A Failed DSCR Is Rarely a Failed Deal. It Is Usually a Failed Structure.

36 minutes ago

22 min read

Bridge under construction.

Summary

On October 1, 2026, SBA raised the coverage floor for most 7(a) changes of ownership from 1.15x to 1.25x and made it a historical test. A business that could support $2.08 million of ten-year debt at 10.00 percent under the old floor now supports $1.92 million. The same earnings carry 8 percent less debt, and projections no longer close the gap.

Of the eight structural levers lenders reach for when coverage fails, four still do most of the work: amortization within the program caps, the capital stack, resizing the loan to what coverage supports, and real, unborrowed equity. Seller paper now helps only on full standby for the life of the loan. Interest-only periods do nothing for an acquisition tested on historical results. Rate concessions are too small to close a real gap: on our case deal, each 100 basis points moves coverage by about 0.05x.

On a $2.5 million acquisition of a business and its building, with $380,000 of adjusted EBITDA, the term sheet as first drafted produces 1.07x. A single 7(a) loan on the 16-year blended maturity the SOP allows produces 1.35x without a dollar of new equity. A 504 for the building plus a ten-year 7(a) for the business produces 1.30x on day one, but holds up better under a 200 basis point rate shock and costs about $420,000 less in total payments.

Written for lenders, CDCs, loan brokers and borrowers working a file that does not cover. 22 minute read.

Key figures

1.25x. Minimum historical DSCR for 7(a) Initial Acquisitions, Owner Buyouts and ESOP or cooperative purchases. Source: SBA SOP 50 10 8.1, Appendix 15. Effective October 1, 2026.

8 percent. Debt capacity lost on the same EBITDA when the floor moves from 1.15x to 1.25x. Source: MMCG Invest calculation. As of November 2026.

10 years. Maximum amortization on the business portion of a 7(a) change of ownership. Source: SBA SOP 50 10 8.1, Appendix 15. Effective October 1, 2026.

50 percent. Share of the required equity injection that full-standby seller debt and minority equity may cover, combined. Source: SBA SOP 50 10 8.1, Appendix 15. Effective October 1, 2026.

3 years. Longest permitted wait before the first full principal and interest payment on a USDA B&I loan. Source: 7 CFR 5001.402(b)(3). As amended September 30, 2024.

7.00 percent. WSJ Prime Rate, which caps 7(a) variable pricing on loans over $350,000 at 10.00 percent. Source: Bankrate, WSJ Prime Rate table. Effective September 17, 2026.

When debt service coverage misses, the business is rarely the problem and the structure usually is. These are the eight levers that fix it, ranked by what still works under SOP 50 10 8.1.

Ask a loan broker what to do when coverage misses and you will usually hear one of two answers: stretch the amortization, or bring in a seller note. Eighteen months ago both answers were defensible. Today one of them still works and the other mostly does not.

The instinct is right. The playbook is out of date. A failed DSCR is rarely a verdict on the business. It is usually a verdict on a term sheet that asks the business to repay debt faster than the rules require, or that loads debt the structure did not need. In most credit memos we have read, and a fair number we have written, the distance between 1.07x and 1.25x is closed by structure, not by better earnings.

What changed is the rulebook. SBA's SOP 50 10 8, effective June 1, 2025, stopped counting a seller note as equity unless it sits on full standby for the life of the loan. SOP 50 10 8.1, effective October 1, 2026, went further for changes of ownership. It set the coverage floor at 1.25x on historical results, capped the business portion at ten years of amortization, removed the election that gave a loan 25 years once real estate reached 51 percent of proceeds, and imputed a ten-year amortization to any seller note that is not on full standby.

This piece takes one acquisition that fails coverage and runs it through eight levers: amortization, loan size, equity, seller paper, interest-only periods and reserves, the capital stack, rate, and the numerator itself. For each we name the controlling rule, label what is rule and what is convention, show the arithmetic, and say what makes an underwriter reject it. Our companion flagship on DSCR in feasibility studies covers how coverage is defined and measured. This one is about what to do when it fails.

The floor moved on October 1

There is no single DSCR minimum in U.S. small-business lending. There are program floors, a regulatory requirement that each bank set its own, and a market convention that most people mistake for a rule. The SBA regulation itself, 13 CFR 120.150, requires only that loans be "so sound as to reasonably assure repayment." The ratios live in the SOP. Before choosing a lever, an underwriter needs to know which floor the file must clear, and on what basis.

Program or transaction	Minimum DSCR	Basis	Rule or convention	Source
SBA 7(a) Standard, over $350,000	1.15x business; 1.00x global	Historical and/or projected; projections must reach 1.15x within 2 years of funding	Rule	SOP 50 10 8.1, Sec. B, Ch. 1
SBA 7(a) Small, $350,000 or less	1.10x	Historical, or projected within 1 year	Rule	PN 5000-876777, effective March 1, 2026
7(a) Initial Acquisition, Owner Buyout, ESOP or cooperative	1.25x	Last fiscal year or 2-year average, historical or adjusted; projections not relied on	Rule	SOP 50 10 8.1, App. 15
7(a) Business Expansion	1.15x	Same as above	Rule	SOP 50 10 8.1, App. 15
SBA 504	1.15x historical (1.00x under SOP 50 10 8)	Operating company; 2 years of projections if history falls short	Rule as reported; wording not yet conformed	SOP 50 10 8.1, Sec. C, Ch. 1
USDA B&I	No general floor; 1.1x historical for majority-refinancing loans.	Lender documents repayment ability; global analysis for affiliates	Rule without a number	7 CFR 5001.202
Bank CRE and C&I	No regulatory number; 1.20x to 1.35x common	The bank's own written policy	Policy required by rule; number is convention	Interagency Guidelines; OCC
Agency multifamily, conventional	1.25x	Amortizing basis, underwritten net cash flow	Rule for that program	Freddie Mac Optigo term sheet

Three claims deserve an audit.

"SBA requires 1.25x." False as a general statement, true since October 1 for one category. Under SOP 50 10 8, in force from June 1, 2025 through September 30, 2026, no 1.25x figure appears. Appendix 15 of SOP 50 10 8.1 introduced it for Initial Acquisitions, Owner Buyouts and ESOP or cooperative purchases. Business Expansions stay at 1.15x and 7(a) Small loans at 1.10x. Borrowers who have heard 1.25x on other SBA loans have heard bank policy layered on top of SBA.

"Banks require 1.20x to 1.25x." No regulator says so. The Interagency Guidelines for Real Estate Lending Policies require each bank to set "minimum standards for net worth, cash flow, and debt service coverage," and stop there. The FDIC examination manual states that its guidance "will not include or allude to acceptable ratios." The closest thing to a regulator's number is a superseded OTS handbook illustration: a 1.2x origination minimum with exceptions down to 1.1x. It was an example, and it became folklore.

"Below 1.00x is an automatic decline." No primary rule we found says this. A sub-1.00x file fails every SBA floor as structured, which is a different thing. SOP 50 10 8.1 names the cure itself: the applicant may contribute additional equity "to reduce the loan amount and corresponding debt service in order to meet the DSC requirement."

The move from 1.15x to 1.25x looks small. It is not. Debt capacity is cash flow divided by the required ratio, so the same EBITDA now carries 1.15 ÷ 1.25, or 92 percent, of the debt it carried in September. Every lever below has to be sized against that 8 percent haircut. The new floor did not get harder for good businesses. It got harder for loose structures.

One deal, eight levers

The case is illustrative, but typical of the acquisitions we see. A buyer is purchasing an operating business together with the building it occupies. Total project cost is $2,500,000: $1,000,000 for the real estate and $1,500,000 for goodwill, equipment and working capital. The buyer injects the 10 percent minimum, $250,000, and requests a single 7(a) loan of $2,250,000. It is an Initial Acquisition, so Appendix 15 applies: 1.25x, historical. The purchase price is below $3,000,000, so no Quality of Earnings report is required.

Adjusted EBITDA for the last fiscal year is $380,000. That figure already adds back $84,000 of rent the business paid to the seller's real estate entity, which the acquisition debt replaces.

The term sheet as first drafted prices the loan at the 7(a) variable cap for loans over $350,000: Prime plus 3.00 percent, or 10.00 percent at today's 7.00 percent Prime. It amortizes everything over ten years. Annual debt service is $356,807, and coverage is 1.07x against a 1.25x floor. To pass, debt service must fall to $304,000, a reduction of $52,807 a year, or EBITDA must rise to about $446,000.

Lenders see this file every week. What follows is what each lever does to it.

Lever 1. Amortization: the cheapest fix is time

Amortization is the most powerful lever because it costs the borrower no cash at closing. It is also the most constrained.

The controlling rule for 7(a) is 13 CFR 120.212: ten years for working capital and most business assets, up to 25 years for real estate plus the time needed to complete construction, and equipment by useful life. SOP 50 10 8.1 tightened how mixed-purpose acquisitions use that rule. The business portion of a change of ownership may not exceed ten years. When real estate is in the deal, the lender either makes two loans, real estate on 25 years and business on ten, or makes one loan on a blended weighted-average maturity, rounded to the nearest year and computed on project cost before equity. The one exception in the Technical Policy Updates is an owner-occupied special purpose property, such as a hotel, car wash or assisted living facility, bought with its business, which may still take a single 25-year loan.

In our case, real estate is 40 percent of project cost. The blended maturity is 10 + 15 × 0.40, or 16 years. The arithmetic surprises some of the lenders we walk through it.

Structure	Annual debt service	DSCR
As drafted: one loan, 10 years	$356,807	1.07x
Two loans: $900,000 on 25 years, $1,350,000 on 10 years	$312,224	1.22x
One loan on the 16-year blended maturity	$282,394	1.35x

The single blended loan beats the split. Splitting isolates the goodwill on its ten-year clock, where each dollar of debt is most expensive to service. Blending spreads the real estate's longer life across the whole balance, and it clears the floor without new equity.

USDA B&I is looser still. 7 CFR 5001.402(a) limits maturity to "the justified useful life of the project or assets used as collateral" and to 40 years, and allows no balloons outside servicing. Conventional banks typically amortize real estate over 20 to 25 years with a five- or ten-year balloon. That is convention, and it is where most conventional files lose the comparison.

What makes an underwriter reject it: a blended term that stretches goodwill or working capital past ten years without the real estate share to justify it, or a B&I term longer than the collateral's useful life. Both are eligibility defects, and SBA tends to find the first one at guaranty purchase rather than at approval.

Lever 2. Resizing: let coverage set the loan

Resizing is the honest lever. It asks what the business can carry and lends that, rather than asking what the purchase needs and hoping coverage follows.

The arithmetic is short. Maximum debt service is cash flow divided by the required ratio, less any other debt service. Maximum loan is that figure divided by the annual payment constant. On the drafted ten-year term, $380,000 ÷ 1.25 gives $304,000 of debt service. At a constant of 0.1586, that supports a loan of about $1,917,000. The request was $2,250,000, so the gap is $333,000.

SOP 50 10 8.1 makes resizing the default outcome on acquisitions. Total 7(a) debt supporting a change of ownership may not exceed the independent business valuation. Where a Quality of Earnings report is required, at a business purchase price of $3,000,000 or more, the QoE earnings figure governs the coverage test. Practitioners who have read the Word version summarize the effect plainly: if coverage does not support the price and structure, the loan is reduced.

Banks typically size to the higher of the SBA floor and their own policy, often 1.25x to 1.35x, and stress variable loans by 200 basis points. That is convention. It is also why a file that clears SBA's floor by 0.02x has not yet cleared credit.

What makes an underwriter reject it: a resize that leaves the borrower short of working capital. The gap has to come from somewhere, which brings us to the next lever.

Lever 3. Equity: the only cure the SOP names

Each dollar of unborrowed equity removes a dollar of debt and its debt service. At a 10.00 percent, ten-year constant, every $100,000 of added equity frees about $15,860 a year. To close our case on the drafted term, the buyer would need $583,000 of total equity, 23 percent of project cost instead of 10.

The floors are rules. SOP 50 10 8, effective June 1, 2025, restored a 10 percent minimum injection for startups and complete changes of ownership. Under SOP 50 10 8.1, the 10 percent on an Initial Acquisition may not be reduced or waived. On a Business Expansion or Owner Buyout it may be reduced where the business shows positive net worth and sufficient liquidity. The 504 borrower contribution is 10 percent, rising to 15 percent for a new business or a limited or single-purpose property, and to 20 percent for both. USDA B&I requires "a minimum of 10 percent balance sheet equity" at closing for an existing business, or a debt-to-equity ratio of no more than 9 to 1, with higher thresholds for new businesses.

What counts matters as much as how much. As practitioners have reported it, Appendix 15 sorts equity into two buckets. Unlimited sources include unborrowed cash, a personal loan to a guarantor repaid from something other than business cash flow, and grants without clawback. Limited sources, meaning full-standby seller debt and non-controlling minority equity, may together cover no more than half the required injection.

Minimum equity gets a deal approved. Adequate equity gets it to stabilization.

What makes an underwriter reject it: a borrowed injection quietly repaid from business cash flow, gift funds without the donor's bank statements, or a minority investor holding a preferred return or redemption right, which disqualifies the money as equity.

Lever 4. Seller paper: standby or nothing

The seller note was the broker's favorite lever for a decade. Under SOP 50 10 7.1, a note on 24 months of standby could count toward the injection. SOP 50 10 8 ended that on June 1, 2025. A seller note now counts only if it sits on full standby, with no principal and no interest paid, for the life of the SBA loan, and only up to half the required injection. SOP 50 10 8.1 kept that rule and added a harder one. Seller debt that is not on full standby sits in the coverage denominator, and if it is interest-only, it is underwritten as though it amortized over no more than ten years.

The arithmetic explains why brokers have gone quiet. Suppose the seller carries $250,000 of the price and the 7(a) loan falls to $2,000,000.

Seller note terms	Debt service counted	DSCR
Full standby for the life of the loan	$317,162 (SBA loan only)	1.20x
Interest-only at 6.00 percent, not on standby	$350,468 (note imputed at 10 years)	1.08x
Same note, if actual interest were counted	$332,162	1.14x

Full standby is the only version that helps, and even it does not close the gap alone. It also asks a seller to wait ten years for a dollar of payment, which most sellers will not do. A non-standby note improves coverage by 0.02x, because the imputed amortization erases the benefit of the lower coupon.

Documentation is a rule: SBA Form 155, or a lender equivalent, with the subordinated note attached. Lenders often also require a 24- to 36-month payment standby on seller notes that do not count as equity. That is convention, not SBA. Under USDA B&I, subordinated debt counts toward balance sheet equity only "when subject to a standstill agreement for the life of the loan."

What makes an underwriter reject it: consulting agreements or side payments that function as disguised debt service, standby agreements that let payments resume on a covenant test rather than at payoff, or seller paper that funds price above the independent valuation without full standby.

Lever 5. Interest-only and reserves: liquidity, not coverage

An interest-only period lowers the payment. It does not lower the debt. In our case, interest-only at 10.00 percent would cost $225,000 a year and show 1.69x. No SBA reviewer will look at that number on an acquisition. Appendix 15 tests historical coverage against the debt service the business will actually carry, and projections may not be relied on. For a change of ownership, an interest-only period changes nothing in the test.

Interest-only earns its place where the business is not yet operating. SBA's standard 7(a) rule asks startups and construction projects to show 1.15x within two years of funding, or of the end of construction. In a 504, interim interest is an eligible project cost under 13 CFR 120.882, and the contingency reserve may run up to 10 percent of construction cost. Where 504 projections show repayment in year two but not in year one, the CDC must show liquidity to cover the shortfall.

USDA B&I is the most explicit of the three programs. Under 7 CFR 5001.402(b)(3), interest must be paid at least annually from the first year, and "the first full principal and interest payment must commence not more than 3 years from the date of the promissory note." For a rural project that covers only after stabilization, that is the most useful sentence in the regulation. The third piece in this series, on B&I coverage under 7 CFR 5001, takes it apart.

Sizing a reserve is convention: average outstanding balance during the draw, times rate, times months to stabilization, plus a 10 to 20 percent cushion. A $3,000,000 construction draw at 50 percent average utilization, 7.50 percent and 12 months needs $112,500 before the cushion. We have not verified a paragraph in SOP 50 10 8.1 that names a funded interest reserve as an eligible 7(a) use of proceeds. Lenders who build one into a 7(a) construction loan do so inside the eligible construction use, and should confirm against the SOP text.

What makes an underwriter reject it: an interest-only period used to mask a permanent shortfall, or a reserve sized for fewer months than a realistic lease-up. Interest-only coverage is a liquidity test. Repayment is tested on the amortizing payment.

Lever 6. The capital stack: 504 is not always the bigger gain

The standard advice is to put owner-occupied real estate in a 504. For a pure real estate project, the advice is right. In an acquisition that mixes real estate with goodwill, it needs arithmetic first.

The 504 structure is rule. The third-party lender's first lien must be at least as large as the CDC loan, and is typically 50 percent of project cost. The CDC second lien is generally capped at 40 percent, and the borrower contributes 10 to 20 percent. The CDC piece is fixed for 10, 20 or 25 years. October 2026 effective rates were 6.97 percent on 20- and 25-year debentures, priced October 8, with some CDCs quoting slightly higher on different fee assumptions. Since July 4, 2026, a borrower may reportedly hold up to $5 million of 7(a) and $5 million of 504 at the same time. A 504 cannot be blended into the 7(a) maturity.

Run the case. The building goes into a 504: a $500,000 bank first lien, assumed at 7.50 percent fixed over 25 years, a $400,000 CDC loan at 6.97 percent over 25 years, and $100,000 of equity. The business portion goes into a $1,350,000 7(a) loan on ten years.

Structure	Annual debt service	DSCR today	DSCR at +200 bp on variable debt	Total payments over life
One 7(a) loan, 16-year blend	$282,394	1.35x	1.20x	$4.52 million
504 for the building, 7(a) for the business	$292,257	1.30x	1.22x	$4.10 million

The single blended 7(a) wins on day-one coverage, because the 504 pulls the real estate out of the blend and leaves the goodwill on its ten-year clock. The 504 stack wins on everything after day one. Its fixed debenture holds up better under a rate shock, and it costs about $420,000 less in total payments. Which structure is right depends on whether the file's problem is clearing 1.25x at closing or surviving the next rate cycle. The stress testing piece in this series takes up the second question.

Two other stack rules matter. In USDA B&I, the guaranteed and unguaranteed portions must be secured by "the same collateral with equal lien priority." And mezzanine or subordinate debt raises senior coverage while lowering total coverage. In an SBA deal it helps only on full standby. Otherwise its payments enter the denominator.

What makes an underwriter reject it: a third-party loan with a demand or early-call feature, which 13 CFR 120.921 prohibits outside default, or current-pay junior debt presented as equity.

Lever 7. Rate: the cap is a ceiling, not a price

The 7(a) variable maximum for a loan over $350,000 is the base rate plus 3.00 percent under 13 CFR 120.214(d). At 7.00 percent Prime, effective September 17, 2026, that is 10.00 percent. Fixed-rate maximums, set in 2022, run Prime plus 5.00 percent for the same loan size, or 12.00 percent. Since March 1, 2026, lenders may also price 7(a) loans off SOFR or the five- or ten-year Treasury. However, "the maximum interest rate that can be charged by the Lender shall not exceed Prime plus the allowed spread for that loan amount." The cap binds on the all-in rate, not on the spread.

That creates room. SOFR was 3.95 percent on October 30, 2026. A lender pricing off SOFR can carry a spread well above 3.00 percent and still land below the Prime-based ceiling. Most term sheets we see are drafted at the cap. Few need to be.

Rate is still a weak lever for a large gap. In our case, each 100 basis points is worth about $14,800 a year, or 0.05x of coverage. To clear 1.25x on the drafted ten-year term, the rate would have to fall to 6.31 percent. Rate concessions close gaps of a few hundredths, not a few tenths.

Under USDA B&I, rates are negotiated but may not exceed what the lender "customarily charges its borrowers for non-guaranteed loans in similar circumstances." Variable rates may adjust no more often than quarterly, and if a swap is used the guarantee covers principal and interest only. In the 504, the debenture rate is fixed at the monthly sale.

Buydowns are convention, not rule. A permanent buydown paid from verified, unborrowed funds is a legitimate cost of capital. A temporary buydown that subsidizes the first two years should not be credited in underwriting coverage. Test at the note rate after the subsidy ends.

What makes an underwriter reject it: an alternative-base 7(a) rate that drifts above the Prime-based cap at a reset, or year-one coverage manufactured by a temporary buydown.

Lever 8. The numerator: fix it honestly or not at all

Every lever so far shrinks the denominator. The eighth asks whether the numerator is right. Sometimes cash flow is understated, and correcting it is legitimate. More often the correction appears only after the shortfall does, and that sequence is the signal underwriters watch for.

SBA defines operating cash flow as EBITDA, and debt service as all future principal and interest on business debt, including the new loan. Appendix 15 allows "prudent adjustments" for savings the transaction will realize, including seller discretionary expenses and owner compensation, and requires the justification for each add-back in the credit memorandum. An owner-compensation adjustment must also be supported by a global cash flow analysis showing the principals can live on the adjusted salary and still meet 1.00x global coverage. USDA B&I defines DSCR more strictly, deducting reasonably expected replacement capital expenditures from EBITDA. We take that definition up in the B&I piece in this series.

Item	Treatment	Status
Depreciation, amortization, interest on refinanced debt	Add back	Rule
Rent to a related party, where the loan buys the property	Add back; replaced by the new debt service	Rule
Owner compensation above a market replacement salary	Add back the excess only; must survive 1.00x global	Rule; the salary benchmark is convention
Documented non-recurring expense	Add back with evidence in the credit memo	Convention, with documentation required by rule
Projected synergies, or growth under the new owner	Not allowed on a change of ownership	Rule
A "one-time" expense that recurred in each of the last three years	Reject	Convention

In our case, suppose the seller paid himself $150,000 and the buyer will take a market salary of $110,000. The $40,000 add-back lifts coverage on the drafted term from 1.07x to 1.18x. It is real money. It is not enough on its own, and it holds only if the buyer's personal budget shows the household can live on $110,000.

Refinancing is the numerator's companion lever: retire expensive debt so the same cash flow serves less. A 7(a) refinance must cut the installment on the debt being refinanced by at least 10 percent, and that debt must have been current for 12 months. For 504, Direct Final Rule 89 FR 79734, effective October 1, 2024, removed the substantial-benefit test for refinancing without expansion. In both programs, a refinance may not shift a lender's existing loss onto SBA.

What makes an underwriter reject it: add-backs that appear only after the shortfall, unreported cash sales, and owner draws normalized away when the global analysis shows the household needs them. A Quality of Earnings report, required at a $3,000,000 purchase price, exists to catch all three.

Here is the case under each lever applied on its own.

Lever	Change to the case	Annual debt service	DSCR	Clears 1.25x
As drafted	One 7(a) loan, 10 years, 10.00 percent	$356,807	1.07x	No
1. Amortization	16-year blended maturity	$282,394	1.35x	Yes
2. Resizing	Loan cut to $1,917,000	$304,000	1.25x	Yes, if the gap is funded
3. Equity	Equity raised to $583,000, or 23 percent	$304,000	1.25x	Yes
4. Seller paper	$250,000 on full standby	$317,162	1.20x	No
5. Interest-only	Interest-only at 10.00 percent	Not credited in the test	1.07x as tested	No
6. Capital stack	504 for the building, 7(a) for the business	$292,257	1.30x	Yes
7. Rate	9.00 percent instead of 10.00 percent	$342,025	1.11x	No
8. Numerator	$40,000 owner-compensation add-back	$356,807	1.18x	No

Four levers clear the floor alone, and two of them, the blended maturity and the 504 split, were available to the borrower before anyone asked for another dollar. That is the pattern we see most often. The file did not fail on earnings. It failed on a term sheet drafted before anyone ran the blended maturity.

Our sequence for any change of ownership tested under Appendix 15 runs as follows. First, fix the numerator to what the tax returns and, where required, the QoE support; this often moves it down rather than up. Second, take the longest amortization the rule allows, comparing the single blended 7(a) against a 504 split on both day-one coverage and a rate shock. Third, resize to the coverage-constrained amount. Fourth, fill the gap with unlimited-category equity. Only then reach for full-standby seller paper and rate. Interest-only belongs to startups, construction and B&I ramp-ups, not to acquisitions.

The order matters because the early levers are durable and the late ones are fragile. An amortization schedule does not get renegotiated in year three. A seller who agreed to wait ten years sometimes does.

What this means if you are building the file

Run the floor before the term sheet. Identify the transaction category under Appendix 15 and the SOP version that governs the file, then compute coverage on the basis the rule requires: historical for acquisitions, projected within two years for startups and construction. A term sheet drafted before that step is a guess.
Show the structures side by side. For any acquisition that includes real estate, put the single blended 7(a), the two-loan split and the 504 stack in one table, with day-one coverage, coverage under a 200 basis point shock and lifetime cost. Credit committees decide faster when they see the trade-off than when they are sold one answer.
Size the loan from coverage, not from price. State the coverage-constrained loan amount in the credit memo, and show where the difference between that amount and the request is coming from.
Document every add-back where it is made. The credit memo is where SBA will look at guaranty purchase. An add-back without support is an add-back that did not happen.
Treat seller paper as a late lever. If the note is not on full standby, model it at the imputed ten-year amortization from the first draft.
Keep interest-only out of the repayment test. Show it as liquidity, next to the reserve that funds it.

An independent feasibility study does most of this work before the lender sees the file. It establishes the cash flow the business can actually support, tests it against the rule that applies, and shows which structure clears it and by how much. That is the work behind our SBA and USDA feasibility studies.

Current as of November 3, 2026

This analysis reflects SOP 50 10 8.1 with Technical Policy Updates, effective October 1, 2026 under Information Notice 5000-882227. That notice superseded the August 14, 2026 version, which never took effect. The two notices describe applicability differently. The August notice applied the SOP to loans issued an SBA loan number on or after October 1. The September notice, as reported, applies it to applications received by SBA on or after that date. For a file that straddles the date, confirm with the lender which version governs.

SBA publishes SOP 50 10 8.1 as a Word document. Our SOP quotations follow practitioner readings of that document where we have not yet conformed the wording ourselves. CFR quotations were checked against eCFR. Practitioner sources disagree on the 504 floor under SOP 50 10 8.1, and we treat 1.15x as reported until the SOP text is conformed.

Rates are as of October 30, 2026: WSJ Prime 7.00 percent and SOFR 3.95 percent. 504 effective rates are as priced on October 8, 2026, and the next 504 pricing is November 5, 2026. All worked examples are illustrative and use the assumptions stated in the text. We refresh this piece whenever SBA issues a notice touching Appendix 15 or the 504 credit standards, and at least quarterly for rates.

Questions lenders and borrowers actually ask

What is the minimum DSCR for an SBA 7(a) business acquisition in 2026? Since October 1, 2026, it is 1.25x for Initial Acquisitions, Owner Buyouts and ESOP or cooperative purchases, and 1.15x for Business Expansions. Coverage is measured on the last fiscal year or a two-year average, historical or adjusted, and projections may not be relied on.

Can a seller note fix a DSCR shortfall on an SBA loan? Only if it is on full standby, with no principal or interest paid, for the life of the SBA loan. A seller note that is not on full standby counts in debt service, and an interest-only note is underwritten as if it amortized over no more than ten years.

Does an interest-only period help an SBA acquisition meet DSCR? No. Appendix 15 tests historical coverage against the debt service the business will carry. Interest-only periods matter for startups, construction projects and USDA B&I loans, where the first full payment may come up to three years after the note.

Is an SBA 504 always better than a 7(a) for DSCR? Not on day one. In an acquisition with goodwill and real estate, a single 7(a) on the blended maturity can show higher initial coverage than a 504 for the building plus a ten-year 7(a) for the business. The 504 usually wins on rate resilience and lifetime cost.

Do bank regulators set a minimum DSCR? No. The Interagency Guidelines, the OCC and the FDIC require each bank to set its own minimum standards for debt service coverage, and deliberately state no number. The 1.20x to 1.35x range is convention.

Is a DSCR below 1.00x an automatic decline? No primary rule says so. A file below 1.00x fails every SBA floor as structured, and SOP 50 10 8.1 names additional equity to reduce the loan as the cure.

What is the longest amortization available on a USDA B&I loan? The justified useful life of the project or collateral, up to 40 years, with no balloon payments outside servicing.

November 2026, by Michal Mohelsky, principal of MMCG Invest, LLC, a national SBA and USDA feasibility study consultancy.

Sources:

13 CFR 120.150, What are SBA's lending criteria?
U.S. Small Business Administration, SOP 50 10 8, Lender and Development Company Loan Programs. Effective June 1, 2025.
U.S. Small Business Administration, SOP 50 10 8.1 with Technical Policy Updates, including Section B, Chapters 1 and 2; Section C, Chapter 1; Appendices 14 and 15. Effective October 1, 2026.
SBA Information Notice 5000-880695, Issuance of SOP 50 10 8.1. August 14, 2026.
SBA Information Notice 5000-882227, Issuance of Technical Updates to SOP 50 10 8.1. September 25, 2026. As reported by NAGGL.
SBA Procedural Notices 5000-875701 and 5000-876777, 7(a) Small Loan underwriting. Effective March 1, 2026. Secondary summary by Starfield & Smith.
13 CFR 120.212, loan maturities.
13 CFR 120.213 and 120.214, fixed and variable interest rates.
13 CFR 120.882 and 120.890, 504 eligible project costs and interim financing; Direct Final Rule, 89 FR 79734, effective October 1, 2024.
13 CFR 120.910, 120.920, 120.921 and 120.930, 504 borrower contribution, third-party loans and loan limits.
7 CFR Part 5001, OneRD Guaranteed Loan Program, including 5001.105, 5001.202, 5001.401, 5001.402 and 5001.450; as amended 89 FR 79725, September 30, 2024.
SBA, alternative base rates for 7(a) loans, Federal Register Doc. 2026-02660, February 10, 2026; Procedural Notice 5000-875051, effective March 1, 2026.
SBA, maximum allowable fixed interest rates for 7(a) loans, 87 FR 46883, August 1, 2022.
Office of the Comptroller of the Currency, Comptroller's Handbook, Commercial Real Estate Lending, Version 2.0, March 2022.
Interagency Guidelines for Real Estate Lending Policies, 12 CFR Part 365, Appendix A, and OCC and Federal Reserve equivalents.
FDIC Risk Management Manual of Examination Policies, Section 3.2, Loans.
Office of Thrift Supervision, Examination Handbook Section 210, Appendix A, Income Property Lending. Superseded; historical illustration only.
Freddie Mac Multifamily, Optigo Conventional Fixed-Rate term sheet.
Bankrate, WSJ Prime Rate table, updated October 30, 2026. Secondary.
Federal Reserve Bank of New York, Secured Overnight Financing Rate, via FRED, October 30, 2026.
Growth Corp and SomerCor, SBA 504 effective rates for October 2026, priced October 8, 2026. Secondary.
Pioneer Capital Advisory, SOP 50 10 8 vs SOP 50 10 8.1: Annotated Comparison, September 2026. Secondary.
Business Plan Firm, SBA Debt Service Coverage Ratio Requirements Under SOP 50 10 8.1, updated October 12, 2026. Secondary.