How do you size agency debt in a multifamily underwriting model?

Run both constraints separately, maximum proceeds at the stated LTV (typically 60–65%) and maximum proceeds implied by a 1.25x DSCR floor at the lender's constant, and take the lower of the two. On stabilized, lower-leverage deals the LTV constraint usually binds; on thinner-margin deals the DSCR floor often binds instead.

What's the difference between GPR and effective rental income in a multifamily model?

Gross potential rent (GPR) is the sum of market or in-place rent across every unit at full occupancy. Effective rental income subtracts vacancy loss, loss-to-lease, and concessions from GPR. Models that skip this bridge and use GPR as if it were collectible revenue overstate NOI from the first line.

How should other income be modeled in a multifamily pro forma?

Build it at a $/unit/year run rate tied to actual trailing data (RUBS reimbursement from the utility billing register, fees from the trailing ledger) rather than as a flat percentage of rental income. A percentage-of-rent shortcut tends to overstate other income on lower-rent assets and understate it on higher-rent ones.

Why does the exit cap matter more than the entry cap in a 10-year hold?

Terminal value is the largest single driver of levered IRR on a long hold, and it's set entirely by the exit cap assumption on Year 11 NOI. Underwriting standard is to hold the exit cap flat or 15–50 bps above the entry cap, assuming compression prices in a market recovery the buyer doesn't control.

What is a real estate underwriting model?

A structured spreadsheet that projects a property's cash flow, financing, and returns under explicit assumptions — the basis for every institutional acquisition decision.

Multifamily · Underwriting Model

Multifamily Underwriting Model

How to build a multifamily underwriting model: unit-level rent schedule, loss-to-lease, RUBS, agency debt sizing, exit cap. Free template.

Get the Excel template Automate it with Moraine

Why the underwriting model looks different for multifamily

A multifamily model is built bottom-up from a unit-type rent schedule, not a single blended rent line. Loss-to-lease has to be bridged from the trailing rent roll to in-place GPR before you can even start the pro forma, because renewals reset at market rather than stepping on a fixed schedule.

Agency debt (Fannie Mae, Freddie Mac) sizes off the lesser of a stated LTV and a 1.25x DSCR floor, and on stabilized deals the LTV constraint usually binds first, which caps proceeds well below what a DSCR-only sizing would suggest. Office and retail debt sizing runs the opposite way more often.

Other income (RUBS reimbursement, pet fees, parking, application fees) has to be built at a $/unit/year run rate tied to the trailing utility billing register rather than assumed as a flat percentage of rent. Overstating it is how multifamily models most often inflate NOI.

The multifamily-specific checklist

1

Build the rent roll

Verify: unit- or tenant-level detail with rent, term, recoveries
2

Project market rent growth

Verify: submarket-specific; avoid flat nationals assumption
3

Model recoveries explicitly

Verify: CAM, tax, insurance broken out, not blended
4

Run sensitivity on exit cap

Verify: ±75 bps and stress the downside
5

Build debt sizing

Verify: LTV and DSCR constraints both checked
6

Compute leveraged IRR and equity multiple

Verify: cash flow to equity, not enterprise
7

Build a unit-type rent schedule anchored to the trailing rent roll

Verify: modeled average effective rent matches T-12 collections within 2%
8

Bridge loss-to-lease from in-place rents to market rents by unit type

Verify: bridge reconciles to total in-place GPR within 0.5%
9

Model RUBS and other income at a $/unit/year run rate

Verify: tied to trailing 3-month utility billing register, not a flat % of rent
10

Build controllable vs. non-controllable opex from the T-12 with per-unit benchmarks

Verify: flag any line item more than 10% off comp-set per-unit average
11

Size agency debt off the lesser of stated LTV and 1.25x DSCR

Verify: confirm which constraint binds before locking loan proceeds
12

Set the exit cap 15–50 bps above the entry cap

Verify: cross-checked against comparable-vintage exit cap surveys for the submarket

Metrics that matter for multifamily

Metric	Target	Calculation
Going-in cap rate	5.0–5.5%	Year 1 NOI / purchase price
Debt yield	>7.5%	Year 1 NOI / loan amount
Loss-to-lease capture	5–10% of GPR	(market rent − in-place rent) / in-place rent, by unit type
Economic vacancy	5–7%	1 - (effective rent / gross scheduled rent)
Expense ratio	35–45% of EGI	opex / EGI
Break-even occupancy	<85%	(opex + debt service) / GSR

Red flags unique to multifamily

Other income modeled above $500/unit/year without a bulk program in place

RUBS plus fees rarely clear $400/unit without bulk internet or covered parking already implemented, inflated other income quietly lowers the effective cap rate
Levered IRR relies on exit cap compression below the entry cap

assuming cap compression across a 10-year hold reverses market-cycle risk onto the buyer instead of pricing it
Debt sizing shown only against max LTV, with no DSCR check

agency lenders size to the lesser of the two constraints, skipping the DSCR test overstates proceeds and understates the required equity check
Loss-to-lease exceeding 8%

signals stale leases or mismanaged rent growth; the in-place roll understates market
Concessions over 1 month average

demand weakness the broker's pro forma almost never prices in
RUBS recovery under 60%

expense inflation flows directly to NOI because recoveries are capped below market
Single-month trailing collections under 96%

rising delinquency typically precedes a 2-3% NOI miss within two quarters

Example — multifamily underwriting model walkthrough

Walk through a 220-unit Class B multifamily asset in Dallas priced at $55.0M, $250,000 per unit. The model starts with a unit-type rent schedule: average in-place rent of $1,750/month across the roster produces gross potential rent (GPR) of $4,620,000. Vacancy loss at 6% and loss-to-lease at 3% strip out $415,800, and modeled concessions at 1% remove another $46,200, leaving net rental income of $4,158,000.

Other income is built next, not assumed. RUBS reimbursement at $185/unit/year and ancillary fees at $325/unit/year add $110,700, bringing effective gross income to $4,270,200. Opex is split controllable ($3,100/unit: payroll, marketing, repairs) and non-controllable ($2,850/unit: taxes, insurance), plus a 3% management fee and a $300/unit replacement reserve, for $1,503,106 in total. Year 1 NOI lands at $2,767,094, a 5.03% going-in cap rate on the $55.0M basis, computed on a fully loaded expense base with the replacement reserve inside opex rather than below the line.

Debt sizing runs both constraints. At 65% LTV, proceeds cap at $35.75M. The 1.25x DSCR test (max debt service of $2,213,675 at a 5.9% constant) implies $37.5M of proceeds. The LTV constraint binds first, so the loan is $35.75M; defaulting to the DSCR-implied number would overstate proceeds and understate the required equity check by $1.75M. Loan sizing therefore gets rechecked whenever NOI moves, since a revised Year 1 figure can flip which constraint binds.

The exit assumption closes the loop. Holding 10 years with 3% annual rent growth carries Year 1 NOI to $3,718,800 by Year 11. Setting the exit cap 25 bps above the 5.03% entry cap, at 5.28%, produces a terminal value of roughly $70.4M. A model that instead assumes cap compression to 4.75% would inflate terminal value by nearly $4M and mask a materially weaker levered IRR under a flat-cap scenario. The same discipline applies down-column: vacancy at 6%, loss-to-lease at 3%, and concessions at 1% are each tied to the trailing rent roll rather than penciled in to hit a target return.