrong tool for a need that closes before the loan can, for a business without the financial statements the file requires, for a business whose last full year does not show the coverage the bank requires even though the last three months plainly would, for an owner unwilling to sign a personal guarantee or pledge collateral where the lender asks for either, and for an amount small enough that the ceiling — 14.0 points at an assumed 7.5% prime for a loan of $50,000 or less[2] — is no longer the bargain the program's reputation suggests. It is also the wrong tool as a bridge: a ten-year loan taken for a sixty-day gap leaves the business carrying nine years and ten months of a facility it did not need, and the interest on that is real even if the rate is low.
Our capital is the wrong tool in every case in the list above, and in three more: for a business whose receipts are lumpy enough that a fixed weekly remittance cannot be sized to the worst week; as a second or third position on top of an existing one, which is stacking and is a decline here, not a discount — this is the one we decline for most; and for a structural loss. A structural loss is a business that spends more than it makes in a normal month; borrowing adds a payment to it, and no rate, ours or the bank's, changes that. When borrowing is the wrong answer is the test to run before either application.
And neither is the right tool alone for the commonest case in this article: a time-sensitive need in a business that will qualify at a bank once it has a year in hand. That business should take the fast capital for the deadline and the loan for the base, in that order, and it should tell each lender about the other. A small-business loan from us on a longer term is sometimes the middle instrument — longer than a twelve-month advance, faster than a bank, sized to the deposits — and the side-by-side comparison of funding options puts all three on one page.
A 7(a) loan is a bank loan with a federal guaranty on part of it: up to $5,000,000, guaranteed 85% to $150,000 and 75% above[1], priced at most 6.5 to 3.0 points over the base rate depending on size[2], and since January 2026 underwritten on the lender's own credit policy in place of the SBSS screen,[3] above a debt service coverage floor of 1.0× for Small Loans of $350,000 or less and 1.15× above.[9] From a complete, eligible file it takes five to thirteen weeks by our working model, and there is no public dataset that says otherwise. Our working capital takes one to three business days from three months of bank statements.
The rate difference is real — $21,613 in a year on $200,000, comparing our 24% cost of capital against the loan at its ceiling — and it is the price of speed. Whether it is worth paying is decided by one division: the margin the business forgoes each month while it waits, times the wait in months, against the rate difference in dollars. Below the break-even, wait for the loan and we will tell you to. Above it, take the capital, run the business, and apply for the loan with the results in the file. The two products are a sequence far more often than they are a choice.
If you want to know where your own numbers land, check what you would qualify for — a minute, soft pull only — or call 518-312-0382 and ask us directly whether you should be waiting for the bank.
The figures in this article are derived from stated inputs, not surveyed. The loan is priced at the regulatory ceiling for its size — 6.0 points over an assumed prime rate of 7.5%, or 13.5%, for a $200,000 loan[2] — and amortized over 120 months, with interest and principal summed over the first twelve months; the ceiling is the conservative assumption in the loan's favour, since lenders may price below it. The prime rate is assumed and the figure that uses it shows 8.5% beside it; the current rate is on the Federal Reserve's H.15 release.[5] The upfront guaranty fee the lender may pass through[1] is excluded because its schedule is set by annual notice and was not reproduced here; including it makes the loan dearer in every comparison. The advance is priced at a 24% cost of capital on a twelve-month term, cost of capital meaning total cost as a share of the amount funded, and the break-even table restates the result at 12%, 18% and 36%. Every published value is recomputed from those inputs by the site's arithmetic audit before publication.
The stage-by-stage timeline is Full Send Funding's working model of a standard 7(a) working-capital loan at a bank, stated as a low and a high number of calendar days per stage and summed. It is an estimate built from what borrowers who began a 7(a) application describe to us; it is not a measured average, and no public dataset of application-to-funding time for the program exists. The SBA publishes its own per-program review turnaround on its lender pages, which was not quoted because no extract of it was confirmed for this article. The decline-risk table treats the probability of a bank decline as an input to vary; the Federal Reserve Small Business Credit Survey figures quoted for scale — 42% of applicants received the full amount, 36% some or most, 22% none — are for all financing and all lender types among employer firms with 1–499 employees, from a nationwide convenience sample fielded in autumn 2025, and are not 7(a) approval rates.[6]
The program facts are stated from the primary documents cited: the SBA's published terms for the 7(a) program[1], the rate-ceiling regulation[2], the two procedural notices that sunset the SBSS screen[3][4], the SBA's announcement of the $10,000,000 cumulative limit[7] and its description of the CAPLines[8]. The eligibility conditions that live in Standard Operating Procedure 50 10 are described in outline only, without section numbers, because the current SOP was not available to confirm for this article. The debt service coverage floors — 1.0× for 7(a) Small Loans of $350,000 or less and 1.15× above — are stated from the SBA's information notice issuing SOP 50 10 8;[9] the statement that the coverage test is the requirement that most often stops a working-capital applicant is the firm's experience, not a measured share, and no public dataset of 7(a) decline reasons exists. The SOP and the loan documents control.
There is no public dataset. By our working model a standard 7(a) working-capital loan runs 34 to 94 calendar days — five to thirteen weeks — from a complete, eligible file, most of it in assembling the file, underwriting and closing rather than in the SBA's own review. A borrower with three filed years, current statements and a lender with delegated authority is at the short end.
For variable-rate loans the regulation caps the spread over the base rate at 6.5 percentage points for loans of $50,000 and less, 6.0 up to $250,000, 4.5 up to $350,000 and 3.0 above that. At an assumed 7.5% prime those ceilings are 14.0%, 13.5%, 12.0% and 10.5%; lenders may price below them, and the upfront guaranty fee is separate.
The lender's own credit analysis, built on a debt service coverage test. Procedural Notice 5000-875701, effective January 16, 2026, and supplemental guidance effective March 1, 2026, sunset the SBSS screen for 7(a) Small Loans, and lenders may now use the same policies and processes they use for their non-SBA loans of the same type and size. SOP 50 10 8 sets the coverage floor at 1.0× for Small Loans of $350,000 or less and 1.15× above, and in our experience that test is what most often stops a working-capital applicant.
When the margin you forgo each month while waiting, multiplied by the wait in months, is less than the rate difference in dollars. On $200,000 against our 24% pricing the rate difference is $21,613, so a need that costs under about $7,200 a month to leave unfunded should wait ninety days for the loan, and a refit that saves $3,000 a month should wait as long as it takes.
Yes, and it is the commonest right answer in this article: take the fast capital for the deadline, run the business, and apply for the loan with the results in the file to carry the permanent base at the bank's rate. Tell each lender about the other, and do not add a second short-term position while the bank file is in underwriting, because it changes the coverage ratio the bank is computing.
At the lower part of our range, yes. A twelve-month advance at a 12% cost of capital costs $24,000 on $200,000, against $26,387 of first-year interest on the loan at its ceiling before the guaranty fee. That is not the usual case — the usual case is the 24% row, where the loan is $21,613 cheaper for the year — but it is a real one.
Travis Yule founded Full Send Funding in 2021 and leads it from Middle Grove, New York. He writes about working capital from the underwriting side of the table — what the numbers actually have to say before a business gets funded.