How to Underwrite a Multifamily Acquisition: A Step-by-Step Guide

A first-year analyst hands you a multifamily underwriting showing a 6.8 percent going-in cap rate and a 12 percent unlevered IRR. The numbers come off the broker pro forma. Property taxes are flat at the current owner's basis. Insurance is unchanged. Rent growth is 4 percent per year for five years. Loss to lease is zeroed out at month 13. Every assumption is the most optimistic one available. The model clears the hurdle rate because it was built to clear the hurdle rate.

This article covers the two source documents every underwriting starts with, how to project revenue without inheriting the seller's assumptions, the expense lines that get understated most often, how to compute NOI and handle CapEx, and the downside stress tests to run before committing.

For the broader mechanics of building an auditable real estate model, see TILT's real estate pro forma guide.

What Documents Do You Need Before You Start Underwriting?

Multifamily underwriting begins with two documents from the seller: the trailing 12-month operating statement (T12) and the current rent roll. Everything else gets built on top. Without both, the underwriting is a guess.

The T12 is the property's profit and loss statement covering the most recent twelve months. It shows historical income line by line and operating expenses line by line, month by month. The right format is monthly, not annual. A monthly T12 lets you see seasonality, one-time spikes, and gaps that the annual roll-up hides.

The rent roll is a snapshot of the property's current rental status as of a specific date. Every unit, unit type and square footage, current tenant name, lease start and end dates, current contract rent, market rent, security deposit on file, concessions, and any delinquency. The right format is a unit-level export from the property management system, not a summary by floor plan.

Read together, the two documents give you history and current status. The underwriting job is to project forward from them without inheriting the seller's assumptions.

How Do You Read a T12 for Deferred Maintenance Signals?

The T12 reveals deferred maintenance before any property tour does. Look for repairs and maintenance line items that are abnormally low, abnormally lumpy, or trending down at a rate that does not match a stabilized asset.

Specific signals to flag when reading a T12:

• Repairs and maintenance that do not reconcile to condition. Compare the T12 with work orders, invoices, age of systems, inspection findings, and the property condition assessment. A per-unit benchmark without asset age, climate, staffing, and accounting policy is not a reliable normalization.
• Make-ready costs that look reasonable but turnover is high. Compare turns, days vacant, scopes, invoices, concessions, and vendor bids. Use the subject property's observed cost per turn and a documented renovation scope rather than a universal per-turn range.
• Utility expenses trending up unevenly month over month. Reconcile invoices to rates, consumption, weather, meter configuration, and service changes. Ask for the last 24 months of utility invoices, not only the T12 roll-up.
• Lumpy capital items inside the operating expense line. A 28,000 dollar charge sitting inside "repairs and maintenance" in month seven is usually a partial roof replacement classified to keep CapEx off the books. Pull those items out before computing normalized expenses.
• Insurance held flat year-over-year. A flat T12 premium is not a quote for the buyer. Obtain a current indication or binder from a broker using the property's construction, loss history, catastrophe exposure, deductibles, and intended ownership structure.

The T12 is also a stress test for the rent roll. Compute trailing collections from the T12 and divide by the rent roll's Gross Potential Rent. If trailing collections are 91 percent of GPR and the seller claims a 95 percent economic occupancy, the rent roll has not been reconciled to the bank deposits.

How Do You Project Revenue Like a Professional?

Build revenue from the ground up: Gross Potential Rent, less loss to lease, less vacancy and credit loss, plus ancillary income equals Effective Gross Income. Each line is its own assumption and each assumption must be defensible against verifiable comps.

Gross Potential Rent (GPR). The total rent the property would generate if every unit were leased at full market rates with zero vacancy. Start by computing in-place GPR from the rent roll (sum of contract rents annualized). Then compute market GPR using the property's market rents, which should be triangulated from three sources: CoStar or Yardi Matrix submarket data, on-the-ground comps from a recent shop of 5 to 10 competing properties, and current asking rents on Apartments.com and Zillow for comparable units. Do not use a single source. A submarket-wide rent average is not the same as the rent your specific unit mix at your specific property quality can achieve.

Loss to lease. The difference between current in-place rents and market rents. If contract rents average 1,800 dollars per month and verified market rents support 2,000 dollars per month for comparable units, loss to lease is 200 dollars per unit per month or 2,400 dollars per unit per year. This is a potential value-add opportunity, not guaranteed upside. Model roll-off from the property's lease expirations, renewal history, turns, downtime, concessions, and verified effective rents, not a fixed calendar.

Vacancy and credit loss. Do not use one combined allowance as a universal standard. Reconcile physical occupancy, economic occupancy, collections, delinquency, skips, evictions, concessions, downtime, and the rent roll. Freddie Mac reported 5.1 percent vacancy for its multifamily series in Q1 2026, while Census reported 7.3 percent for all rental housing. The difference illustrates why the property, market definition, and data series matter. Use a documented subject-property base case, then stress the components separately.

Ancillary and other income. Often understated in seller pro formas. Build each line from the property's records and legal operating plan:

• Parking: Use signed leases, current ledgers, physical counts, and comparable properties with similar parking supply. Separate occupied, reserved, covered, and free spaces.
• Pet rent and pet fees: Use the executed lease schedule and trailing collections. Check local fee limits and fair-housing compliance before assuming new fees.
• RUBS and utility reimbursements: Use the last 24 months of bills, the proposed allocation method, collections, lease language, and current state and local rules. Do not treat a projected reimbursement as NOI until it is legally collectible and operationally supported.
• Laundry, application, administrative, late, storage, and amenity income: Model each from the rent roll, fee schedule, historical collections, unit and amenity counts, and a conservative collection rate.

Do not plug in a national ancillary-income range. The result of these property-specific lines is Effective Gross Income (EGI), the actual revenue the property can generate under documented operating conditions.

What Are the Common Rent Growth Assumption Traps?

A recurring multifamily underwriting error is using national or metro-level rent growth averages instead of submarket-specific data and applying smooth annual rent growth to a market with material new supply hitting in the next 24 months.

Four traps that show up repeatedly:

Trap 1: MSA-level data masking submarket weakness. An MSA average can combine strong and weak submarkets. Pull effective-rent and concession evidence for the property's competitive set and a documented radius appropriate to the asset, then adjust for the specific property class, unit mix, age, condition, and location.

Trap 2: Smooth annual growth in a lumpy supply environment. Underwriting 3 percent annual rent growth for five years assumes new supply is absorbed evenly. Check the local permit, starts, units-under-construction, completion, absorption, and concession data before choosing a path. A national forecast cannot establish the delivery schedule for the property's submarket.

Trap 3: Inheriting the seller's loss-to-lease roll-off pace. A seller may model loss-to-lease rolling off on a calendar. In a softening market, the asking rent on a vacated unit may be lower than the contract rent it replaced. Underwrite mark-to-market on the property's lease expirations, renewal history, turn pace, downtime, concessions, and verified comparable units, not on a universal turnover percentage.

Trap 4: Forgetting concessions. Asking rent is not necessarily collected rent. A property advertising 1,950 dollars per month with one month free on a 12-month lease has average collected rent of 1,787.50 dollars per month before other concessions or bad debt: 1,950 x 11 / 12. Underwrite net effective rent, and check whether comparable properties are advertising concessions that the current rent roll does not reflect.

Source check: For current context, see Freddie Mac's Q1 2026 multifamily presentation, the Census Q1 2026 vacancy release, and Census and HUD new-construction data. These series have different populations and definitions. Use local permits, completions, absorption, leases, and concessions for the subject property.

How Should You Normalize Operating Expenses?

"Normalizing" the T12 means adjusting every operating expense line to reflect what the property will actually cost to run under your ownership, not what it cost the current owner under their tax basis, their insurance carrier, their management arrangement, and their deferred-maintenance posture.

The lines that need the most scrutiny:

Property taxes. Start here, but do not generalize across jurisdictions. Determine whether the transfer triggers reassessment, how assessed value is defined, which assessment ratio and tax rates apply, when the new value becomes effective, and whether abatements, PILOTs, exemptions, or caps transfer. California's Board of Equalization, for example, says a change in ownership generally triggers reassessment to current fair market value, subject to exclusions. The seller's bill is evidence of the current assessment, not necessarily the buyer's forecast.

Property management fee. Obtain the proposed management agreement and price the actual services, staffing, supervision, and incentive fees. If the property is self-managed, add the internal labor and overhead that new ownership will actually bear. Do not substitute a universal percentage for a management proposal.

Repairs and maintenance. Look for consistency in the T12. Reconcile R and M to work orders, invoices, inspection findings, system age, staffing, and accounting policy. Classify betterments, restorations, and adaptations consistently with the property's accounting policy and applicable tax guidance. Value-add properties need a scope-based budget, not a generic per-unit benchmark.

Insurance. Get a current quote or indication from your broker for the property under your name and ownership structure. Do not use the seller's premium as the forecast. Price construction, loss history, catastrophe exposure, deductibles, limits, lender requirements, and policy terms.

Utilities. Pull the actual 24-month utility invoice history, rate schedules, meter configuration, weather, consumption, and service contracts. Use the applicable utility or regulatory forecast where available, then model the property's usage separately from any legally collectible recovery.

Payroll. For properties with on-site staffing, payroll is a major line. Adjust for your management company's wage structure and benefits load, not the seller's.

Marketing and turnover. Use the property manager's budget, recent invoices, lead-to-lease conversion, traffic, vacancy days, concessions, and planned lease-up scope. Do not use a universal per-unit marketing allowance.

Inflate each expense line using its own evidence. Taxes, insurance, utilities, payroll, contracts, and supplies have different drivers. Show a base case and sensitivities, and document the source and effective date for every escalation.

How Do You Estimate the Property Tax Reassessment?

Property tax reassessment practice varies dramatically by jurisdiction. Estimating year-one taxes correctly requires understanding whether the state reassesses on sale, what assessment ratio applies, and what the local mill rate is.

Four jurisdiction archetypes to know:

Transfer-related reassessment. Some jurisdictions reassess after a qualifying transfer, but the trigger, valuation standard, timing, assessment ratio, exemptions, and appeal process are local questions. California is a documented example. Do not assume the purchase price, a one-to-two-year timing rule, or a particular assessment ratio applies elsewhere.

Periodic or mass-appraisal systems. Some jurisdictions use periodic valuation cycles or other rules. Confirm the next valuation date, interim rules, appeal rights, and effective date with the assessor or governing authority.

Caps and exemptions. Assessment caps and exemptions often depend on use, ownership, filing, and asset class. Do not import a homeowner rule into a commercial multifamily model. Verify the rule for the subject parcel.

Negotiated abatement and PILOT states. Some jurisdictions offer payment-in-lieu-of-taxes (PILOT) or tax abatement programs that fix taxes for a defined period. If the property is currently under a PILOT or abatement, verify whether the agreement transfers to new ownership, when it expires, and what taxes step up to after expiration. This is a major underwriting input that frequently gets missed.

The practical workflow: call the county assessor's office directly. Most assessors will provide an estimated post-sale assessment based on a stated purchase price. Use that estimate, not the seller's current bill, in the model. For larger deals, retain a property tax consultant in the jurisdiction to validate the estimate and identify appeal opportunities.

Keeping the seller's property tax basis in the year-one pro forma without checking the local rule creates a false sense of precision.

For tax classification and capital-versus-repair treatment, see the California Board of Equalization change-in-ownership guidance as an example of jurisdiction-specific rules and IRS Publication 527 for the federal repair, improvement, and depreciation distinction. Neither replaces local tax counsel or the assessor's current guidance.

How Do You Calculate NOI and Properly Handle CapEx?

Net Operating Income equals Effective Gross Income minus total operating expenses. Capital Expenditures are tracked separately below the NOI line because they are not recurring operating costs.

The clean equation:

• Gross Potential Rent
• Less loss to lease
• Less vacancy and credit loss
• Plus ancillary income
• Equals Effective Gross Income
• Less normalized operating expenses (taxes, insurance, management, payroll, utilities, R and M, marketing, admin)
• Equals Net Operating Income

NOI excludes CapEx, debt service, depreciation, and income taxes. NOI is the metric cap rates are applied to and the metric that determines whether the property covers debt service.

CapEx covers items with a useful life of multiple years: roof replacement, HVAC replacement, parking lot repaving, exterior paint, major appliance replacements, water heater bank replacement, plumbing risers, and unit interior renovations in a value-add scope. Size replacement reserves from the property condition assessment, engineer or contractor bids, useful-life schedule, lender requirements, and the timing of planned work. Freddie Mac's K-Deal materials describe a program-specific reserve floor generally equal to the greater of an engineer's recommendation or 250 dollars per unit. That is not a universal reserve benchmark.

CapEx belongs below the NOI line because investors and lenders need NOI as a clean comparable metric. Burying CapEx inside operating expenses (a frequent seller-pro-forma trick) makes the operating cap rate look higher than it actually is. Pull every capital item out of the seller's T12 before computing your own NOI.

How Do You Stress Test Multifamily Underwriting?

A defensible multifamily underwriting includes a base case, a downside case, and a break-even sensitivity that shows where the deal stops working. In a softening market, the downside case matters more than the base case.

The minimum stress-test stack for a 2026 multifamily acquisition:

• Rent and concession downside based on observed local effective-rent volatility, lease expirations, competing deliveries, and the property's exposure. Use a wider sensitivity grid where evidence supports it rather than declaring a universal 10 to 15 percent decline.
• Line-by-line expense downside using the property's insurance indication, tax estimate, utility history, payroll plan, and contract renewals. Apply a separate sensitivity to items with genuine quote or rate uncertainty.
• Vacancy, credit loss, and downtime stress based on the property's collections and leasing history, local supply, concessions, and lease-up assumptions. Keep physical vacancy, economic vacancy, credit loss, and concessions visible.
• Exit cap rate expansion tested as a sensitivity, with the range tied to current transaction evidence, lender maturity analysis, and the uncertainty of the NOI forecast. Holding NOI constant, a move from 5.5 percent to 6.0 percent reduces value by 8.33 percent because 1 - 5.5 / 6.0 = 0.0833.
• Interest rate shock if the deal includes floating-rate or short-term debt. Model the rate cap, the strike, and what happens if the cap expires before refinance.
• Renovation overrun for value-add deals. Use contractor bids, contingency policy, scope completeness, schedule, procurement exposure, and a documented contingency sensitivity. Do not assume a universal overrun percentage.

The downside case does not have to be the likely case. It has to be one the deal can survive. If the property still covers debt service there, it can hold through a recession. If it misses debt service, expect a capital call.

Lender sizing is product-specific. For example, Fannie Mae's conventional term sheet lists 80 percent maximum LTV, 1.25x minimum DSCR, and typically 90 percent stabilized occupancy for 90 days. Freddie Mac's current materials show different requirements by product and rate structure. Use the actual lender term sheet, including its NOI definition, replacement reserves, vacancy, taxes, insurance, interest-rate floor or cap, amortization, and maturity analysis. See Fannie Mae's conventional properties term sheet and Freddie Mac's current Guide resources.

What Are the Most Common Multifamily Underwriting Mistakes?

The most damaging multifamily underwriting errors involve trusting the seller's pro forma, missing the property tax reassessment, using overoptimistic rent growth, mis-classifying CapEx as operating expense, and skipping downside stress testing.

1. Anchoring on the seller's pro forma. The seller's broker built the pro forma to sell the asset. Treat it as a marketing document, not an underwriting input. Build the model from the T12 and rent roll yourself.
2. Keeping the seller's property tax basis. Covered above, and the most expensive error on this list. Always get a post-sale estimate from the assessor.
3. Using metro-level rent growth in a softening submarket. Pull submarket-level data and adjust for property class. A 2 percent metro average can mask a minus 5 percent submarket.
4. Burying CapEx inside operating expenses. Inflates apparent NOI and apparent cap rate. Pull capital items out before computing normalized NOI.
5. Ignoring insurance and tax evidence. Do not apply a generic escalation to a quoted insurance premium or local tax rule. Use current property-specific evidence and a sensitivity for uncertainty.
6. Skipping the downside case. A base case on its own does not tell you what the deal can absorb. Run the downside case and the sensitivity grid too.
7. Not reconciling the rent roll to bank deposits. If reported economic occupancy does not match T12 collections divided by GPR, the rent roll is wrong, the T12 is wrong, or both.

Why Does the Financial Model Have to Be Dynamic?

A pro forma that takes the seller's numbers and lightly edits them is still the seller's marketing document. The model has to support the full workflow from initial screening through closing diligence through hold-period management.

A multifamily acquisition model that earns its place in the workflow handles:

• Unit-level rent roll import with current and market rent comparison
• Monthly T12 import and expense normalization with override columns
• Loss-to-lease roll-off on actual renewal pace, not calendar
• Property tax reassessment based on purchase price, assessment ratio, and mill rate
• Ancillary income build with line-by-line $/unit/month inputs
• Operating expense projections with line-by-line inflation assumptions
• CapEx tracking below the line with reserve funding and timing
• Sensitivity tables on rent growth, exit cap, expense inflation, and renovation overrun
• LP and GP waterfall distributions including preferred returns and promote tiers
• Debt scenarios covering fixed-rate, floating-rate with rate cap, supplemental loans, and refinance

Holding NOI constant, moving from a 5.5 percent to a 6.0 percent exit cap reduces indicated value by 8.33 percent. The model cannot hide that sensitivity.

Frequently Asked Questions