The price is a decision. The loan is arithmetic.

A smaller number arrives

An investor won a twelve-unit building near Flagler after two rounds of best and final. The price was full and he knew it, and the building deserved it: a clean rent roll, a newer roof, tenants paying on the first. The term sheet in his file said up to 80%, and the plan was built on 75.

Three weeks later the loan came back at 60% of the price. Nothing had gone wrong with the building, the appraisal, or the borrower. The lender had run three computations against the same operating statement and taken the smallest answer, which is exactly what the term sheet said it would do.

The words that carried the file were up to.

In its most common shape, agency and institutional lending on stabilized buildings, a commercial mortgage is not a percentage of the price. It is the lowest of three tests, and the three answer different fears.

1. Loan-to-value. The loan measured against the appraised value. This is the famous percentage, and it is a ceiling.
2. Debt service coverage. The income measured against the payments. The building has to earn its own debt service with room to spare.
3. Debt yield. The income measured against the loan itself, with no interest rate and no amortization anywhere in the formula.

The loan is the smallest of the three answers. Everything else in this piece is that sentence, worked out.

That is the widest lane of the market, not the whole road. Bridge lenders, private money, and banks lending on their own balance sheet each run their own arithmetic. Some of them price the story and the exit more than the trailing income. The three tests are the math a stabilized building meets first, and the discipline they teach travels to every desk.

The three tests, and who sets them

Loan-to-value has a published limit. Fannie Mae’s fixed-rate conventional term sheet states it plainly: 80% maximum for conventional properties. The denominator is the appraised value, not the contract price, so the test moves with the appraisal and stands still when the market moves the price.

Coverage has a published limit too. The same term sheet sets the minimum debt service coverage at 1.25x for conventional properties: the underwritten income has to cover a full year of payments 1.25 times. The payment carries the interest rate inside it, and that detail does the quiet work. When the rate rises the payment rises, the income does not, and the loan that satisfies 1.25 gets smaller. Nothing about the building changed.

Debt yield is the lender’s own brake. Income divided by the loan, with no rate and no schedule in the formula. It answers a plainer fear: if this loan came back to the lender tomorrow, what would the income earn on it. The floor is not an agency constant. It is printed on the individual term sheet. On the sheets that carry it, it is the test that refuses to be argued with: there is no rate assumption inside it to argue about.

The same sheet says one more quiet thing: escrows for replacement reserves, taxes, and insurance are typically required. The income in these tests is income after the building has been made to fund its own future.

The arithmetic, worked in the open

The numbers below are illustrative. The inputs are round and hypothetical. The two agency limits are real and cited above.

The exercise: a building with an underwritten net operating income of $180,000, under contract at $3,000,000, appraised at the price. The rate is 7% on a 30-year amortization, stated here as an input rather than a quote.

The value test: 80% of $3,000,000 is $2,400,000.

The income test: $180,000 divided by 1.25 leaves $144,000 a year available for payments. At 7% on a 30-year schedule, $100,000 of loan costs about $7,980 a year. The income carries $144,000 of payments, so the loan that fits is about $1,800,000.

The loan is the smaller answer: $1.8 million, which is 60% of the price. The plan said 75, and 75% of the price is $2,250,000. The gap between the plan and the file is $450,000, and it arrives as equity, dollar for dollar.

The inversion is where the pricing information lives. At this income, the price where a 75% plan actually works is $2,400,000. And for the value test to be the one that binds at $3,000,000, the underwritten income would have to reach roughly $239,000. Between those two numbers sits every conversation about this building’s price.

Whose operating statement is in the formula

The income in every test above is the lender’s number, not the listing’s. Underwriting rebuilds the operating statement the way the next owner will live it, and the rebuilt version is usually smaller than the one in the brochure. Not because anyone inflated anything: the two documents answer different questions. A listing describes the building as it has been run. A loan file describes the building as it is about to be owned.

Two lines move the most in Miami. Property taxes go in as the next owner will pay them, because the assessment does not travel through a sale: the cap resets, and our tax piece works that arithmetic in the open. Insurance goes in at the quote in the file, not at the premium the seller has been renewing for years.

The rest of the rebuild is quieter and runs the same direction. A vacancy assumption is applied even where the rent roll is full, because the sheet underwrites stabilized occupancy rather than a good month. Management is priced even where the owner manages the building personally, because the next owner might not. Reserves are set aside before income counts as income, which is the escrow line from the term sheet doing its work.

An owner reading a smaller NOI than the one they live with has not been corrected. They are reading a different document about a different year: the buyer’s first one.

Three chairs

The seller reads the price against the market. Comparable buildings traded at comparable numbers, and a full price on a clean building is not wishful, it is the market speaking. That position is real and often right. What the seller’s chair cannot see is which test binds inside the buyer’s file. When the income test is the one that binds, every dollar of price above what the debt supports is funded entirely from the buyer’s equity. Equity is the scarcest money at the table. The price can be right and the pool of buyers who can reach it narrow at the same time.

The buyer reads the loan against the last deal. The 75% in the plan came from real closings, not from optimism. The strongest version of that chair is discipline: a leverage assumption that worked repeatedly is a reasonable planning number. What that chair has trouble seeing is that the percentage was always an output. It was produced by a rate, and at a different rate the same building at the same price produces a different loan. Carrying the old percentage forward is pricing the memory of a rate.

The developer reads the building as a bridge. From that chair the purchase is about the lot, the income is temporary, and refinement of an operating statement feels like bookkeeping on a building that is coming down. The honest strength there is patience: the value being bought really is in the future. What the developer’s chair cannot see is that the acquisition loan is sized on today’s income all the same, and the years between closing and permits are lived on that loan. A thin NOI that read as a footnote in the pro forma becomes the entire carry.

Where they collide

The collision is that a price negotiation is partly a negotiation about a spreadsheet nobody at the table controls. The seller is arguing with the market, and the market is real. The buyer is arguing with a credit file, and the file is real too. The two can disagree about the same building without anyone being wrong.

The version we run into most often is a single sentence with its second half missing. A strong price gets quoted with the lender’s percentage attached: buyers can finance 80%. The sentence is true with the second half restored: 80% of the lower of the appraised value or what the income carries. On a building where the income lags the price, those are two different numbers, and the distance between them is somebody’s equity.

There is no villain in that gap. A seller quoting the ceiling is repeating what the sheet says. A buyer discovering the income test three weeks into a contract is discovering it exactly when the file does. The only expensive version is the one where both find out late.

Our read

The title is the read. A price is decided by people, and people can be persuaded, anchored, and out-waited. The loan is computed from an operating statement, and arithmetic does not negotiate. Confusing the two is how a closing table ends up short six figures with everyone acting in good faith.

The preparation that actually pays, on either side of the table, is running the three tests on the real statement before any number is spoken. Ten minutes of arithmetic names the binding test, prices the marginal dollar, and predicts who will show up. When the debt supports the ask, the bids arrive shaped like financing. When it does not, the bids arrive shaped like equity, fewer and slower and better capitalized. A seller who knows that in advance reads the offer stack correctly instead of reading it as weak demand.

For a seller, the useful question before pricing is what loan this statement supports at today’s rates, because that number is where the financed bids will cluster. An offer above it is a statement about the buyer’s equity, worth reading as one. For a buyer, the discipline is underwriting the lender before underwriting the seller: the term sheet in hand, the three limits named, and the plan built on the smallest answer rather than the familiar percentage.

Who to ask

The sizing itself belongs to a lender, and the document that answers is the term sheet. A mortgage banker or the lender’s originator will run the actual statement against the actual tests in a day, and the useful question to bring them is not what rate, it is which test binds at today’s rate. The answer to that question is the shape of the whole deal.

The insurance number belongs to a licensed insurance agent, and it belongs in the file before the offer rather than after, because in Miami it is one of the two lines that moves the underwritten income the most. The quote, not last year’s premium, is what the file will carry.

The tax line is public arithmetic. The Miami-Dade Property Appraiser’s records show the current assessment, and what a sale does to it is the subject of our tax reset piece, linked below.

What we do is read the building and its statement the way the file will, and say which test binds before anyone prices anything. We last checked the cited Fannie Mae term sheet on August 10, 2026. Term sheets move with the market, and the sheet in force on the day of an application is the one that governs.