How do you underwrite a multifamily deal? You analyze the property’s income, expenses, debt, and value to see if it meets your criteria. You end with one number: the most you can pay and still hit your returns.
In 2015 I was stationed in Hawaii, active duty Air Force, underwriting a multifamily deal in Michigan from a laptop after my shift. I couldn’t drive by the property. All I had were the documents and the will to rebuild every number myself. I bought 20 units that way, 4,000 miles away, and the deal worked because the underwriting was done before I wired a dollar.
One thing before we start. Underwriting is not a tool for making a deal work. It’s a tool for killing bad ones. Most deals you look at should die on your desk. That’s the system working.
I’ll walk you through how to underwrite a multifamily deal using one real example: a 50-unit Class B property, built in 1985, asking $4.5 million. Every number below comes from that same deal. By the end you’ll see why the broker called it a 7 cap, and why my math said otherwise.
Know Your Buy Box, Then Screen Fast
Set your criteria before any deal is on the table. Market, unit count, class, return targets, max price per unit. When you fall for a property, you’ll bend the numbers to make it work. Your buy box is the referee, not you.
You also don’t owe every deal four hours. Screen it in 30 minutes with rough numbers. Ask one question: does this deal earn more of my time? Most don’t. The full underwrite below is for the ones that pass.
Step 1: Get the Right Documents
Two documents start every multifamily deal. The T12 is the trailing 12-month operating statement. Get the monthly version. A yearly roll-up hides the games sellers play before a sale. The rent roll is a unit-by-unit snapshot dated within 30 days: rent, lease dates, deposits, delinquency.
The T12 tells you what happened. The rent roll tells you what’s happening now. Your model tells you what happens next, under you. Also get tax bills, 24 months of utility bills, the insurance page, capex history, and current leases. Then verify: bank statements to confirm rent hits the account, and the seller’s tax return against the T12. Occupancy gets inflated before a sale more than people think.
If the seller can’t produce clean records, don’t walk right away. Some of my best deals were mom-and-pops with books in a shoebox. But price the risk. I add 50 to 100 basis points to my cap rate until the gaps close.
Step 2: Read the T12 Like a Detective
The T12 shows deferred maintenance before you ever tour the property. Here’s what I look for:
- Repairs under $600 to $900 per unit per year on a stabilized 1980s building. Lower usually means the owner starved the property to dress up the books.
- A capital cost hiding in the expense line. A $28,000 “repair” in month seven is a roof job, not a repair.
- Insurance that hasn’t moved in years. Nobody re-quoted it. Yours won’t be flat.
- Utility bills climbing with no rate hike. That’s a leak.
Our 50-unit deal: repairs showed $360 a unit and insurance hadn’t changed in three years. Two flags before I left my desk.
Step 3: Verify Income and Vacancy
Never start with the seller’s pro forma. It’s a sales pitch. Build income yourself from the rent roll: 30 two-beds at $950 plus 20 one-beds at $775 comes to $528,000 a year in-place.
Then check it. Rent roll times 12 should roughly match the T12’s collections. A gap past 5 percent has four suspects: vacancy, concessions, bad debt, or a stale rent roll. Make the broker name which one.
Now pull market rents from three sources: data services, live listings, and calls to nearby properties. Use net effective rent, not the asking price; a unit at $1,050 with a free month collects about $960. Our deal supported $1,050 and $850 at market, so the gap between in-place and market rent, the loss to lease, sits at $54,000 a year. That’s the upside, and Step 7 shows how to model it honestly.
Last, set your vacancy loss. Use 5 to 7 percent for a stable property, 7 to 10 for average conditions, 10 to 15 for heavy value-add. Never go below 5, no matter what today’s rent roll says. And use submarket data, not the metro average. A metro at plus 2 percent can hide a submarket at minus 5 if new supply just landed nearby.
Our deal: 7 percent vacancy, or $37,000 off the top. With $30,000 of verified other income, Effective Gross Income lands at $521,000.
Step 4: Rebuild the Expenses Yourself
This is where first-time buyers get hurt, so I won’t sugarcoat it. The seller’s numbers describe the seller’s ownership. Rebuild every line for yours:
- Management: 3 to 5 percent of income, even if you plan to self-manage. The seller’s labor was a real cost that never hit the books.
- Repairs: $600 to $900 per unit for this vintage. The seller showed $360. I used $700.
- Insurance: get your own quote. Ours came back 36 percent above the seller’s premium.
- Capex reserves: $250 to $400 per unit per year, held separate from operating costs.
And the biggest one: property taxes. Most states reassess at your purchase price once you close, and keeping the seller’s old tax bill in your model is the costliest mistake in this business. The fix is one phone call: give the county assessor your expected price and ask for a post-sale estimate. Ours moved from $38,000 to $58,000.
A quick honest note. Every article quotes different expense benchmarks, because a number without property class and location attached is a guess. Get real local quotes on the big lines. The sanity check: total expenses usually run 40 to 55 percent of income for Class B and C. A seller at 35 percent on an old building is leaving something out.
Our deal: the seller’s T12 claimed $208,000. My rebuilt number: $265,000.
Step 5: Calculate NOI and Value the Deal
NOI is income minus operating expenses, nothing else. Run it twice: in-place on today’s rents and your rebuilt expenses, and stabilized once your plan lands. You pay for in-place. You get paid for stabilized. A buyer who pays for stabilized NOI is paying the seller today for work he hasn’t done yet.
Our deal: $521,000 minus $265,000 is $256,000 of in-place NOI. Comps trade around a 6.5 cap, so value comes to $256,000 divided by 0.065, about $3.94 million. The ask is $4.5 million.
Now look at the broker’s math. Same income, his $208,000 expense line, gives $313,000 of “NOI.” Against his ask, that’s a 7 cap on the flyer. On my rebuilt numbers, the same price is a 5.7 cap. Same building. The only difference is whose numbers you trusted. One more rule for any multifamily deal: set your exit cap 25 to 100 basis points above your entry cap. Never bet on the market getting better.
Step 6: Model the Upside Honestly
Two rules keep the upside real. First, loss to lease rolls off on the lease calendar, not yours. A full mark-to-market takes 12 to 18 months, and not every tenant stays. I modeled $45,000 of our $54,000 actually landing.
Second, renovation premiums show up late. Renovate 30 units over 15 months chasing a $125 bump, and the unit you finish in month 14 barely earns anything in year one. Expect 40 to 50 percent of the premium in year one, with the full number arriving after the program wraps. A pro forma showing full premiums on day one was written by someone selling something. Budget the renovation 15 to 30 percent over, because that’s how it usually goes.
Step 7: Model the Debt
Two tests size your loan. DSCR, your income over annual debt payments, needs 1.25x minimum, on in-place NOI, at the lender’s stressed rate. LTV caps how much they’ll lend against value. The smaller of the two numbers wins.
Our deal at my $3.9 million price: a 65 percent loan of $2.55 million at 6.5 percent runs about $193,000 a year. DSCR comes to 1.32x. It passes. Cash-on-cash on roughly $1.5 million of equity lands near 4 percent in year one. Modest, and honest. Value-add deals often start there and build. A lot of people ran heavy bridge debt into 2022 and got burned pretty bad. The ones on lower, fixed-rate debt are still standing.
Step 8: Stress-Test It, Then Break It on Purpose
Run the tests one at a time first: rents down 5 to 10 percent, expenses up 10 to 20, vacancy at 12, exit cap up 100 basis points. Then run them all together as the worst case. The deal doesn’t need to shine there. It needs to cover its debt. One that breaks even in the worst case survives a downturn. One that misses payment is a capital call waiting on a closing date.
The most useful tool I use is a simple grid: rent decline across the top, expense increases down the side, DSCR in every box. Find the box where coverage drops below 1.0x, then ask how far away it sits. Our deal holds until rents fall 10 percent while expenses climb 8 at the same time. That’s real room. If it broke at rents down just 4 percent, I’d walk, because down 4 is just an ordinary bad year.
Step 9: The Buy Right Decision
This is the first pillar of the Buy Right, Finance Right, Manage Right framework: you make your money the day you buy. My go or no-go rules for any multifamily deal:
- Your underwriting sets the max price. The seller’s asking price is irrelevant.
- Set your hurdles first: for value-add, a mid-teens levered IRR, 1.6 to 2x equity multiple over five years, 1.25x coverage from day one.
- Negative leverage is an automatic no. If your cap rate sits below your interest rate, the debt eats your return from the start.
- Plan two ways out. A deal that only works with a refinance in two or three years is a bet on the debt markets cooperating on your schedule. Plenty of people made that bet in 2021. Ask them how it went.
The verdict: worth $3.9 million to me, asking $4.5 million. I offered my number with the backup to defend it. The seller said no, so I passed. That’s not a loss. That’s the tenth deal this quarter that died on my desk, and it cost me a few hours instead of a few hundred thousand dollars.
The Metrics Cheat Sheet
| Metric | Formula | What I look for |
|---|---|---|
| Gross Potential Rent | Units × rent, yearly | Figure in-place and market rent separately |
| Loss to lease | Market rent minus in-place rent | Capture over 12 to 18 months, not day one |
| Vacancy loss | % of gross rent | 5 to 7% stable. Never under 5% |
| Expense ratio | Expenses ÷ income | 40 to 55% for Class B/C. Under 35% means missing lines |
| NOI | Income minus expenses | Buy on in-place. Project on stabilized |
| Cap rate | NOI ÷ price | From closed comps, on post-sale taxes |
| Exit cap | Entry cap + 25 to 100 bps | Never bet on the market improving |
| DSCR | NOI ÷ debt payments | 1.25x minimum, on in-place NOI |
| Cash-on-cash | Year-1 cash flow ÷ equity | Low year one is normal for value-add |
| Capex reserve | Per unit, per year | $250 to $400. Renovation budget is separate |
Want the full breakdown of cap rate versus cash-on-cash versus IRR? Read our metrics guide. [LINK to Topic 4]
Build the Skill Before You Buy
Underwrite 20 deals cold before you offer on one. No pressure to close, just the rep. Write down your max price each time, then watch what the property actually sells for. That gap between your number and the market teaches more than any course.
When you’re ready to buy, start smaller than you think. Go buy a 5 unit or a 10 unit first. Sit in every seat it takes to run the place. The tuition is cheaper at that size, and the discipline you build there protects your family’s money for the next 30 years.
I built this whole process into a one-page checklist: every document, every benchmark, every stress test, in order. Grab it here. [LINK to checklist download] And if you want to learn how to underwrite a multifamily deal next to people doing it every week, our students have closed 90,000+ units and over $5 billion in deals. Apply when you’re ready to put in the work. [LINK to community application]
FAQ
How long does it take to underwrite a multifamily deal? A first screen takes 30 minutes. A full underwrite with stress tests takes 2 to 4 hours once you have the documents. If a deal fails the screen, stop there.
What DSCR do lenders want? Most agency lenders want 1.25x minimum, sized at a stressed rate. I underwrite to 1.25x on in-place NOI, not on projections.
What’s a normal expense ratio for a multifamily deal? 40 to 55 percent of income for Class B and C properties, higher for older buildings. Under 35 percent almost always means missing lines.
How do I pick the right cap rate? Closed sales of comparable properties in the same submarket, checked against broker reports, then recalculated on post-sale property taxes.
How much should I set aside for capex? $250 to $400 per unit per year in reserves, kept separate from operations, with renovations budgeted on top and assumed to run 15 to 30 percent over.er unit per year in reserves, below the NOI line, with renovations budgeted separately and assumed to run 15 to 30 percent over.
