SBA 7(a) Method: What 13 CFR 120 Tests and How a Feasibility Study Answers Each Test

A 7(a) loan is a bank's or another participating lender's loan that carries a federal guarantee, and the guarantee arrives with conditions written into regulation. SBA's own lender page says that its eligibility and credit standards begin with the requirements in 13 CFR part 120 ^[1]. This article walks that part, with the statute behind it in 15 U.S.C. 636(a), one test at a time. For each test it asks two questions. What evidence can a feasibility study place in the lender's file? And what can the study never decide? The second question deserves as much attention as the first. A study can show that demand exists, that an operating projection hangs together and that cash flow carries the debt under stated conventions. It cannot make a business eligible, certify that credit is unavailable elsewhere, or approve anything.

The regulation is not the whole rulebook, and it says so. Part 120 defines Loan Program Requirements to include, beyond statute and regulation, SBA's standard operating procedures, Federal Register notices and official SBA notices and forms ^[2], and it obliges every SBA lender to comply with those requirements and keep current with them as they are revised ^[3]. This article stays with text published as law or as agency guidance on SBA's own pages. Where the site's SBA program page states a coverage floor, the sentence says that the floor is the site's statement.

Two archived posts on this site cover neighbouring ground. The site's post SBA Feasibility Study Requirements: What SOP 50 10 8 and 8.1 Say, What Lenders Order Anyway, and How Much Loan Volume Runs Through a Study covers the agency's guidance and lender practice, and the site's SBA program page reads that guidance as setting 1.15x business coverage for standard 7(a) loans ^[4]. The other post, The Importance of Feasibility Studies for SBA 7(a) and 504 Loan Programs, explains why studies matter in both programs. This piece is narrower and more literal: it reads the regulation section by section and pairs each section with the evidence that answers it. Its companion on the 504 program does the same for subpart H.

Where 13 CFR 120 Sits in a 7(a) Decision

SBA describes 7(a) as its primary business loan program for small businesses ^[5], yet the borrower deals with a lender, not with the agency. Its 7(a) page tells applicants that they will work directly with their lender and not with SBA ^[6]. The statute frames the arrangement the same way. SBA may lend on its own account, or it may join banks and other institutions through participation agreements, and those agreements come in two kinds, immediate and deferred, the deferred kind being a guarantee ^[7]. In a 7(a) loan SBA takes the deferred kind, and the credit analysis is the lender's ^[8].

The regulation then sets out who answers for what. Among its conditions of participation, a lender must itself be in safe and sound condition, and the lending policies, procedures and standards it uses must be commercially reasonable, of the kind prudent lenders employ ^[9]. Lenders and their contractors are independent contractors, responsible for their own actions on a 7(a) loan ^[10]. Under the Preferred Lenders Program, the delegated lender carries every decision on eligibility, size included, and on creditworthiness ^[11]. Even so, SBA's lender page says that SBA makes the final determination of the applicant's eligibility on the guaranty requests lenders submit ^[12].

A feasibility study enters this structure through one sentence. Section 120.160 lists the conditions SBA normally imposes on business loans ^[13], and its paragraph (b) is where the study appears: SBA may make one a condition of the loan, in the same breath as professional appraisals and a survey ^[14]. That sentence matters for two reasons. It makes the study a recognised condition of a loan, not a marketing extra. And it names the study as one instrument and the appraisal as another, a useful reminder that the two do different jobs and that a study does not value property.

Most of the rest of part 120 is a set of tests applied by the lender, and in some cases by SBA, to evidence drawn from the application ^[15] and from other reports; where the project depends on a market that has not yet proven itself, one of those reports may be a study ^[14]. Table 1 maps the tests this article covers against the evidence a study can supply and the party that decides each one. Read across a row: the third column is the study's contribution, and the fourth column is a reminder of whose decision it supports.

Eligibility: Facts a Study Can Document, a Status It Cannot Confer

The first gate is the shortest to state. To be eligible, an applicant must be an operating business (an Eligible Passive Company is the exception), organised for profit and located in the United States ^[16]. It must be small under the size rules in part 121, counting its affiliates ^[37], and it must be able to demonstrate a need for the credit it seeks ^[38]. SBA's 7(a) page puts the same idea in plainer words: the key eligibility factors are what the business does to earn its income, its credit history and where it operates ^[39].

Section 120.110 then lists businesses that cannot borrow at all. The list opens with non-profit businesses, though their for-profit subsidiaries may qualify, and with financial businesses primarily engaged in lending ^[40]. It excludes the passive holding business of a landlord or developer that will not itself operate from the financed property, unless the borrower qualifies as an Eligible Passive Company ^[17]. Further down, among other types, it bars a business that derives more than one-third of its gross annual revenue from legal gambling ^[41], and it bars speculative businesses, giving oil wildcatting as its example ^[42].

None of these is a market question, and a study does not answer any of them. But several of them turn on facts that a well-built study already collects, and a lender reading the study should be able to find those facts without hunting.

The operating test is one. A study describes the business the loan will finance: what it sells, to whom, through what facilities and with what staff. That description is the factual basis for saying the borrower is an operating concern and not a landlord holding a building for rent ^[17]. If the real estate sits in an Eligible Passive Company that rents it to an operating company, the study should model the tenant business's cash flow, because that is the business whose earnings repay the loan.

Revenue composition is another. A study's revenue build shows each line of income separately. For most projects that is simply good practice. For a business with any gaming revenue at all, it is also the schedule from which a lender can see how much of the total legal gambling contributes, which is the number the one-third bar turns on ^[41]. The study does not rule on that share; it makes the share visible.

Speculation is a third. The regulation does not define a speculative business beyond its example ^[42]. A study cannot classify a borrower, but it can show whether the revenue case rests on observed demand in a defined trade area or on a hope that demand will appear. That evidence is what a study can put in front of a lender weighing whether a novel concept is a going business or a bet.

Need for credit is the last, and it is the closest to the study's ordinary work. A sources and uses schedule shows what the project costs, what the borrower brings and what is left to finance ^[38]. That gap is the need. A study that sizes the project from a sourced cost estimate gives the lender a defensible number for it.

What the study must not do is assume eligibility and carry on. Size is measured under part 121 with affiliates counted ^[37], and that is a calculation for the lender and SBA, not for a market analyst. The final call on eligibility belongs to the delegated lender or to SBA ^{[11] [12]}. A study whose conclusion declares the borrower eligible has stepped outside its lane. A study that states which business the borrower operates, and which space it will occupy, has given the lender something it can use.

Credit Elsewhere: The Test Behind Every Guarantee

The statute's first restriction on 7(a) lending is blunt: no financial assistance may be extended if the applicant can obtain credit elsewhere ^[19]. The definition that gives the phrase content sits in 15 U.S.C. 632(h). Credit elsewhere means credit available to the individual applicant on reasonable terms and conditions from non-government sources, judged by factors associated with conventional lending practices ^[43]. The factors named are the applicant's industry, whether it has operated for two years or less, the adequacy of collateral ^[44] and the loan term needed to reasonably assure repayment from actual or projected cash flow ^[45]. A residual factor covers anything else about the application, documented in detail by the lender, that cannot be overcome except with a federal guarantee under prudent lending standards ^[46].

The regulation repeats the test in lender language. The lender has to certify, or show in some other way, that the applicant cannot get the credit it wants on reasonable terms without SBA's help ^[18]. The factors are the same: industry, operation for two years or less, adequacy of collateral ^[47] and the term needed for repayment from actual or projected business cash flow ^[48]. Submitting the application is itself the certification that the lender examined the availability of credit ^[20], and the lender must keep substantiation for that certification in its file ^[49]. One reason is closed off by statute: SBA may not guarantee a loan when the borrower lacks other credit only because the lender's own liquidity depends on selling the guaranteed portion ^[50].

This is the test a study answers least directly and supports most broadly. Take the factors one at a time.

Industry is a factual question about risk. A study documents the trade area, the competition and the demand the project will draw on, and it can set the project against the loss history of the asset class. The loan file shows that classes differ widely; the data section below prints the charge-off floors for three classes this site's worked examples use. A lender that cites industry as a reason conventional terms are unavailable needs that evidence in the file ^[49].

Operating history is a matter of record. A new business has none, and the regulation names operation for two years or less as a factor in its own right ^[47]. For a start-up, a study should expect to carry most of the weight: for most business loans the application calls for historical statements or tax returns for the past three years ^[51], and a new business has none of its own, so the projection has to stand on outside evidence.

Collateral is the appraisal's subject, not the study's. A feasibility study does not value property and should not offer an opinion of value. What it can do is describe the property's use, its fit to the business and the business's dependence on it, so that a lender reading a separate appraisal understands why a single-purpose building may fall short as security ^[47].

Term is where the study's projection does direct work. The statute and the regulation both name the term needed to repay from projected cash flow as a factor ^{[45] [48]}. A study should therefore test the projection both at the term a conventional offer would carry, where the borrower has one, and at the longest term the regulation allows for the assets financed, because the gap between the two is what this factor is about. The sizing table later in this article shows how far coverage moves between a ten-year and a 25-year term on the same cash flow; that comparison, set out in a study, is the kind of evidence a lender can place behind a credit elsewhere certification.

Market conditions matter too, and they change. Banks answering the Federal Reserve's survey of senior loan officers, published in October 2026, said that over the third quarter their standards for commercial and industrial loans to large and middle-market firms were basically unchanged, while standards for such loans to small firms tightened ^[52]. The survey counts a firm with annual sales under $50 million as small ^[53]. Banks also reported tightening, on net, across every commercial real estate loan category ^[54]. A study may cite such evidence for context. It should not draw the lender's conclusion.

There is a trap on the other side. A study written as advocacy can describe a project as so strong that a reader wonders why it needs a guarantee at all. A study is evidence, not a brief for either test: it should report demand, costs and coverage as they are, and leave the certification to the lender whose name is on it ^[20].

Use of Proceeds and Occupancy: The Project Must Fit the Program

A 7(a) loan must be used for sound business purposes ^[21]. Within that frame, proceeds may buy land, existing buildings and their renovation, new construction and fixed assets ^[55]. A 7(a) loan can also carry inventory, supplies, raw materials and working capital ^[56]. Its proceeds may also refinance certain existing debts ^[57] and fund a change of ownership, whether of one owner's stake or of the whole business ^[58]. The regulation also says what proceeds may not do. They may not finance real or personal property held primarily for sale, lease or investment, outside the stated exceptions ^[22], and, among the other restrictions, they may not serve a purpose that does not benefit the small business ^[59].

The study's contribution here is a disciplined sources and uses schedule. Each line should name what it buys, its source in the cost estimate and the category of use it falls into. A study should lay the schedule out so that it can be read line by line beside the regulation. A vague line (a large soft cost figure, an unexplained reserve, a payment to a related party) is a question the study should have answered. A study that ties each dollar of the request to an identified cost leaves no such question open and exposes any line that might not qualify, before the file reaches SBA.

Real estate brings an occupancy test. For a newly constructed building financed by SBA, the borrower has to occupy and use, permanently, no less than 60 percent of the building's rentable space, and may lease out at most 20 percent on a permanent basis ^[23]. The rule also looks forward: some of the space that is neither occupied at once nor permanently leased has to be planned for the borrower's own use inside three years, and the rest inside ten ^[60]. An existing building bought or renovated with SBA money carries a lower bar: the borrower occupies at least 51 percent and may lease up to 49 percent ^[24].

These are measurements, and a study is often where they are first made against a plan. The space program should state the rentable area, the area the operating business will use, the area offered to tenants and any area held for growth, and it should show the arithmetic against the applicable threshold ^[23]. If the projection counts rent from a tenant, the study should say so plainly and show that the leased share stays within the limit, because rental income from more space than the regulation allows is not a cash flow the loan can rely on.

Construction adds a further requirement. A 7(a) loan that pays for construction obliges the borrower, unless SBA waives it, to post a payment and performance bond covering the whole contract and to carry builder's risk insurance ^[61]. A study does not obtain bonds, but its development budget should carry their cost, and its opening date should allow for the time a bonded contractor needs.

The three 7(a) worked examples on this site show how the shape of a project changes the work. Worked Example: SBA 7(a) Gas Station Feasibility Study in Tennessee is ground-up construction, so the budget, the bond and the opening ramp all matter. Worked Example: SBA 7(a) Assisted Living Feasibility Study in Utah is the purchase of an operating facility, a change of ownership under 120.202 ^[58], so the seller's earnings come first. Worked Example: SBA 7(a) Gym Feasibility Study in Hawaii finances leasehold improvements, equipment and working capital in leased space, so no real estate is bought and the occupancy test does not arise, but the term does.

Reasonable Assurance of Repayment: The Test a Projection Must Carry

The statute requires every 7(a) loan to be of such sound value or so secured as reasonably to assure repayment ^[26]. The regulation puts its credit standard in two short sentences: the applicant has to be creditworthy, and each loan has to be sound enough that repayment is reasonably assured ^[25]. It then tells lenders how to judge that. They must use prudent, generally accepted commercial credit analysis consistent with what they use on their similarly sized loans without a guarantee ^[8]. In approving a loan, lenders, CDCs and SBA "may consider (as applicable) the following criteria: credit score or credit history of the applicant (and the Operating Company, if applicable), its Associates and any guarantors; the earnings or cashflow of applicant; or where applicable any equity or collateral of the applicant" ^[62]. SBA's 7(a) page adds the practical detail that most 7(a) term loans are repaid in monthly payments of principal and interest from the cash flow of the business ^[63].

Cash flow is where a feasibility study earns its place. Where a business has a record, a study should start from that history. For a start-up, an expansion or a project whose revenue depends on a market the borrower has not yet served, the study's projection is the case, and it is only as good as the demand evidence beneath it. The study's chain runs in order: the trade area and its demand, the share the project can capture against named competition, the revenue that share produces, the operating costs to deliver it, and the cash flow left for debt service. Each link should rest on a cited source or a labelled assumption.

Conventions. This article follows the conventions of the Loan Analytics worked examples. Coverage is tested against the program floor the site's program page states, where the program has one, and against a stated 1.25x lender assumption, and the higher of the two binds. Equity is measured at the loan the cash flow supports, with fees recomputed on that loan. An acquisition is tested on the seller's historical earnings, not on projections.

The floor itself is not in part 120. The site's SBA program page says that the agency's current procedures, SOP 50 10 8.1 in the page's words, put standard 7(a) loans at a minimum of 1.15x business coverage, with a start-up allowed two years from funding to get there ^[4]. The site's page then splits changes of ownership by type. For an initial acquisition, an owner buyout or an ESOP, the site's page asks for 1.25x on historical earnings, the same multiple the worked examples adopt as their lender assumption, while a business expansion needs 1.15x ^[64]. For those first three types, the site's page adds, a projection cannot satisfy the floor ^[65], whereas expansions, like start-ups and ground-up construction, are underwritten on projections that must be independently supported, in the site's words ^[66]. On the site's methodology page, the multiple a study tests against is whatever the program and the lender require, generally somewhere from 1.15x up to 1.25x, the upper end being the lender assumption this article uses ^[67]. Each of these is the site's statement of a floor; this article does not restate the agency's procedures beyond them.

Table 2 lists the inputs of one assumed request, and Table 3 tests it five ways. The project costs an assumed $5,600,000 before fees, the request is set at the program maximum of $5,000,000 ^[27], and the stabilized cash flow available for debt service is an assumed $620,000. The rate is prime at 7.00 percent ^[68] plus an assumed lender spread, and the term is 25 years because the example finances real estate ^[31].

Rounded as shown; lines may not foot by one unit in the last place.

The base case shows why the conventions line tests coverage twice. On the full request, coverage is 1.16x. That clears the 1.15x floor the site's page states ^[4] but not the 1.25x lender assumption, so the lender assumption binds and the loan the cash flow supports is $4,638,267. Equity is then measured at that loan, with the guarantee fee recomputed on it, and comes to $1,089,684, or 19.0 percent of cost plus fee.

The second column moves the spread to the regulatory ceiling for a loan above $350,000 ^[32]. The rate rises to 10.00 percent, coverage on the full request falls to 1.14x, below the floor, and the supported loan shrinks again. The third column cuts cash flow by the assumed shortfall. Coverage drops to 1.04x, and the borrower would need $1,540,466 of equity to close the gap at the binding coverage. A study that shows only the base case leaves the lender to build these columns alone.

The fourth column keeps everything but the term, which falls to ten years, the ceiling when the loan buys neither real estate nor equipment that lasts beyond a decade ^[30]. Coverage falls to 0.79x. The same cash flow supports only $3,160,760 on that term. This is why the term question appears inside the credit elsewhere test ^[45] and why a study must state which assets the loan finances: the answer drives the amortization, and the amortization drives the loan.

The fifth column treats the same request as an initial acquisition of an existing business, one of the three types for which the site's page sets the historical test ^[64]. The test moves to the seller's last fiscal year, an assumed $575,000, and the floor the site's page states for an initial acquisition ^[64], which happens to equal the lender assumption of the conventions. Coverage on the request is 1.08x and the supported loan is $4,301,619. For a deal of those types, a projection of better results under the new owner does not rescue the test on the site's stated reading ^[65]; a study's job shifts to testing the seller's history and the add-backs, and its projection becomes a forward view of risk rather than the basis for sizing.

Two further parts of the regulation bear on the same question. Holders of at least 20 percent of the business generally must guarantee the loan personally ^[36], which puts the owners' resources behind the projection. And for most business loans the application itself must contain, among other things, a description of the history and nature of the business, the amount and purpose of the loan, the collateral and current financial statements ^[15], with historical statements or tax returns for the past three years, IRS tax verification and a business plan when applicable ^[51]. A study does not replace the business plan. It tests the plan's revenue and cost assumptions against outside evidence and reports where they hold and where they do not.

What the study never does is make the credit decision. The lender's analysis governs ^[8], and the delegated lender answers for creditworthiness ^[11]. A study that reports 1.16x coverage has reported a number under its stated conventions; whether that number, with the guarantors, the collateral and the credit history, is enough is the lender's call.

Loan Size, Guarantee Share and Term

The regulation sets the size cap in one line: "The maximum loan amount for any one 7(a) loan is $5,000,000." ^[27] A separate limit caps SBA's exposure to one borrower. Counting every loan to the borrower and its affiliates, SBA's portions "must not exceed a guaranty amount of $3,750,000, except as otherwise authorized by statute for a specific program" ^[73]. The statute states both limits in one clause ^[29]. A study bears on size through the cost basis and through the loan the cash flow supports, which may be well below either cap, as the sizing table shows.

The guarantee share follows the loan amount. Loans of $150,000 or less may carry a guarantee of up to 85 percent, and larger loans up to 75 percent unless law authorizes otherwise ^[28]; the statute sets the same shares against the balance outstanding at disbursement ^[74]. The share can be lower. The statute lets a lender ask for a reduced guarantee ^[75], and SBA's lender page describes SBA Express as trading a lower guaranty percentage for the use of the lender's own processes ^[76]. SBA's lender page defines Standard 7(a) loans as 7(a) loans above $350,000 other than 7(a) Small, SBA Express, Export Express, CAPLines, Export Working Capital Program, International Trade and pilot program loans ^[77]; the sizing example is a loan of that kind.

The loan file lets the guarantee rule be checked against the program as a whole. Table 4 divides the guaranteed dollars of FY2025 approvals by their gross dollars.

The program's guaranteed share, 73.76 percent, sits just below the 75 percent maximum for loans above $150,000 ^[28]. That blend is consistent with the rules: small loans can carry more than 75 percent, Express loans less, and any lender can request less ^{[75] [76]}. For a single project, the share matters because the upfront fee is charged on the guaranteed portion, not on the whole loan ^[71].

Term is the other structural choice, and the regulation states it as a principle before it states a limit. The first rule is that the term be the shortest one appropriate to what the borrower can repay. The second caps it at a decade, unless the money buys or refinances real estate, or equipment expected to last longer than that ^[30], and the same paragraph adds: "The term for a loan to finance equipment and/or leasehold improvements may include an additional reasonable period, not to exceed 12 months, when necessary to complete the installation of the equipment", a rule that covers finishing leasehold work as well ^[78]. The ceiling is 25 years including extensions ^[31], with added construction time allowed for a real property portion, and the statute sets the same 25-year bound ^[79]. A loan of 15 years or more also triggers the statute's prepayment fee rules ^[80]: a voluntary prepayment of more than 25 percent of the balance in one of the first three years after disbursement costs a recoupment fee of 5 percent in the first year, stepping down after that ^[81].

A study supports the term decision by separating costs. The lender needs to know how much of the request buys land and buildings, how much buys equipment and of what kind, and how much is working capital or a change of ownership payment. A cost schedule built that way lets the lender assign a term to each part and compute a blended maturity, and it lets the study test coverage at the debt service that blend produces rather than at a single assumed term. The gym example's ten-year term, and the gas station example's two terms, 25 years on the real estate portion and ten on equipment, inventory and working capital, both come from that allocation.

Rates and Fees the Projection Must Carry

A 7(a) loan may carry a reasonable fixed rate, with SBA publishing the maximum allowable rate in the Federal Register from time to time ^[35]. A variable rate floats over a base rate, which is prime or SBA's Optional Peg Rate ^[82], unless SBA has permitted one of the alternative base rates widely adopted in small business commercial lending, which it announces in the Federal Register ^[83]. The spread the lender may charge over that base falls as the loan grows. Table 5 sets the bands side by side with the guarantee shares and the statutory fee ceilings.

With prime at 7.00 percent ^[68], the variable rate ceiling on a loan above $350,000 is the 10.00 percent printed in the second column of the sizing table. SBA does not allow balloon payments ^[87], so a projection cannot assume a low payment followed by a refinancing; it must carry full amortization. And SBA's 7(a) page warns that on a variable rate loan the lender may require a different payment when the rate changes ^[88]. A study that tests coverage only at today's quoted rate has tested one point on a moving line. Testing at the ceiling spread shows the widest margin the regulation permits on that loan size; the base rate itself can still move.

Fees follow the statute. SBA collects an upfront guarantee fee on each loan of more than a year, payable by the lender and chargeable to the borrower, of not more than 2, 3 or 3.5 percent of the guaranteed share as the total loan rises through the $150,000 and $700,000 thresholds ^[71]. Above the highest band the statute adds "a guarantee fee equal to 0.25 percent of any portion of the deferred participation share that is more than $1,000,000" ^[72], so the added charge applies to no more than the excess of the guaranteed share over that amount. The regulation words the added fee differently, as 0.25 percent of the guaranteed portion of any loan whose total amount exceeds $1,000,000 ^[33]. The difference is modest in dollars, but it touches every loan above $1,000,000 ^[33], and for a loan just above that amount whose guaranteed portion is still below it, one reading charges the added fee and the other does not ^[72]. A study should compute the fee the lender quotes and show its basis, rather than adopt either reading silently.

The ceilings are ceilings. The statute's fee rule in 636(a)(23)(C) applies when SBA determines that guarantee fees paid by lenders and borrowers may be reduced, consistent with reducing to zero the cost to SBA of making the guarantees ^[89]. In that case SBA must first consider cutting the borrower fees as far as possible, and borrower fees may not be raised above the levels in effect on December 8, 2004 ^[90]. The annual service fee is a lender cost: it is capped at 0.55 percent a year on the guaranteed balance still outstanding, and the lender may not pass it to the borrower ^[34]. And if a loan of more than a year is never disbursed, SBA refunds the upfront fee when the lender asks in writing for the refund and the cancellation of the guarantee ^[91], a rule worth remembering when reading the cancellation figures below.

For the projection, the point is mechanical. The upfront fee is a use of funds, which the borrower may pay from loan proceeds, though not through a first disbursement made solely or primarily for it ^[92], and it changes with the loan. In the sizing table the fee is recomputed in every column at the loan the cash flow supports, which is why the fee falls when the supported loan falls. A study that holds the fee at the requested loan while sizing a smaller one overstates the equity required, or understates it, depending on which way the error runs.

Conditions SBA May Impose, Including the Study Itself

Section 120.160 opens by listing requirements SBA normally imposes on all business loans ^[13]. The first is personal guarantees, which holders of at least 20 percent of the business generally must give ^[36]. The second is the one this series is about, the power to call for a feasibility study, set in the same sentence as appraisals and a survey ^[14]. The third is hazard insurance, which SBA requires on all collateral once a 7(a) loan is greater than $500,000 ^[93].

Read closely, paragraph (b) settles less than it might seem to. It gives the power to SBA, while a lender, which must run its own prudent credit analysis ^[8], can ask for a study under its own credit policy. It names a study without defining one, so nothing in part 120 says what a study must contain. That silence is why this article proposes judging a study by whether it answers the tests in the rest of the part.

A study fit for that condition has a few recognisable features. It is independent of the outcome: the analyst has no stake in whether the loan closes. Its assumptions trace to sources a reviewer can check, and the ones that cannot be sourced are labelled as assumptions. It states its coverage conventions, as the conventions line above does, so that a credit officer can see which floor was used and why. It tests the case against the variables that move it: rate, term, volume and, for an acquisition, the seller's verified history. And it stops where its evidence stops. A study can say that the projected cash flow covers the debt at a stated multiple; it cannot say that the borrower is creditworthy, that the collateral is adequate, or that SBA will guarantee the loan.

What can go wrong is usually a mismatch between the study and the file. The cost estimate in the study differs from the one in the application. The study models a 25-year term on a request that is mostly equipment. The revenue case includes rent from more space than the occupancy test allows ^[24]. The coverage is tested against a floor that does not apply to the transaction type, such as a projection offered for an initial acquisition, an owner buyout or an ESOP, where the site's SBA program page states that the floor cannot be met on projections ^[65]. Each of these is avoidable when the study is built against the regulation from the start rather than fitted to it afterwards.

What the Loan File Shows About the Tests

The SBA's FOIA loan files, as compiled into this site's dataset, show the program the regulation governs. SBA's 7(a) approvals dated in FY2025 number 78,078 ^[94], worth $37.29 billion ^[69] gross, a count that keeps approvals later cancelled, following SBA's own convention. By September 30, 2026, 14,124 ^[95] of those FY2025 approvals had been cancelled before any disbursement. That count is young: the cohort has had little time to season, and more cancellations may follow.

Across the FY2017 to FY2024 approval cohort, the cancellation rate is 11.24 percent ^[96]. Like every rate in this dataset, it is an approval-cohort floor, because approvals still undisbursed may yet cancel. A cancelled approval is not a loss; it is a loan that never funded, and the fee refund rule above can reach it ^[91]. A companion piece treats that distinction at length.

Losses are measured on the FY2017 to FY2021 cohort, loans now five to ten years old. To date, 6.33 percent ^[97] of the cohort's disbursed loans have been charged off. Read that as a floor on this approval cohort: open loans keep their status confidential, and the cohort is still accruing charge-offs. Holding the horizon fixed, 3.72 percent ^[98] of the cohort's disbursed loans had reached charge-off inside five years from approval, a measure that compares cleanly across cohorts. Why a published rate of this kind can only understate gets its own article in this series.

The asset classes behind this site's three 7(a) worked examples sit far apart. On the same cohort basis, and each as a floor, the charge-off rate to date is 1.44 percent ^[99] for assisted living, 1.16 percent ^[100] for gas stations and truck stops, and 9.48 percent ^[101] for fitness centers. Industry is one of the factors the credit elsewhere test names ^[47], and a spread of that size is the kind of evidence a lender may cite when it explains why conventional terms are not available on reasonable terms. It is not a forecast for any one project; a study's own market evidence has to carry the case for the project in hand.

Median loan size also places each class within the rules. The median 7(a) loan is $246,000 ^[101] for fitness centers, inside the band where the spread ceiling is six points ^[85] and the guarantee is capped at three quarters ^[28]. It is $752,450 ^[99] for assisted living, above the $700,000 threshold where the highest fee band begins ^[71]. For gas stations and truck stops it is $1,095,000 ^[100], a loan just above $1,000,000 whose guaranteed share at the maximum would still sit below $1,000,000, which is exactly where the statute's and the regulation's wording of the added fee part ways ^{[72] [33]}. Each class has a 7(a) data page carrying its full cell, with every figure's key and basis beside it, reachable from the asset-class index.

Loan Data Behind This Article

Every figure in this table comes from the Loan Analytics loan dataset, analytics-loan-data freeze-3 (commit 360c856), and carries its key. SBA rates are approval-cohort floors and are withheld when fewer than 30 loans stand behind them.

A Checklist for Lenders and Borrowers

A lender or borrower reading a 7(a) feasibility study can test it against the regulation with a short set of questions, each tied to the section it serves.

Does the study describe an operating business, its revenue lines and who occupies the property, so that the eligibility facts under 120.100 and 120.110 can be read from it ^{[16] [17]}? Does it show the revenue mix plainly enough that any gaming share is visible ^[41]?

Does it give the evidence a credit elsewhere certification can rest on: the industry's risk, the operating history and the term the projected cash flow needs ^[45]? Does it leave the certification itself to the lender ^[20]?

Does every line of the sources and uses schedule name an eligible use ^[21]? For real estate, does the space program show the occupied and leased shares against the applicable threshold ^{[23] [24]}?

Does the coverage test show both multiples, the floor from the site's SBA page and the lender assumption, and does the stricter of the two decide the loan ^[4]? For an initial acquisition, an owner buyout or an ESOP, is the test run on the seller's history, as the site's SBA page states ^[65]? And is the equity figure computed on the loan the cash flow can carry, with the guarantee fee worked out again on that smaller loan?

Does the cost schedule separate real estate, equipment and working capital, so that the term follows the assets ^[30]? Is coverage shown at the quoted rate and at the regulatory ceiling for the loan size ^[32], with no balloon assumed ^[87]?

Does the study stop at its evidence? It should report demand, the projection and coverage under stated conventions, and it should not opine on eligibility, creditworthiness or the value of collateral. Those belong to the lender and to SBA ^{[11] [12]}, and to the appraisal the regulation lists beside the study ^[14]. The data above, and the companion article on conventional bank lending, show what public evidence can add to that file, and where it runs out.

Related Pages on Loan Analytics

• The Loan Analytics methodology series (all articles)
• How a feasibility study is done: the Loan Analytics methodology
• SBA feasibility studies (7(a) and 504)
• Loan data by state
• Loan data by asset class
• Worked examples (case studies)
• Worked Example: SBA 7(a) Assisted Living Feasibility Study in Utah
• Worked Example: SBA 7(a) Gas Station Feasibility Study in Tennessee
• Worked Example: SBA 7(a) Gym Feasibility Study in Hawaii
• SBA Feasibility Study Requirements: What SOP 50 10 8 and 8.1 Say, What Lenders Order Anyway, and How Much Loan Volume Runs Through a Study
• The Importance of Feasibility Studies for SBA 7(a) and 504 Loan Programs
• SBA 504 Method: The Job Test, the Project Split and What a Study Must Show the CDC
• Cancellation Is Not Charge-Off: The Two Ways an SBA Approval Can End Without Repayment

Commission a Feasibility Study

Loan Analytics prepares lender-grade feasibility studies for SBA 7(a), SBA 504, USDA Business and Industry, conventional and EB-5 financing, built on the methods this series sets out. A study is not an appraisal: it tests whether a project's market, costs and cash flow can carry the debt a lender is asked to approve. To discuss a study for your project, contact Loan Analytics.

Daniel Smith, MAI, Loan Analytics