Case Study 1: The Guaranty

How the SBA 7(a) program solves the restaurant collateral problem — and what it charges the borrower for doing so

Background

Start with the market failure the program exists to fix, because everything else follows from it.

A commercial bank makes money by lending at a spread and getting repaid. It manages the risk of not being repaid in two ways: by underwriting the borrower's cash flow, and by taking collateral it can sell if the cash flow fails. For most small businesses that second mechanism does some real work. A trucking company has trucks. A dental practice has chairs, equipment, and a patient list. A manufacturer has machines and receivables.

A new restaurant has almost none of it. Its largest single expenditure is construction that becomes part of a building somebody else owns. Its second largest is equipment that sells at auction for a fraction of what it cost, some of which — hoods, ductwork, make-up air, grease interceptors, anything site-built — is not equipment at all in a lender's eyes but a leasehold improvement that cannot be detached. It has no receivables, because guests pay before they leave. Its inventory is perishable and worth less every day. And it has no operating history, because it has not opened.

Underwrite that file on collateral alone and no one gets a loan. Which means, without some intervention, restaurant ownership belongs almost exclusively to people who already have money — or whose families do.

The Small Business Administration's 7(a) loan guaranty program is one of the interventions that changed that.

What the program actually is

The single most widely misunderstood fact about the program is what it does not do. The SBA does not lend the money.

Under 7(a), a participating lender — a bank, a credit union, or a non-bank SBA lender — makes the loan. That lender takes the application, underwrites the credit, approves or declines, funds, and services the loan for its whole life. You send your payments to the lender. What the SBA supplies is a guaranty: a promise to cover a portion of the lender's loss if the loan defaults, provided the lender followed the program's rules in making it.

That single structural feature does all the work. It does not make a weak restaurant strong. It changes the shape of the lender's loss, and in doing so it makes a category of credit bankable that would otherwise be declined on collateral grounds alone. The bank still wants to be repaid. It simply no longer needs the liquidation value of a used combi oven to be the answer.

Around that core sit the program's other features: eligible uses broad enough to cover leasehold improvements, equipment, working capital, and refinancing; maturities longer than a conventional commercial bank would ordinarily offer a business this size; rates quoted as a base rate plus a spread, within maximums the program sets; and a guaranty fee, together with the documentation regime that comes with any federal program.

A note on precision. Eligible uses, maximum loan size, maximum maturity, the guaranty percentage, the guaranty fee schedule, spread caps, and injection requirements are all set by SBA rules, and all of them have changed — sometimes more than once in a decade, and in some periods quite sharply. This case study describes the program's architecture, which is stable. It quotes no current parameters, and neither should you. Get them from a participating lender before you build a number into a plan.

The operating issue

For an operator, the program creates three practical problems that have nothing to do with whether the loan is a good idea.

Problem one: you are not applying to the SBA, so there is no single answer. Because participating lenders layer their own credit policy on top of the program's rules, the same file genuinely can receive different answers from different institutions. Lenders vary in how much restaurant experience they have, how much they like the category, how large a loan they want to make, how fast they move, and whether they hold delegated authority to approve within the program without a separate SBA review. An applicant who treats a declination as a verdict on the business rather than a verdict from one credit committee has drawn the wrong conclusion.

Problem two: the documentation is a real project. The package described in §5.3 — projections, personal financial statements, years of personal returns, résumés, the lease, bids, quotes, verification of the injection, program forms — takes weeks to assemble and generates follow-up requests for things you believe you already sent. Meanwhile the lease clock runs, the contractor's schedule slips, and the liquor-license timeline (Chapter 8) does not pause either. Operators underestimate this consistently, and the cost is not the paperwork. It is rent on a dark building.

Problem three — and this is the one that matters — the guaranty is not free, and you pay for it. This is the part of the program that goes unexamined in most enthusiastic descriptions of it. The guaranty protects the lender, not you. The program's rules direct lenders to take available collateral, and generally to obtain personal guarantees from owners at or above a specified ownership stake; depending on the loan and the circumstances, that can extend to liens on personal real estate where equity exists, and to a spouse's guarantee.

So the trade at the center of the program is this: a federal guaranty stands behind part of the lender's exposure, and the borrower's personal balance sheet stands behind the rest.

What it shows

Access to capital is engineered, not natural. Restaurants are not financeable on their own terms. The reason a line cook with fourteen years of experience and \$150,000 in savings can borrow \$335,000 is that a public program deliberately restructured the risk so a private lender would say yes. That is a policy choice, and it has visible consequences: independent restaurants exist in numbers they otherwise would not, and their owners are personally liable in ways that owners of better-collateralized businesses frequently are not.

The guaranty solves the lender's problem and relocates the borrower's. Nothing about the structure reduces what the borrower owes or what they have pledged. It makes the loan possible. The chapter's Figure 5.3 is the honest picture: \$620,000 of spending, almost none of it salable, and \$335,000 advanced against it. Something has to fill that gap. Part of it is federal. The rest is you.

The paperwork is not bureaucracy for its own sake. Much of what the program requires — verified injections, documented projections, tax returns, résumés — is exactly the material a competent lender would want anyway on a start-up with no history. The program formalizes it. An operator who resents the file is usually an operator who has not yet built one, and the building of it is genuinely useful: the assumptions register, the sources-and-uses statement, and the coverage arithmetic are management tools before they are application exhibits.

Outcome

The program has been a durable feature of American small-business finance for decades, and food service is consistently among the industries that use it heavily — which is exactly what you would predict from a program designed to bridge collateral gaps, applied to the least collateralizable small business there is.

Its parameters have moved repeatedly. Fees have been raised, lowered, and in some periods temporarily waived; guaranty percentages and size limits have been adjusted; underwriting standards have tightened and loosened with the credit cycle and with policy. During the COVID-19 period, temporary programs and subsidies altered the picture again. The architecture has been stable; the numbers never have. That is the single most important operational fact about the program for anyone planning to use it.

Lesson

Understand who is solving what, and the entire conversation gets easier.

The SBA is solving a market failure: private lenders will not lend against restaurant collateral. The lender is solving for repayment, and cares about your downside rather than your dream. You are solving for capital at a cost you can carry. The guaranty is the mechanism that lets those three interests overlap — and the personal guarantee is the price of the overlap.

Which produces three concrete instructions:

1. Shop the lender, not just the rate. Ask how many restaurant loans they closed last year, how they compute coverage, how draws work, and how long a complete file takes. The answers vary widely, and the differences cost more than a quarter point.
2. Get every number from the lender, in writing, and never from a book. Rates, spreads, fees, injection requirements, collateral policy. All of it changes.
3. Read the guarantee before you read the loan agreement. The loan agreement describes what the business owes. The guarantee describes what you owe, for as long as it takes, whatever happens to the restaurant. It is the most consequential document in the package and it is the one that gets the least attention.

Discussion questions

1. The program exists because private lenders will not lend against restaurant collateral. Is a federal guaranty the right instrument for that problem? Name one alternative and say what it would do differently.

2. Reconstruct the trade at the center of 7(a) in your own words: what the guaranty covers, what the personal guarantee covers, and who bears which risk. Then argue whether the trade is fair.

3. Chapter 1's discussion guide asked what a lender's declination rate for new restaurants should be. Revisit that question now that you understand the guaranty. Does a public guaranty make the access problem better, or does it mostly move the risk from institutions onto individuals who can least absorb it?

4. An applicant is declined by one participating lender. What are the three most useful things they can do next, and what is the one thing they should not do?

5. The case argues that assembling the application package is useful independent of the loan. Test that: which three exhibits would you still build if you were funding the restaurant entirely with your own money, and why?

6. Program parameters change constantly while the architecture stays stable. What does that imply about how a business plan should cite financing terms — and what should a plan say when the number it needs will not be knowable until an actual lender quotes it?