Owner-Occupied Retail Underwriting Method: Sales Capture, Occupancy Rules and the Retail Loan Record

A store that owns its building asks a lender to finance real estate whose most important tenant is the borrower. The rent roll that carries an investor's shopping center loan is missing. In its place stand the store's own till, the margin it earns on each sale, and the share of the building the business fills with its own shelves. The sections below take that loan apart in four steps: a sales capture model that turns the spending of a trade area into the sales of one store, the measures a credit officer reads to judge those sales, the federal occupancy rules that decide whether SBA money can buy the building at all, and the record the SBA loan files hold for store operators.

That record needs a plain warning before anything else. The Loan Analytics dataset has no single retail class. Retail trade is divided into many NAICS industries ^[1], and the asset map turns exactly two of them into classes: the liquor store and the grocery store. Those two are used here as the retail classes the dataset does hold. Their figures describe lending to liquor store and grocery store operators; they are not a measure of owner-occupied retail as a whole, and the text says so wherever they appear.

The general demand methods behind the capture model are explained in Trade Area and Demand Methods: Capture Rates, Penetration Rates and Gravity Models, and the coverage ratio itself in Debt Service Coverage in a Feasibility Study: How the Ratio Is Built, and Which Floor Binds. The full worked example for this asset type is Worked Example: SBA 504 Owner-Occupied Retail Feasibility Study in Michigan; the short model near the end of this article is a sketch, not a substitute for it. Readers who came for retail development on behalf of an investor will find the archived post The Ultimate Guide to Retail Development Feasibility Studies: A Financing-Led Framework closer to their question; here the subject is the merchant who buys or builds the box it trades from.

Retail in the Loan Dataset Is Two Store Types

NAICS does not treat retail as one industry. It sorts retail establishments into industries chiefly by the lines of merchandise they carry ^[1], and it pictures the store as a fixed point of sale placed and designed to pull in a large volume of walk-in customers ^[2]. That second description explains why a store building is collateral of a particular kind: the building's value to the business rests on location and frontage, and the business's value rests on what passes through the door.

The asset map that assigns SBA loans to classes works from those codes, a method set out in From NAICS Code to Asset Class: How SBA Loans Are Assigned, and Where Two Industries Share One Code. Among retail codes it keeps two whole. Code 445110, the grocery store class, takes supermarkets and other grocers whose main business is a general line of food, and it leaves convenience retailers out by name ^[3]. A corner market selling a short line of milk, bread, soda and snacks without fuel pumps belongs to 445131 ^[4], which no class in the dataset holds. Put fuel pumps in front of the same food mart and NAICS moves it to 457110 ^[5], which sits inside the gas station and truck stop class; that store type has its own article, Gas Station Underwriting Method: Fuel Gallons, Inside Sales and the Class's SBA Record. Code 445320, the liquor store class, takes establishments selling packaged beer, wine and spirits ^[6]. A business that sells packaged liquor and also pours drinks for customers on the premises is a drinking place under 722410 ^[7], outside the class.

Table 1 lays the boundaries side by side, because the first job of a study on a store loan is to say which code the borrower carries and therefore which record, if any, speaks to it.

Two other places in the asset map touch retail and neither helps the owner-occupier. Clothing retailers (458110) are listed among the large SBA industries the map did not propose as classes. Investor-owned retail buildings sit in the nonresidential lessor class with flex and office buildings, the landlord's side of the market, which the archived Flex Space Feasibility Study page covers for its own property type.

One more limit applies to both cells. The code on an SBA loan names the borrower's business. It does not say whether a liquor store borrower bought its building, leased it, financed inventory or paid for an existing store, and the loan file has no field that would settle the question, as What the SBA Loan File Cannot Tell You: The Limits of FOIA Data in Underwriting explains. The two cells are a record of SBA lending to two kinds of store operator. They bear on owner-occupied retail real estate only indirectly, because the file does not separate the loans that financed buildings from the rest.

The Sales Capture Model, Written as Four Equations

A store's revenue projection should be traceable from the people who live near it to the dollars at the register. The capture model does that in four steps, each with one input a reviewer can challenge on its own.

1. Resident spending through stores of the type: E = R × s × i, where R is the trade-area population, s is national sales per resident through stores of that kind, and i is a local spending index.
2. Sales won from residents: Sr = E × c, where c is the subject's capture of that spending.
3. Store sales: S = Sr ÷ (1 - f), where f is the share of the store's sales that come from customers living outside the trade area.
4. Productivity check: S ÷ A, sales per square foot of the space the store occupies.

The population term comes from the Census Bureau's Population Estimates Program, whose estimates run from the national total down to counties, cities and towns ^[8]; drawing a drive-time boundary and filling it with block-group counts is the trade area article's subject, and nothing about it changes for a store. The national per-resident figure s is where the store type enters. The Census Bureau's Annual Retail Trade Survey (ARTS) publishes annual sales of retail firms by kind of business. Supermarkets and other grocery stores other than convenience stores sold 808,759 million dollars in 2022 ^[9], and package stores of beer, wine and liquor sold 70,915 million dollars ^[10]. Divided by the population the survey's own per capita table uses, 333,271 thousand residents as of July 1, 2022 ^{[11] [12]}, those totals give the per-resident figures in the capture table below. For grocery the ARTS line chosen is the supermarket line rather than the wider grocery store total, because ARTS runs on the 2017 NAICS ^[13] and its supermarket line is the nearest match to the 2022 code the dataset's grocery class uses.

Three features of the anchor need saying in the study, not discovered later by a reviewer. ARTS covers firms with paid employees ^[14] and its final release is employer-only ^[15], so stores with no payroll are outside it. Its dollars are not adjusted for price changes ^[16], so a 2022 figure carried to a later study year needs a stated inflation or sales-growth step. And it reports the sales of retail firms by their kind of business ^[17], with each firm's establishments summed into one sampling unit ^[18], so beer and wine sold by a supermarket firm sit in the grocery line, and the liquor store line holds what package store firms sell. Where the local mix of outlets differs from the national mix, that difference belongs in the local index i, with its evidence, not in a quiet change to the capture share.

The local index i converts a national average into a local one. It is an assumption in the table, labelled as such, and in a real study it would be derived from local income relative to the nation and from any local evidence on prices and spending patterns. Setting it at 95 percent says the trade area spends a little less per head through these stores than the nation does; the study should show the arithmetic behind whatever figure it uses.

The capture share c carries most of the risk. It is the part of resident spending through stores of the type that the subject wins against every other outlet that sells the same goods. Its evidence is a count of competitors and their sizes, which the Census Bureau's County Business Patterns supports with annual establishment counts by industry and employment size ^[19], although that series covers only establishments with paid employees ^[20]; the owner-run shops it misses appear in Nonemployer Statistics, which covers businesses with no paid employees by industry and geography ^[21]. A capture share should survive a simple test: the shares claimed for the subject and for every existing store, plus the spending that leaks to stores outside the trade area, cannot add to more than all of it. A study that gives the subject 14 percent should show who keeps the rest. The supply side of that test is the subject of Supply and Comparable Analysis: Counting Competitors, the Pipeline and the Properties a Study Leans On.

The inflow share f lifts resident sales to total sales. A store on a commuter route or near a regional attraction draws customers from outside any boundary drawn around homes. The evidence is the store's own records where they exist: card transactions by billing postal code, loyalty accounts, a customer survey. For a new location the inflow should come from a comparable store in a comparable position, never from a round number.

Table 2 runs both store types through the same four equations. The grocery example starts from 14,000 residents and ends at $4,911,492 of sales. The liquor store example needs a larger population for a much smaller sales figure, because national sales per resident through package stores are less than a tenth of the supermarket figure. The last row is a check, not an input. Sales per square foot occupied fall out of the model, and a study should compare them with the operator's existing stores and with comparable stores before it accepts the capture share that produced them. None of the Census tables used here publishes sales per square foot by store type, so that comparison rests on operating records the study gathers, and a figure well above what comparable stores achieve is the clearest sign that the capture share is too generous.

Six Measures a Credit Officer Reads in a Store File

For an owner-occupied store the property's income is the business's income, so the measures that matter are trading measures, read beside one real estate measure that only the SBA rules impose. Table 3 lists them with the evidence a study assembles and the public series or rule that anchors each one.

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.

Sales come first because every other line is a share of them. For an operating store, a study should reconcile three records that are kept for different masters: the income tax return, the state sales tax filings, and the point-of-sale system. A gap between them is not automatically a warning sign, since the three records can define a sale differently, but it has to be explained before any sales figure is projected forward.

Gross margin is sales less the cost of the goods sold, as a share of sales. ARTS reports it by kind of business: 28.0 percent for grocery stores and 31.1 percent for beer, wine and liquor stores in 2022 ^{[27] [28]}. A store's own margin moves with its mix (fresh food against packaged goods, spirits against beer), with its buying terms, and with shrink, which an inventory count exposes and an income statement can hide.

The operating expense ratio is where national and local figures part company most easily. ARTS collects total operating expenses from retail firms, and Table 4 divides them by sales: the result is a national figure for firms of every size, not a budget. The budget a lender can trace is built line by line from the store's own ledger, as Operating Expense Evidence: Building an Expense Budget a Credit Officer Can Trace describes, and the national ratio is a test of whether the budget is plausible.

Read across, Table 4 shows why a study should not underwrite the two store types alike, even though both sell consumables. The liquor store line earns a wider gross margin and spends a smaller share of sales on operations, which leaves a larger share of sales after total operating expenses as ARTS counts them. It also holds far more stock per dollar of sales: the cost of goods sold passes through end-of-year inventory about 12.4 times a year in the grocery line and about 4.9 times in the liquor line, both computed from the ARTS totals ^{[24] [29]}. A liquor store therefore ties more of the owner's cash in bottles on the shelf, and the inventory question becomes a financing question. SBA's 504 program page says a 504 loan cannot pay for working capital or inventory ^[30], while the 7(a) page lists short and long term working capital among 7(a) uses beside real estate ^[31]. A study on a 504 building purchase should show where the stock is financed.

The ARTS ratios are firm aggregates and should be read as such. They are published for the wider grocery store line, 4451, which contains the supermarket line used for sales per resident ^{[32] [9]}. The survey's frame is employer firms whose main business is retail ^[14], and in its sample design the largest firms in each industry are all selected with certainty ^[33]. A ratio of national dollar totals moves with the firms that hold the most dollars, so the grocery line reflects large multi-store firms at least as much as it reflects a single family-owned market. The national figure is the right check on whether a projected margin is plausible and the wrong number to put in a projection without the store's own history.

Last comes coverage. Loan Analytics' own SBA program page sets 1.15x as the floor for a standard 7(a) loan, allowing a startup two years after funding to reach it ^[25]. The same page of the site gives 504 loans 1.15x as well, to be shown on history or on a projection two years out ^[34]. For the owner-occupier one derived measure is worth printing beside coverage: annual debt service as a share of store sales. It plays the part that rent as a share of sales plays for a tenant store, and it shows at a glance how much of each sale the building takes.

The Occupancy Rules That Decide SBA Eligibility

Every owner-occupied store loan under an SBA program starts from the passive business rule. The bar in 13 CFR 120.110(c) reads: "Passive businesses owned by developers and landlords that do not actively use or occupy the assets acquired or improved with the loan proceeds" ^[35]. Holding companies have one way around that bar, the Eligible Passive Company, taken up below. A store that occupies its building is not passive. A store owner who buys a strip center and fills one bay of it may be, and the occupancy rules draw the line.

The measuring stick is what the regulation calls Rentable Property: "the total square footage of all buildings or facilities used for business operations" ^[26]. The share is taken over the whole building, so a mezzanine, a loading area or a suite let to a tenant counts in the total. A study should measure from plans and say whose plans they are.

For an existing building bought, renovated or rebuilt with 7(a) or 504 money, the test is 51 percent: the borrower must itself occupy and use, on a permanent basis, at least that share of the Rentable Property, and may then let up to 49 percent to others on permanent leases ^[36]. A new building faces 60 percent and a timetable. Permanent leasing is capped at 20 percent there, and is allowed only when the borrower occupies and uses 60 percent or more on a permanent basis ^[37], intends to take up, within three years, part of the space it neither fills at once nor lets permanently ^[38], and intends to take up within ten years everything not let on a permanent lease ^[39]. The rule therefore does more than set a percentage for new building: it asks the study to show a growth plan that fills the vacant remainder, and a store with no plan to expand into it has a problem the percentage alone does not reveal.

Ownership is sometimes split, with a holding company owning the building and renting it to the store. The regulation permits the split through what it calls an Eligible Passive Company, on conditions a study should tick off in turn; the ones that matter most to a store loan follow. If such a company leases an existing building in full to operating companies, the occupancy test is applied to the operating companies taken together ^[40]. The lease between holding company and store has to be in writing and must rank behind SBA's lien or security interest ^[41]. Rent is capped: it may cover the loan payment and the holding company's direct costs of owning the building, with maintenance, insurance and property taxes given as examples, and nothing more ^[42], so the store cannot be charged above what the debt and the building's direct costs require. The lease must not end before the loan does, renewals that the store alone may exercise included ^[43], and the store must sign either as guarantor or as co-borrower ^[44]; if the 7(a) loan also finances the store's working capital or other assets, only the second of those roles will do ^[45]. For underwriting, that last condition matters most: the store's cash flow stands behind the loan either way, which is why the coverage test is run on the store and not on the lease.

Three other SBA eligibility rules bear on these two store types. Businesses that take more than one third of gross annual revenue from legal gambling are ineligible ^[46], and a liquor store or grocer with lottery sales or gaming machines should report that revenue as a share of gross revenue, so the lender applies the test to a figure rather than to an impression. SBA also refuses any business engaged in an activity that federal, state or local law makes illegal ^[47], so the state license (next section) is an eligibility document as well as an operating permit. And the program chosen changes what can be financed. Under 13 CFR 120.212(c) the longest 7(a) term is 25 years with extensions counted, and the part of a loan spent acquiring or improving real property may add the time that construction or improvement takes ^[48]. A 504 debenture comes in 10, 20 or 25 year terms ^[49]. On a 504 project the borrower's own money must reach 10 percent of cost at a minimum when nothing unusual applies ^[50]. A young business, two years old or less (for an Eligible Passive Company, the age of the operating company counts), raises that minimum to 15 percent ^[51]; so does a limited or single purpose structure ^[52]; and the two together raise it to 20 percent ^[50]. The section does not say who decides whether a store building is limited or single purpose, so the study states which treatment it assumed and why.

A store may also stand on a ground lease, for example on a pad site in front of a larger center. Under 13 CFR 120.870(a), 504 proceeds may finance buildings and improvements on land leased from an unrelated lessor if the remaining lease term, counting renewal options the lessee alone can exercise, at least equals the debenture term ^[53], provided too that the CDC receives an assignment of the borrower's lease interest, with SBA able to take it by reassignment, and that a recorded lien on the leasehold estate, plus other collateral where needed, secures the loan ^[54]. A study should print the lease's remaining term with options beside the proposed debenture term.

A store outside the cities may also look to USDA's Business and Industry guarantee. The OneRD rule lists the purchase and development of land and buildings for commercial enterprises among eligible project types ^[55] and requires the project to be in a rural area, with two exceptions: loans to cooperatives, and local foods projects, which may also sit outside rural areas ^[56]. The rural test itself shuts out cities and towns above 50,000 inhabitants and the urbanized areas next to them ^[57], but not every part of that fringe: an area there judged rural in character stays inside the definition ^[58], and each such judgment is for USDA's Under Secretary of Rural Development to make, one project at a time ^[59]. The local foods exception carries its own conditions. It covers projects that process, distribute, aggregate, store or market locally or regionally produced food ^[60], and such a project may sit outside a rural area only if it meets each of four tests: it expands or preserves the availability of staple food in underserved areas with moderate and low income populations ^[61], by keeping or adding retail or institutional outlets that offer healthy perishable and staple foods ^[62]; it creates or retains quality jobs for low income residents, with a significant amount of the food locally or regionally produced and sold ^[63]; and it carries an agreement with the retail and institutional buyers to tell consumers the food is local or regional ^[64]. Those tests turn on staple food, so they can reach a grocery store that sells locally produced food on those terms; a package store selling beer, wine and liquor does not fit them. Where a project spans rural and non rural areas, the guarantee covers only the portion in the eligible rural area ^[65]. Its gambling test is stricter than SBA's. B&I turns away any entity drawing over 15 percent of annual gross revenue from gambling, with rent from leased space or machines included in the count ^[66]; the rule leaves out State or Tribal authorized lottery or gaming proceeds that the Agency approves and that are conducted to raise funds for the approved project ^[67]. A rural grocer looking for a different USDA program should know that the rule lists grocery stores among the inherently commercial enterprises ineligible for Community Facilities financing ^[68], a bar that applies to that program only ^[69]. The program tests in full are in SBA 7(a) Method: What 13 CFR 120 Tests and How a Feasibility Study Answers Each Test, SBA 504 Method: The Job Test, the Project Split and What a Study Must Show the CDC and USDA B&I Guarantees Under 7 CFR 5001: What the Agency Tests and How a Study Answers.

Licenses and Food Assistance Authorization

Both store types depend on permissions that sit outside the loan programs, and a study that treats them as paperwork misses part of the demand and supply picture.

For a liquor store the license is a condition of trading and, where a state caps the number issued, a scarce asset. Florida is one example of how a state can tie the number of licenses to population. Section 561.20(1) of its statutes allows a county no more licenses of the classes in section 565.02(1)(a) to (f) than one per 7,500 residents ^[70], a group of license taxes that applies to vendors permitted to sell alcoholic beverages of any alcoholic content ^[71] and that opens with the vendor whose place of business sells only in sealed containers for consumption off the premises, the package store ^[72]. The same subsection keeps at least three such licenses available in any county that may approve the sale of intoxicating liquors ^[73], and section 561.20(3) says the population limit does not apply to existing licenses or to their renewal or transfer ^[74]. The cap counts every license in that group, so package stores and on-premises vendors draw on the same county total ^[70].

Table 5 applies the ratio to an illustrative county. The result is not a count of package stores: it covers every license class in the group, and the statute also issues a new license for each 7,500 residents of growth above the county's April 1, 1999 population estimate ^[75], so the licenses outstanding in a real county reflect its history as well as its present size. A study in a state with a cap like this should treat the cap as a limit on supply and count the licenses actually issued near the site. Where a new store needs an existing license from another holder, the price of that license belongs in the sources and uses of the project, as Equity Injection and Sources and Uses: How a Study Shows Where Every Dollar Comes From lays out for any cost. The Florida cap is one state's rule, and a study should cite the statute that governs its own site rather than assume Florida's.

For a grocery store the parallel permission is authorization to accept SNAP benefits. Under 7 CFR 278.1(b)(1), a retail food store that sells food for home preparation and consumption normally qualifies if it offers, on a continuous basis, a variety of qualifying foods in each of the four staple food categories, with perishable foods in at least three of them (Criterion A) ^[76], or if more than 50 percent of its total gross retail sales are staple foods (Criterion B) ^[77]. Criterion A is specific: on any day of operation, no fewer than seven varieties in each of the four staple categories, with a depth of stock of three stocking units for each qualifying variety ^[78]. The agency may base its eligibility decision on visual inspection, sales and purchase records, stockkeeping unit counts and other customary inventory or accounting records, among other factors ^[79]. A grocer that grows a large hot food and prepared food counter should also know that firms with more than 50 percent of total gross sales in foods cooked or heated on site are treated as restaurants for this purpose ^[80].

For the study, SNAP authorization is part of the revenue evidence. A study should state whether the store is authorized, report the share of sales tendered in benefits from the store's own records, and test whether the proposed layout and stocking plan keep the store inside the criterion it relies on. Losing the authorization would remove a tender some customers depend on, and the projection should not assume it away.

What the Liquor and Grocery Cells Record

The national cells, PA-C37 and PA-C40 for each program, hold approvals for fiscal years 2017 to 2026. Table 6 gives the size of the record and Table 7 its course year by year. Every figure is the liquor store or grocery store record; neither is a figure for retail.

Three readings follow from Table 6, and none of them needs a cause. First, 7(a) approvals far outnumber 504 approvals in both store types; the table does not say why, and the study should not guess. What the program texts do settle is that the two programs pay for different things: SBA's 504 page names "Working capital or inventory" among the things a 504 loan may not fund ^[30], while the uses SBA lists for 7(a) loans run from real estate and buildings to working capital ^[31]. For the owner-occupied question, then, the cells for the 504 program are the nearer record, since SBA describes 504 as financing for major fixed assets ^[85], and the 7(a) cells mix building loans with every other use. Second, the 504 medians sit well above the 7(a) medians in both store types, and the 504 figure is the debenture alone, so the project behind each 504 approval is larger still. Third, franchise codes are rare in both classes, so most approvals went to borrowers without one, and a study on such a store will be judging the operator's own history rather than a franchise system's averages.

The year by year counts in Table 7 show lending to both store types in every year of the window, in both programs, with no year in which either class disappears. The 504 counts are small enough that a single year's movement says little; the 7(a) counts are large enough to show a level. No cause for the movement from year to year appears in the dataset's report, and a study should not invent one from the table.

Below, the loan-data block gathers the dataset figures used above and gives each one its key, so that any of them can be traced.

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.

How the Cohort Floors Compare With the Program

The cancellation and charge-off figures for the two store types are approval-cohort floors, as The Approval-Cohort Charge-Off Floor: Why a Published SBA Loss Rate Can Only Understate explains: open loans are exempt from status disclosure and the cohorts are still accruing losses, so each printed rate can only understate. Cancellation, an approval that ends before money moves, is a different event from charge-off, as Cancellation Is Not Charge-Off: The Two Ways an SBA Approval Can End Without Repayment sets out. Table 9 sets the class floors beside the program floors from the dataset's national figures; as before, the columns are the liquor store and grocery store records, not a record for retail.

Every rate in Table 9 rests on more than 30 loans, so none is withheld, and every one is a floor. The two 7(a) store types diverge. The grocery store floor of 7.64 percent ^[83] stands above the 7(a) program floor of 6.33 percent ^[86], while the liquor store floor of 2.48 percent ^[81] stands well below it. The measure that holds the horizon fixed, charge-offs within five years of approval, tells the same story more mildly for grocery, which sits close to the program, and keeps liquor stores well under it. Cancellation floors for both store types run a little above the 7(a) program figure. In the 504 program neither store type shows a charge-off to date, against a program floor of 0.35 percent ^[87], and the liquor store 504 cancellation floor of 15.48 percent ^[82] stands above the program's.

A floor printed beside a program floor invites a question the table cannot answer by itself: is the gap more than the chance variation a cohort of that size would show? Table 10 does the arithmetic. It applies the program floor to each class's disbursed loans to get the charge-offs the program rate would give, takes the difference from the class count, and divides by the binomial standard deviation at the program rate.

The answer differs by row. The liquor store 7(a) gap is many standard deviations wide, more than chance variation in a cohort of that size would plausibly give. The grocery store 7(a) gap is about two standard deviations, large enough to note and small enough that a modest change in the open loans could move it. The two 504 zeros are well inside chance: at the program floor, a cohort of that size would be expected to show less than one charge-off, so a count of zero says almost nothing about whether either class is safer than the 504 program as a whole. Why a Rate on Fewer Than 30 Loans Is Withheld: Small Samples and the Suppression Rule sets out why small cohorts need this kind of reading.

What the floors cannot say matters as much. The arithmetic treats each loan as an independent draw at the program rate, which loans in one store type, approved in the same years and exposed to the same conditions, may not be. The comparison is floor against floor, and the shares of open loans differ between the cells and the program, so part of any gap may close as the open loans resolve. The cells mix uses of proceeds, so a charge-off in the grocery 7(a) cell may be a business acquisition or a working capital line rather than a building loan. And nothing in the file gives a reason for a gap: no cause should be read into these figures, by a study or by a lender, unless the dataset's report gives one. A study should use the floors to decide which assumptions deserve the hardest stress test, never as a probability of default for the store in front of it.

A Short Model: A Grocer Opens a Second Store in a Building It Buys

The full worked example for owner-occupied retail is the Michigan 504 case linked at the start, which builds a purchase, a renovation, a ramp and a sensitivity grid. The model here is deliberately short. An established independent grocer, operating its first store for well over two years, buys an existing 15,000 square foot building in a second trade area, occupies 11,000 square feet as a new store, and leases the remaining suite to an unrelated tenant. The building purchase is financed with a 7(a) loan at a variable rate. Every project figure is an assumption labelled in the inputs table; the national anchors are sourced and cited there.

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

Table 12 answers the eligibility question first. The store occupies 73.3 percent of the Rentable Property, above the 51 percent that 13 CFR 120.131(b) requires for an existing building ^[36], and the leased suite is within the 49 percent the same paragraph allows. Had the grocer bought a center several times the size of its store, the share would fall below the line and the study would stop there.

The cash flow then runs from the capture table's sales through the ARTS grocery margin, a store expense budget and a market wage for the owner, adds the suite's net rent and deducts the costs of the whole building. Coverage comes out at 1.73x. Of the two tests, the 1.25x lender assumption is the stricter, so it binds; the 7(a) floor on the site's SBA program page is lower, at 1.15x ^[25]. The cushion looks wide until it is turned into the inputs that produce it. The store's sales could fall to the break-even figure in Table 12 before coverage meets the minimum, which means a capture share of 10.9 percent instead of 14.0. Two other rows deserve a lender's eye. Debt service takes about 4.2 percent of every dollar of sales, the figure a grocer would compare with the rent it pays at its first store. And stocking the new store at the national grocery ratio of sales to inventory would take about $285,951 of inventory, money the building loan does not provide and that a study should show coming from somewhere: the owner's cash, a separate 7(a) working capital loan, or supplier terms.

Table 13 applies four single shocks. A capture share three points lower brings coverage to the edge of the minimum. A gross margin two points lower lands almost exactly on it, because in a grocery store a point of margin on several million dollars of sales is a large share of the cash left for the lender. A two point rise in prime, which a variable rate passes straight to the payment, and a vacant suite each cost less. The ordering is the lesson: for this store type the margin and the capture share are the inputs the study must prove, and the building's own rent is secondary. Combined shocks and the choice of shock sizes are covered in Sensitivity and Stress Testing: Break-Even Points, Shock Sizes and What a Lender Should See.

Checks Before a Store Loan Goes to Credit Committee

Everything above comes down to a handful of questions. A borrower can answer them before ordering a study, and a credit officer can ask them of any store study that arrives.

• Which code, which record. Does the study state the borrower's NAICS code, and does it use the liquor store or grocery store record only where the code matches, saying plainly that neither is a retail-wide figure ^{[3] [6]}?
• A capture share with a ledger. Is the capture share supported by a competitor count with sizes, and do the shares claimed for every store in the trade area add to no more than the spending available ^[19]?
• National anchors used as checks. Are the ARTS margin, expense and inventory ratios used to test the store's own figures rather than to replace them, and is the price and year adjustment from 2022 stated ^[16]?
• Occupancy measured on the whole building. Is the occupied share computed on all space used for business operations, with the existing building or new construction test applied and, for new construction, the three year and ten year plans shown ^{[26] [38] [39]}?
• The related-party lease, if there is one. Where an Eligible Passive Company holds the building, does the lease meet the rent cap, the term and the subordination conditions, and is the store a guarantor or co-borrower, with co-borrower the required role if the 7(a) loan stretches to the store's working capital ^{[42] [43] [44] [45]}?
• Gaming revenue as a number. Is revenue from lottery sales or gaming machines reported as a share of gross revenue, so the SBA one third test or the B&I 15 percent test can be applied ^{[46] [66]}?
• The permission to trade. Is the liquor license identified by class, with its transfer status under the governing state statute, and is the grocery store's SNAP authorization and the criterion it relies on stated ^{[74] [76]}?
• Inventory financed. Does the sources and uses show where the opening or expanded stock is financed, given that a 504 loan cannot pay for inventory ^[30]?
• Floors read as floors. Where the study cites the cells, does it call the rates approval-cohort floors, keep the 504 zeros in proportion, and assert no cause for the gaps?

A study that answers each of these can show a credit officer the path from residents to sales to the cash that pays the building loan, and the points on that path where the loan would fail.

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)
• USDA feasibility studies
• Bankable feasibility studies for conventional loans
• Loan data by state
• Loan data by asset class
• Worked examples (case studies)
• Worked Example: SBA 504 Owner-Occupied Retail Feasibility Study in Michigan

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