How to underwrite a multifamily deal.

01Gather the documents

Three documents drive everything: the rent roll (a unit-by-unit snapshot of every lease — rent, status, move-in and expiration dates), the T-12 (the trailing-twelve-months operating statement showing what the property actually collected and spent, month by month), and the offering memorandum for context on the story, the unit mix, and the submarket.

Reconcile before you model. The rent roll's total scheduled rent should tie to the T-12's rental income line within a few percent; if it doesn't, find out why before going further — concessions, delinquency, or a stale rent roll all change the deal. And treat the OM's pro forma as marketing: the seller's numbers are the starting point for questions, never the inputs.

02Build the revenue picture

Start from gross potential rent — every unit at market rent, fully occupied — then walk down to what the property actually collects. The two deductions that matter: loss to lease (in-place rents below market, recovered only as leases roll) and vacancy (physical plus economic — use the submarket's stabilized rate, not the seller's best month). Then add other income: fees, laundry, parking, pet rent, utility reimbursements.

That bottom line is effective gross income, and it is the base for everything downstream — including the management fee, which is why building it correctly matters twice.

03Normalize the expenses

The T-12 tells you what the seller spent; underwriting asks what you will spend. Go line by line: keep the operating history where it's credible (utilities, contract services), and replace it where ownership changes the number. The three lines that change most:

• Property taxes — most jurisdictions reassess at sale. Underwrite on your price and the local millage, not the seller's assessed value. In the example: about $216,000 (1.2% of the $18.0M purchase).
• Management fee — a percentage of collected income, typically 2.5–4% of EGI. Here: 3% × $2,000,400 ≈ $60,000.
• Insurance — get a quote; trailing premiums rarely survive a sale, and in some states they have doubled in two years.

Convention note: replacement reserves (commonly $250 per unit per year) sit below NOI in most institutional models but inside expenses in some lender models. Either is fine — just be consistent, and know which one a quoted cap rate implies.

04NOI and the going-in cap rate

Net operating income is the property's earnings before debt and capital items, and the going-in cap rate — year-one NOI over purchase price — is its price tag: $1,100,400 ÷ $18,000,000 ≈ 6.1%. Compare it to recent trades in the submarket. A purchase below the market cap rate needs a story (true upside you can execute); above it, ask what the market knows that you don't.

05Size the debt

The loan is sized by whichever constraint binds first: loan-to-value (typically 60–70% for agency debt) or DSCR — NOI over annual debt service, with lenders generally requiring at least 1.20–1.25×.

In the example: 65% LTV puts the loan at $11.7M. At a 6.25% rate with 30-year amortization, annual debt service runs about $864,000 — so DSCR is $1,100,400 ÷ $864,000 ≈ 1.27×. The LTV constraint binds, the coverage clears, and the equity check is $6.3M plus closing costs and any capital budget. If DSCR had come in under 1.25×, the loan shrinks until it clears — which is how a debt market quietly reprices a deal the seller still thinks is worth asking.

Watch the interest-only trap: an IO period boosts early cash-on-cash, but coverage is quoted on the amortizing payment. Model both, and check the DSCR again at the exit refinance if your plan depends on one.

06Project the hold and the exit

Now run the operating model forward — month by month, not annual shortcuts: rents grow (2–3% is a defensible stabilized assumption; double-digit submarket growth is not a plan), expenses inflate, loss to lease burns off as leases roll, and any renovation program phases in unit by unit with its own downtime.

The exit is its own underwrite. Apply an exit cap rate to the year after sale's NOI — and underwrite it at or above your going-in cap, because betting on multiple expansion is timing, not skill. In the example: NOI growing ~3% a year reaches about $1.28M by year six; at a 6.5% exit cap the sale is roughly $19.6M, less ~2% selling costs, less the loan balance. Most of the equity's profit arrives in that single cash flow, which is exactly why the exit cap deserves the most skepticism of any assumption in the model.

07Run the equity waterfall

Project-level returns aren't what the LP earns. Distributions flow through the waterfall: preferred equity (if any) gets its coupon first, then LPs earn their preferred return (commonly 7–9%), then the sponsor's promote kicks in — for example 80/20 above an 8% IRR hurdle, stepping to 70/30 above 14%.

Two discipline checks. First, the waterfall must reconcile: every tier's distributions, summed, must equal levered cash flow to the dollar — by construction, not by a plug. Second, look at the GP/LP split of total profit at the base case and one downside case; a structure that pays the sponsor well in scenarios where the LP loses money will not survive diligence.

08Read the returns — then stress them

Three numbers summarize the deal, and they discipline each other: IRR (time-weighted — rewards getting capital back sooner), equity multiple (total cash out over cash in — indifferent to timing), and cash-on-cash (annual distributions over invested equity — what the hold feels like year to year). A high IRR with a thin multiple is a quick flip; a fat multiple with a low IRR is patience priced in. Value-add deals in a normal market tend to pencil in the mid-teens IRR with a 1.8–2.2× multiple over five years.

Then stress everything. Build a sensitivity grid — exit cap on one axis, rent growth on the other, with each cell a full re-run of the model — and find where the deal breaks: the exit cap that pushes IRR below the pref, the vacancy that breaks 1.0× DSCR. If the base case sits one notch from a cliff, the market is telling you the price already assumes your upside.

Common mistakes that sink underwriting

Glossary — the terms that matter