Global Cash Flow: Testing the Borrower, Its Owners and Its Affiliates as One Repayment Source

A small business loan is seldom repaid by one legal person acting alone. The company that signs the note may own nothing except a building. The company that earns the money may be that building's tenant. The people who own both may have guaranteed the loan, may carry mortgages of their own, and sometimes control a second business that borrows as well. Global cash flow is the discipline of adding all of that up once, and only once. The federal banking and credit union agencies put the idea in one line: institutions use global cash flow to assess the combined cash flow of a group of people or entities, so as to see the group's ability to service its debt ^[1]. Their definition of the matching ratio is just as short. Global debt service coverage takes in the cash flows of both the borrowers and the guarantors and sets against them the combined obligations of both, with contingent obligations counted ^[2].

This article explains how such an analysis is assembled for an SBA loan, where the program texts settle much of the structure before any spreadsheet is opened. SBA's business loan regulation requires the applicant, an Operating Company included, to be creditworthy, and it requires each loan to carry a reasonable assurance of being repaid ^[3]. It lists the credit history of the applicant's Associates and guarantors among the criteria a lender may weigh, next to the applicant's earnings or cash flow ^[4]. It provides that "Holders of at least a 20 percent ownership interest generally must guarantee the loan" ^[5]. And it lets a passive holding company borrow to acquire or improve property that it leases to an operating business, on terms that cap the rent in advance ^[6]. Each of those rules decides which entities belong inside the combined view and how the money that moves between them has to be treated.

How the coverage ratio itself is built, from cash flow available for debt service to the payment function and the floor that binds, is the subject of Debt Service Coverage in a Feasibility Study: How the Ratio Is Built, and Which Floor Binds in this series, and the 7(a) tests are taken one at a time in SBA 7(a) Method: What 13 CFR 120 Tests and How a Feasibility Study Answers Each Test. The question here is narrower and harder: how the figures of several entities and several people become one figure without any dollar being counted twice. The worked example later on follows the conventions of the site's worked examples.

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.

What a Combined View Adds to One Borrower's Coverage

A coverage ratio computed for one company answers one question: can this company's cash flow pay this company's debt? A global analysis answers a second question that anyone holding guarantees has to face: can the whole group meet everything it owes, the new loan included, after its members have paid their taxes and their household bills? The two answers can point in opposite directions. A project can clear its own test with room to spare while its owners are stretched by home mortgages and by guarantees of a sister company's loans. A project can also miss its own test while an owner holds enough surplus to carry it through a weak year. Neither result stands in for the other, and an analysis that reports only one of them hands the reader half a picture.

The supervisory texts treat the combined view as part of judging repayment. Among the factors examiners weigh when they look at a commercial borrower's ability to repay, the 2023 interagency policy statement lists the protection that cash flow from operations or collateral provides on a global basis, taking account of the total debt obligations of borrower and guarantor alike ^[7]. The same statement says the coverage analysis should include realistic projections of the borrower's available cash flow and an understanding of how continuous and how accessible each source of repayment is ^[8]. Its reach needs a careful reading. The statement was written to guide the modification of commercial real estate loans to borrowers who are, or may become, unable to meet their payments or fully repay ^[9], and it adds that certain of its principles also apply generally to commercial loans secured by real property or by other business assets ^[10]. It is workout guidance carrying general principles, not an origination rule for SBA loans, and this article uses it only for what it says about combined cash flow and guarantors.

The agencies' real estate lending guidelines point the same way for banks. Secondary sources of repayment and credit enhancements such as guarantees appear on their list of factors that sound real estate underwriting should reflect ^[11]. When the loan finances development, construction or a finished commercial building, the guidelines want the bank's policy, in proportion to how large the project is and what kind, to name a floor for net worth, for cash flow and for coverage, measured on the borrower or on the property ^[12], and to say what guarantor support it demands ^[13]. Those are expectations about a bank's own policy. They tell a reader that guarantor strength is something a bank is expected to measure against a standard of its own, and they leave the content of that standard to the bank.

For SBA loans, the site's own SBA program page reads the current agency guidance in the same direction: it states that, under the SOP, a start-up's projections must show business coverage of 1.15x by the end of the second year after funding, and that coverage of the owners' obligations, taken globally, must reach at least 1.0x ^[14]. That sentence is the site's reading, cited as the site's. The regulation's own lending criteria name earnings or cash flow without fixing any ratio at all ^[4]. Everything below rests on the text of 13 CFR, on the agencies' published statements, and on arithmetic that the tables recompute.

The Holding Company and the Operating Company Under 13 CFR 120.111

The program finances businesses that occupy and use what the loan buys. The regulation shuts out passive businesses whose developer or landlord owners do not use or occupy the assets the loan pays for, with one exception, the Eligible Passive Company ^[15]. The regulation defines that entity by what it does not do: it carries on no regular and continuous business activity, and it leases real or personal property to an Operating Company for use in that company's business, under the conditions of section 120.111 ^[16]. The Operating Company, in turn, is an eligible small business actively conducting, or about to conduct, its operations on real property the passive company owns, or using personal property the passive company owns ^[17]. In its simplest form the owners hold the building in one company and run the business in another, and the loan finances the building.

Section 120.111 then fixes the relationship between the two companies in ways that matter directly to a combined analysis. Table 1 sets out each condition and what it implies for the arithmetic.

The fourth row carries the most weight. Read as arithmetic, the cap says that rent is at most the loan payment plus the direct cost of holding the property. Subtract those two items from the rent and the most the holding company can have left is nothing. If the rent is set at the cap, the holding company ends the year at exactly zero after its debt service; if the rent is set below the cap, it ends the year short and needs cash from somewhere else in the group. Either way, its lease with the operating company can never make the holding company a second, independent source of repayment for this loan. Its income is the Operating Company's expense, and its debt service is paid with money the Operating Company earned.

SBA said as much when it rewrote its affiliation rules in 2023. Answering a comment that its revisions would leave holding companies and operating companies unaffiliated, the agency wrote that an Eligible Passive Company exists only to hold assets for an eligible Operating Company, which it called the qualifying entity on which the loan's cash flow and repayment are based ^[26]. That sentence is the starting point for any combined analysis of the structure. The two companies are one project for repayment purposes, and a figure that treats them as two independent earners has already gone wrong.

The 504 program uses the same structure. Among its basic eligibility rules, a 504 borrower must use the Project Property, an Eligible Passive Company may lease it to an Operating Company, and the borrower together with its Affiliates must meet one of the size standards in 13 CFR 121.301(b) ^[27]. How the 504 project is split between the bank, the development company and the borrower is covered in SBA 504 Method: The Job Test, the Project Split and What a Study Must Show the CDC. How the loan file records the industry of such a loan, whether under the occupant's code or a lessor's code, is a separate problem taken up in From NAICS Code to Asset Class: How SBA Loans Are Assigned, and Where Two Industries Share One Code. The site's 504 owner-occupied retail example in Michigan applies the logic directly: it builds coverage from the store's own earnings and adds only what the two small tenant suites net after the building's costs, and it notes that where a passive company holds the building, the rent the store pays to itself is capped by section 120.111(a)(3) ^[6].

Affiliates: The Circle the Size Test Draws and the Circle Credit Needs

Affiliation enters an SBA file for two stated reasons. SBA's 2023 rule explains that it determines affiliation to make sure an applicant is small enough for SBA assistance and to make sure that the applicant, its affiliates included, stays within the maximum guaranty amount ^[28]. The first reason appears in the basic eligibility rule, which requires an applicant to be small under part 121 with its affiliates counted ^[29]. The second appears in the statutory loan limit, which caps the SBA portions of all loans to a single borrower, counting affiliates as 13 CFR 121.301(f) defines them, at a guaranty amount of $3,750,000, with an exception where a statute provides otherwise for a specific program ^[30].

Part 121 opens with general principles. Concerns are affiliates when one controls or has the power to control the other, or when a third party controls or can control both, and the power is enough whether or not it is ever used ^[31]. SBA looks at the totality of the circumstances and may find affiliation even where no single factor would suffice ^[32]. When it measures size, it counts the receipts, employees or other measure of the concern and of all its domestic and foreign affiliates, whether or not those affiliates are run for profit ^[33]. Then comes the instruction that governs every business loan: for applicants in the Business Loan, Disaster Loan and Surety Bond Guarantee programs, the size standards and the bases for affiliation are the ones set out in 13 CFR 121.301 ^[34].

That section now works through ownership, because the 2023 rule removed control without ownership from the consideration of affiliation ^[35]. An applicant that owns more than 50 percent of another business is affiliated with it ^[36], and a business that owns more than 50 percent of the applicant is affiliated with the applicant ^[37]. An individual who owns more than 50 percent of the applicant and more than 50 percent of another business in the same 3-digit NAICS subsector makes the two businesses affiliates ^[38]. Where nobody holds a majority of the applicant, an owner of 20 percent or more who holds more than 50 percent of another business in the same subsector creates the same link ^[39]. Ownership by spouses and minor children is combined ^[40], and ownership held through entities is counted pro rata ^[41]. Among the older routes the 2023 rule took out, two bear directly on a combined analysis. It dropped affiliation based on identity of interest, which had required close relatives with no ownership stake to hand over financial statements for the size determination ^[42], and it provided that a franchised applicant is not affiliated with its franchisor or with other franchisees except through ownership ^[43].

Once the affiliates are known, the size test itself offers two routes for 7(a) and 504 applicants. The applicant may meet the industry standards of paragraph (a) ^[44], which carry two conditions: the applicant alone, without affiliates, must stay within the size standard for its primary industry ^[45], and the applicant with its affiliates must fit under whichever of two standards is higher, its own industry's or the group's ^[46]. Or it may take the alternative route, which sets a ceiling of $20 million on tangible net worth and a ceiling of $6.5 million on average net income after Federal income taxes, measured over the two completed fiscal years before it applies, with affiliates counted in both ^[47]. The mechanics of the industry route, and why a code can say less than it seems to, are covered in the series article on NAICS codes linked above.

None of this defines the group a lender needs to see for credit, and SBA's own answer to a comment on the 2023 rule points the same way. A commenter on the 2023 rule argued that lenders use affiliation as a guide to which entities they analyze for credit, and that tying affiliation to the applicant's industry code would skew that analysis ^[48]. SBA's answer was that the criteria for underwriting are in section 120.150, which lenders have followed and must keep following when they analyze a loan for credit purposes ^[49]. Read plainly, the answer keeps the two questions apart: affiliation settles size and the loan limit, and the credit analysis follows section 120.150. The size circle is drawn by ownership thresholds and industry subsectors. The credit circle is drawn by money: whoever can be called on to pay, and whatever claims can be made on those people.

The two circles can differ in both directions. An owner may hold a majority stake in a business outside the applicant's subsector. That business is not an affiliate under the ownership rules quoted above, yet if the owner has guaranteed its loans, its debt is a contingent obligation of a guarantor, which the agencies' definition of global debt service coverage brings into the count ^[2]. In the other direction, a business in the same subsector in which the applicant's majority owner also holds a majority is an affiliate for size even if it carries no debt and pays its owner nothing; it adds receipts to the size test and may add nothing at all to the cash available for the new loan. A study, or any analysis placed in the credit file, should therefore name both lists separately: the affiliates as the lender determines them for size, and every entity whose cash or whose claims reach the borrower or a guarantor. Merging the two lists hides exactly the businesses a global analysis exists to find.

Owners and Guarantors: Personal Cash Flow and Contingent Claims

The guarantors arrive in the file through the loan conditions. Holders of at least a 20 percent ownership interest generally must guarantee, and SBA or a lender acting under delegated authority may require other individuals or entities to give full or limited guarantees where it judges that necessary for credit or other reasons ^[5]. In the 2023 rule SBA called the personal guaranty from owners of 20 percent or more a prudent and reasonable measure for mitigating risk ^[50], and it explained that the disclosure of an applicant's owners exists to identify who must guarantee under the 20 percent rule ^[51]. With a holding company, the circle widens, because each holder of at least 20 percent of either company must guarantee ^[24].

The regulation also says what personal information the application carries. Among other things, the application carries three years of historical financial statements or tax returns and IRS tax verification for the business, and principals of the applicant, and of the Operating Company where there is one, provide personal histories and financial statements ^[52]. SBA describes its Form 413, the Personal Financial Statement, as the form used to assess the repayment ability and creditworthiness of applicants for 7(a) and 504 loans, among other programs ^[53]. Its Form 1919 is completed by the business applying for a 7(a) loan and handed to the participating lender ^[54], and it collects information about the applicant and its owners, the loan request, existing indebtedness, and current or previous government financing, among other topics ^[55]. Those documents, not a feasibility study, are where the personal side of a global analysis comes from.

This article builds the personal side in a fixed order. It starts with what the owner actually receives in cash from outside the circle of entities already counted: a salary from the operating company, wages a spouse earns elsewhere, rent from property the owner holds personally, interest and dividends. It subtracts the household's living costs and the owner's income and payroll taxes. The result is the owner's cash flow before personal debt service, and only then are the home mortgage, vehicle loans, credit card minimums and any other personal debts set against it. Contingent claims come last: every guarantee the owner has signed, measured by the payments of the debt it supports. The agencies' definition is explicit that contingent obligations belong in the combined obligations ^[2], and their policy statement adds that an effective assessment asks whether a guarantor can meet the total number and amount of guarantees it has already extended ^[56].

Cash flow is only one of the things the agencies want known about a guarantor. They say institutions with enough information on a guarantor's global financial condition, income, liquidity, cash flow and contingent liabilities are better placed to judge whether the guarantor can perform ^[57]. Liquidity appears in that list beside cash flow, not inside it, and a global analysis should keep it there: cash in the bank can cover a shortfall for a time, but it is a stock that runs down, not a flow that repeats, and adding it into annual cash flow would overstate the coverage of every year after the first. The statement also asks about conduct. Examiners should consider whether a guarantor has shown willingness to meet its current and past obligations, whether it has an economic incentive, and whether it has a significant investment in the project ^[58]. The statement treats as important whether any earlier performance on a guarantee was voluntary or the result of the lender's enforcement ^[59]. Sponsors with no legal obligation may resemble guarantors in ability, willingness and incentive ^[60], but an analysis that relies on a sponsor should say plainly that nothing obliges the sponsor to pay.

The regulation's own credit criteria tie these strands to SBA lending. They allow a lender to weigh the credit history of the applicant's Associates and guarantors ^[4], and an Associate of a small business includes an officer, a director, a key employee and an owner of more than 20 percent of the equity ^[61]. Those criteria name cash flow without saying how a personal cash flow is built or what ratio it must reach. The method is left to the lender's credit analysis, which the same section requires to be prudent and consistent with what the lender uses for similarly sized commercial loans without an SBA guarantee ^[62].

Where the Double Counts Hide

Adding entities together is easy. Adding them together correctly depends on finding every dollar that moves from one member of the group to another and making sure it is counted on one side only. Four items need that check.

Rent between related companies. The operating company pays rent to the holding company. If the operating company is shown after rent and the holding company is shown with its rent income, the rent nets out of the sum and nothing is wrong. The error appears when the operating company is presented before rent, on the reasoning that the rent goes to a related party, and the holding company's rent is also counted as income. The same money then appears twice. Under the cap, the size of the error is known in advance: it is the whole rent, which is the loan payment plus the direct holding costs ^[6]. The cure is to eliminate the rent once, either by showing each company on its own consistent basis or by combining the two as a single project in which the rent does not appear at all. Rent paid to an unrelated landlord is different: it leaves the group, so it stays in the expenses and is never eliminated. The site's 7(a) gym example in Hawaii carries the rent on its leased flex space among operating expenses in just that way.

Distributions and pass-through income. An owner's tax return carries the business's profit whether or not any cash was paid out. The IRS explains that S corporation shareholders report the corporation's flow-through income and losses on their personal returns ^[63], and that each partner reports a share of the partnership's income or loss on a personal return ^[64]. A personal cash flow built from the owner's return therefore already holds the profit that the business's own statements show. Distributions raise the same problem in cash form: they are money moving from a company already counted to an owner. The rule of method is to start the personal side from wages and from income that comes from outside the counted entities, never from the pass-through income or the distributions of a company whose cash flow is already in the total. For an interest that sits outside the circle, such as a minority stake in a business the analysis does not consolidate, the measure is the cash actually distributed, because pass-through income that was never paid out is not money the owner can spend.

Owner compensation. A partner's guaranteed payments are payments made without regard to the partnership's income ^[65]; the partnership generally deducts them as a business expense ^[66], and the partner reports them as ordinary income on Schedule E together with the partner's distributive share of the partnership's other ordinary income ^[67]. Counted as an expense of the business and as income of the owner, a salary or a guaranteed payment is treated consistently. It is counted twice when the business's cash flow is raised by adding the owner's pay back, while the personal side still lists the same pay as household income. Where a business's figures are restated to a market salary for the owner, the personal side has to use that same restated salary. How such restatements are documented for an acquisition is the subject of Buying a Business With a Loan: Why Coverage Is Tested on Historical Earnings, Not the Buyer's Plan.

Debt, counted twice or not at all. A guarantor's personal financial statement lists guarantees as contingent liabilities. If the guaranteed loan's payments already sit in the global debt service through the business that owes them, adding the same payments again as personal debt service counts the debt twice and understates coverage. The opposite mistake is more dangerous: bringing an affiliate's cash flow into the total before its own debt service while leaving its loans out, which overstates coverage by the full amount of that debt service. The agencies' definition requires both halves together, the cash flows of borrowers and guarantors and the combined obligations they carry ^[2]. An affiliate enters with its debt or not at all.

A global figure that has survived all four checks can be read. One that has not can be off by more than the loan's whole annual payment, as the worked example shows, and all but one of the errors described here make the borrower look stronger than it is.

A Worked Global Cash Flow, With the Rent and the Distributions Eliminated

The example is illustrative, and every input other than prime, the term and the site's stated floor is a labelled assumption. The owner holds a building in a holding company that qualifies as an Eligible Passive Company and runs an operating business in it. A 7(a) loan to the holding company buys the building, and the operating company joins it as a guarantor or co-borrower, as the regulation requires ^[23]. The owner also controls a second business that has loans of its own, and the owner has guaranteed them. Table 2 prints each input with the source or assumption behind it.

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

The loan is priced at prime plus an assumed spread. In the H.15 release, prime printed at 7.00 percent for every weekday from Monday, November 2, 2026, to that Friday ^[68], and the assumed 2.75 percent spread stays inside the cap of prime plus 3.0 percentage points that applies to 7(a) loans above $350,000 ^[69]. The term is the 25-year maximum available to real estate ^[70]. The first step is to lay out each entity on its own consistent basis, with the rent set at the regulatory cap, as Table 3 does.

At a rate of 9.75 percent the loan costs $256,648 a year. The rent at the cap is that payment plus $58,000 of direct holding costs, or $314,648, and the holding company is left with $0 once it has paid its lender, exactly as the arithmetic of the cap predicts ^[6]. The operating company keeps $105,352 after rent. The owner's household shows $8,000 before personal debt and is $36,000 short after it, while the affiliate keeps $50,000 after its own loans. Read entity by entity, each line is plausible on its own, and nothing is yet combined.

Table 4 shows what happens when the parts are gathered the careless way and then corrected.

Gathered as presented, the operating company appears before rent, the holding company appears with its rent income, and the owner's distributions appear as household income. The total is $934,648 against $410,648 of combined debt service, a coverage of 2.28x. Two eliminations remove the rent, counted once as operating cash and again as holding company income, and the distributions, which were paid out of cash flow the operating company figure already holds. What remains is $530,000, and global coverage is 1.29x. The figure as gathered counted $404,648 twice, more than the SBA loan's entire annual payment. Nothing in the figure as gathered is invented; every line exists on some statement in the file. The error is purely one of consolidation.

The last step separates the project from the support around it. Table 5 holds the rent and every personal and affiliate figure fixed and lowers the operating company's cash flow by one and then two of the assumed 15 percent shocks.

In the base case the project, meaning the operating and holding companies taken together, covers the SBA loan 1.41x and clears the binding minimum, which here is the stated 1.25x lender assumption because it sits above the site's 1.15x floor ^[71]. The cash flow supports a loan of $2,708,149 at that minimum, more than the loan requested. One shock takes project coverage to 1.17x, which still clears the site's stated floor but falls $21,809 short of the binding minimum. The owner and the affiliate together have $14,000 of surplus after their own obligations, so even if all of it were committed to the project, $7,809 of the gap would remain. Two shocks take project coverage to 0.92x and global coverage to 0.98x: at that point the whole group, counted once and counted correctly, does not earn enough to pay what it owes.

Two lessons follow. The first is that the global figure moves more slowly than the project figure, because the affiliate and the household are unaffected by the shock; a global ratio that still looks acceptable can coexist with a project that no longer covers its own loan. The second is that the outside support in a combined analysis can be small next to the project's own cash flow. Here the owner's household is in deficit before the affiliate's surplus is added, and the whole cushion outside the project is a fraction of the debt service. A global analysis earns its place by showing that, not by producing a larger number.

What a Feasibility Study Supplies, and What It Leaves to the Lender

The regulation names a feasibility study, beside professional appraisals and a survey, among the things SBA may require ^[72]. In a global analysis the study's contribution is the part of the combined figure that depends on the market: the project's own cash flow, built from evidence and cleanly separated from everything else in the file. That separation is the most useful thing a study can provide here, for a simple reason. The rest of the global figure, the owners' households, their guarantees and their other businesses, comes from documents the lender holds and the study does not, and every one of those figures is only as good as the project figure it is added to.

Concretely, a study should do five things on this subject. It should project the operating company's revenue and expenses from market evidence, with the owner's work paid at a stated market salary inside operating expenses, so that no owner compensation is left to be added back later. It should state the rent between the holding company and the operating company, show that it sits at or below the cap ^[6], and present the two companies as one project in which the rent cancels. It should report project coverage against the binding minimum before any outside support is counted. It should show the years before stabilization separately, because any shortfall in those years has to be met from somewhere outside the project. And it should run the downside cases on the project alone, so that a reader can see how quickly any outside surplus would be used up.

The gym example in Hawaii shows why the ramp years matter. It is a start-up in leased space whose early years fall below the assumed lender minimum, and whose two-year shortfall is larger than its ramp reserve, leaving the remainder for the sponsor to fund. A study that states the size of that gap tells the lender how much outside support the plan quietly assumes. The site's 7(a) gas station example in Tennessee shows the other limit. Its base case sits below the program floor as well as the assumed lender minimum, and the example's open question is whether location, fuel and store margins, or a different loan structure, can make up the difference. A global analysis could show whether the owners could carry the difference, but it would not change the project's own coverage, which is the number the conventions of these examples test.

A study does not gather, verify or judge the owners' personal financial statements; the application requires them from principals ^[52], and SBA uses its Form 413 to assess repayment ability and creditworthiness ^[53]. It does not decide which entities are affiliates or whether the applicant is small; under the Preferred Lenders Program, for example, eligibility questions including size, and the credit decision itself, rest with the delegated lender ^[73]. It does not weigh a guarantor's willingness to pay ^[58]. And it does not value the building, because a feasibility study is not an appraisal. The study's honest role in a global analysis is to supply one clean, sourced input and to say plainly where its work ends.

Which Members of the Group the Loan Record Shows

A global analysis gathers a group; the public SBA record keeps one row per approved loan, describing the borrower of record, the lender, the amount, the dates and terms, the location and the industry code, as What the SBA Loan File Cannot Tell You: The Limits of FOIA Data in Underwriting describes. No column names a guarantor, marks a borrower as a holding company leasing to an operating business, or lists the other businesses its owners control. The approval counts therefore measure how many loans this method could apply to, not how those loans were structured: FY2025 brought 78,078 ^[74] gross 7(a) approvals and 6,762 ^[75] gross 504 approvals, cancelled approvals included, and the record cannot sort the holding company loans out of either total.

Industry codes come closest to the structure, and only at one remove. The class built on lessors of nonresidential buildings holds 1,486 ^[76] gross 7(a) approvals and 81 ^[77] gross 504 approvals across FY2017 to FY2026. A lessor's code is the kind a company that only leases property to an operating business might report ^[16], yet the class also holds lessors of every other sort, while holding company loans reported under the occupant's industry sit in other classes altogether. The series' article on NAICS codes, linked earlier, treats that ambiguity in detail.

The franchise code is the one field that touches the group directly. It appears on 46.68 percent ^[78] of 7(a) gas station and truck stop approvals and on 48.47 percent ^[79] of 7(a) fitness center approvals. A franchisor in such a loan is not the borrower and, unless it has signed a guarantee, stands outside the credit circle this article draws: none of its cash is available to repay the loan, and what it brings into the analysis is its royalties and fees, which are operating costs of the borrower and belong in the project's expenses. Size is a separate question. For size, SBA's 2023 rule treats a franchised applicant as unaffiliated with its franchisor except through ownership ^[43], a rule that postdates many of the approvals behind those shares.

Outcomes are the last thing the record shows, and they are blunt. A charge-off says that a loan was written off; it does not say whether the operating business, the guarantors or the collateral was the part that gave way. Over the FY2017 to FY2021 approval cohorts the floor stands at 6.33 percent ^[80] for 7(a) and 0.35 percent ^[81] for 504, at 1.16 percent ^[78] for the 7(a) gas station and truck stop class, and at 9.48 percent ^[79] for the 7(a) fitness center class. All four are approval-cohort floors that can only understate, as The Approval-Cohort Charge-Off Floor: Why a Published SBA Loss Rate Can Only Understate explains. The dataset offers no reason for the gap between the two classes, so none is offered here, and the record does not show whether any guarantor paid after a charge-off.

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.

Questions a Lender or Borrower Should Ask of a Global Cash Flow

A combined analysis is easy to present and hard to check. These questions test whether one was built correctly.

• Which entities are inside the total, and is the list of affiliates for size kept separate from the list of entities that matter for credit ^[49]?
• Where a holding company owns the property, is the rent stated, is it at or below the cap of loan payment plus direct holding costs ^[6], and does it appear on one side of the combination only?
• Does the personal side start from wages and outside income, or from tax return income and distributions that the business figure already holds ^[63]?
• Is the owner's pay treated the same way in the business figure and on the personal side, and is any restated salary carried through to both?
• Does every affiliate whose cash flow is counted bring its own debt service with it, and is every guarantee the owners have signed accounted for, including the total number and amount of guarantees outstanding ^[56]?
• Is liquidity reported beside coverage instead of being added into a year's cash flow ^[57]?
• Is the project's own coverage, against the program floor and the lender's minimum, reported before any outside support is added, and does the downside case show how fast that support would be consumed?
• Does the study say what it did not do: verify personal statements, determine affiliation or judge a guarantor's willingness to pay ^[58]?

An analysis that answers each of these in writing gives the credit officer one number that can be trusted and the parts that produced it. One that cannot answer them has produced a large number and no evidence that it means anything.

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) Gas Station Feasibility Study in Tennessee
• Worked Example: SBA 7(a) Gym Feasibility Study in Hawaii
• Worked Example: SBA 504 Owner-Occupied Retail Feasibility Study in Michigan
• Debt Service Coverage in a Feasibility Study: How the Ratio Is Built, and Which Floor Binds
• SBA 7(a) Method: What 13 CFR 120 Tests and How a Feasibility Study Answers Each Test

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