Most guidance on how to underwrite commercial real estate is written as a spreadsheet tutorial: build the rent roll, net the expenses, apply a cap rate, layer in debt, solve for return. That sequence is correct and beside the point. Underwriting is not arithmetic. It is a chain of five judgment calls, and the model is the machinery that carries them to a number. Models rarely fail because someone summed a column wrong. They fail because one of the five decisions was wrong, and the spreadsheet restated that error to two decimal places.
Key Takeaways
Underwriting a commercial property is five decisions: in-place income, expenses at your operating standard, what the submarket supports, what the capital structure costs and demands, and what the exit is worth.
Model errors are decision errors. Arithmetic mistakes get caught by a second reader. Bad assumptions get formatted, printed, and defended.
Errors do not add across the chain, they multiply. An 11 percent income error combined with an 8 percent exit cap error produces roughly an 18 percent error in exit value.
The seller's expense history describes the seller's operation. Your tax basis after reassessment and your management fee are the numbers you will pay.
Debt sized off unverified income converts income error into a larger equity check at closing.
What does it mean to underwrite a commercial real estate deal?
Underwriting a commercial real estate deal means making five judgment calls: what the in-place income is, what the expenses are at your operating standard, what the market supports on rent and absorption, what the capital structure costs and demands, and what the exit is worth. The model performs the arithmetic that carries those five decisions.
Framing it this way changes what you check. If underwriting is a procedure, quality control means auditing formulas. If it is a set of decisions, it means auditing the evidence behind each one and who supplied it.
Decision | The question | Who usually supplies the answer | Failure signature |
|---|---|---|---|
1. In-place income | What does it collect today? | Seller rent roll and OM | One-time income treated as recurring |
2. Expenses | What will it cost at my standard? | Seller operating statements | Taxes not reassessed, self-management carried over |
3. Market | What do rent and absorption support? | Broker comp set | Metro averages used for submarket reality |
4. Capital structure | What does debt cost and require? | Lender term sheet | Proceeds sized off unverified income |
5. Exit | What is the residual worth? | The underwriter | Exit cap set equal to entry cap |
Note column three. Four of the five decisions are seeded by someone else's document. Only the exit is yours, which is why it produces the largest single swing in value.
How do you verify in-place income and expenses when underwriting a property?
Verify income against the trailing 12 months of collections and the current rent roll, not the offering memorandum. Then rebuild expenses at your operating standard: your tax basis after reassessment, your insurance quote, your management fee. The seller's expense history describes the seller's operation, not the one you will run.
The income side fails in a specific way. Gross potential rent is rarely fabricated, because it is checkable. What gets inflated is the bridge from gross potential to effective gross: a vacancy factor below trailing actuals, concessions netted out of view, and other income carrying items that will not repeat. A lease termination fee, an insurance settlement, and a utility rebate all deposit into the same account as rent. None is income you can underwrite.
The expense side fails differently. Most sellers are not hiding costs. They are reporting costs that are true for them and false for you. A long-held asset carries a tax assessment anchored to an old basis, and in many jurisdictions the sale resets it. A self-managing owner books management at a fraction of a third-party fee. None of that is deception, and all of it lands on your statement.
The test: for every line, ask whether it describes the property or the owner. Reset the second category. See what a T-12 reveals about the seller's story for how these items surface month by month.
How do you decide what the market supports on rent and absorption?
Decide it from the submarket, not the metro. Market rent is what a comparable space in comparable condition within a comparable commute leased for in the last 90 days, and absorption is how long it took. A rent conclusion without a lease-up schedule attached is half an assumption.
Rent and absorption are one decision, not two, and separating them is the most common way this call goes wrong. An underwriter concludes market rent is 8 percent above in-place, marks the pro forma to that number, and assumes the gap closes on the natural expiration schedule. The comp set may support the rent. It says nothing about the vacancy and concession it took to get there.
Two questions discipline the decision. What is the leasing velocity in this submarket for this product, expressed in units or square feet per month rather than a stabilized occupancy percentage? And what did the concession package look like on those comps, because a rent achieved with three months free is not the rent in the model? Submarket data beats metro data on every underwriting decision, and rent is where the difference is most expensive.
Volume determines how thin your comp set is allowed to be. The Mortgage Bankers Association forecast in February 2026 that total commercial mortgage originations would rise 27 percent to 805.5 billion dollars in 2026. In a market that active, a three-comp conclusion is a choice.
What does the capital structure decision cost, and what does it demand?
Debt costs a coupon and demands a covenant. The coupon sets your cash flow. The covenant, expressed as debt service coverage, loan-to-value, and debt yield, sets your proceeds and your behavior for the hold. Sizing debt off an income figure you have not verified transfers the error straight into your equity check.
Work the mechanics. Your verified NOI is 920,000 dollars. At a 1.25x coverage floor the loan can carry 736,000 dollars of annual debt service. At 6.5 percent on a 30-year amortization schedule the annual constant is roughly 0.0758, so proceeds come to about 9.7 million dollars. On the seller's NOI of 1,032,000 dollars, the same covenant appears to support 10.9 million. The lender underwrites the verified number, and the 1.2 million dollar difference becomes equity you find between signing and closing.
Credit conditions do not remove this discipline, they change how visible it is. The Federal Reserve reported in its July 2026 Senior Loan Officer Opinion Survey that banks generally eased standards on commercial real estate loans, the widest easing since early 2022. Easier credit means a wrong income number is more likely to get funded, not less likely to be wrong. See how lenders size a loan.
How much does a small underwriting error change the value at exit?
More than operators expect, because errors multiply rather than add. Value at exit equals exit NOI divided by exit cap rate, so an 11 percent NOI error combined with an 8 percent cap rate error produces roughly an 18 percent value error. On a mid-size deal, that spread can exceed half the equity in the transaction.
Worked from stated inputs on a 17 million dollar acquisition. Gross potential rent is 1,800,000 dollars with vacancy and credit loss at 6 percent, giving effective gross income of 1,692,000 dollars. The offering memorandum adds 40,000 dollars of other income from a lease termination fee received last year. Seller expenses are 700,000 dollars; reassessment adds 55,000 dollars of tax and a third-party management fee adds 17,000 dollars. Both cases grow NOI 3 percent annually to a year-five exit, the optimistic at the entry cap of 5.75 percent and the disciplined at 6.25 percent.
Line | Optimistic case | Disciplined case |
|---|---|---|
Effective gross income | 1,732,000 | 1,692,000 |
Operating expenses | 700,000 | 772,000 |
Year 1 NOI | 1,032,000 | 920,000 |
Year 5 NOI at 3% growth | 1,161,500 | 1,035,500 |
Exit cap rate | 5.75% | 6.25% |
Exit value | 20,200,000 | 16,568,000 |
The gap is 3,632,000 dollars. No line in that table is outrageous. Isolated, each is a rounding argument. Chained, they consume 18 percent of exit value. At 65 percent loan-to-value, equity here is 5,950,000 dollars, so the swing is roughly 61 percent of the money at risk, produced by four defensible-sounding judgment calls and zero arithmetic errors.
A model is not an opinion. It is the arithmetic that carries five opinions, and it will carry a wrong one with exactly the same precision as a right one.
Two habits follow. Run the chain in both directions, because sensitivity analysis separates a model from a guess by showing which decision your return depends on. And treat the residual with suspicion, since the exit cap makes or breaks the return and no document supports it.
Frequently Asked Questions
What are the steps in the commercial real estate underwriting process?
Build the rent roll, normalize the operating statement, set market rent and absorption, size the debt, and solve for return at an exit assumption. Each step encodes a judgment call, and the judgment determines whether the output is worth anything.
Should the exit cap rate be higher than the entry cap rate?
In most underwriting, yes. The asset is older at sale, and setting the exit equal to the entry assumes the market prices it identically after a full hold. Expanding it by 25 to 50 basis points is a common convention, though the right number depends on asset age, plan, and entry pricing.
Why do underwriting models overstate value so consistently in one direction?
Because four of the five decisions are seeded by seller or broker documents, and those present the highest defensible number on every line. The errors that survive a casual read are the ones that inflate value.
Conclusion
Underwriting a commercial property is not a spreadsheet exercise with judgment sprinkled on top. It is five judgment calls with a spreadsheet attached to compute their consequences: in-place income, expenses at your standard, submarket rent and absorption, the cost and covenant of the capital structure, and the exit. Get those right and an ordinary model produces a defensible number. Get one wrong and the finest model in the market will render the error in clean formatting and multiply it against the other four.
The operator's discipline is to stop auditing formulas and start auditing decisions: for each of the five, name the evidence, who produced it, and what you changed. That record is the underwriting. Everything else is arithmetic.