Multifamily underwriting, rent roll to exit cap

What a model is, and what it is not

Underwriting is the process of forming a view about what a property will earn, what it will cost to own, what it can be financed with, and what someone will pay for it later. The output is a model, and the model is the document investors are shown.

A model is not a forecast. It is a set of assumptions arranged so that they produce a number, and the number inherits every weakness of the assumptions without displaying any of them. Two sponsors can model the same building and reach returns that differ by half, entirely through choices that are each individually defensible.

The useful skill is therefore not building a model. It is locating the four or five assumptions that carry the result and asking what each one is based on. In most multifamily deals those are the rent premium, the expense growth, the exit cap rate, and the debt.

The rest of this page is a line-by-line account of where those assumptions sit, and which free public dataset lets you check each one yourself without taking anyone's word for it.

The income statement, line by line

Every multifamily model has the same skeleton, and knowing the order lets you find where an assumption has been made rather than disclosed.

It begins with gross potential rent, which is every unit let at the assumed market rent for a full year, occupied or not. That figure is theoretical by construction.

From it are deducted loss to lease, the gap between in-place rents and market; vacancy; concessions, meaning free rent used to fill units; bad debt, being rent billed and never collected; and non-revenue units such as a model apartment or staff unit. What remains is net rental income.

Then other income is added: parking, pet rent, utility reimbursement, laundry, storage, fees. This line deserves scrutiny because it is easy to grow on paper and harder to grow in practice, and a model leaning on aggressive other-income growth is often disguising a weak rent assumption.

The total is effective gross income. Deduct operating expenses, which include payroll, repairs and maintenance, contract services, marketing, administrative, utilities, management fee, insurance and property taxes, and you reach net operating income.

Note what net operating income excludes: debt service, capital expenditure and depreciation. It is a property-level measure, not an investor-level one, which is why a strong NOI can coexist with no distribution at all.

The rent assumption, and how to check it

The rent premium is where most optimism lives. A renovation plan typically assumes that a renovated unit achieves a stated premium over an unrenovated one, and the entire business plan compounds off that number.

Three questions test it. How many units have actually been renovated and re-leased at the premium, as opposed to modeled at it? Over what period, and in what leasing season? And what were the comparable properties, specifically, including their age, amenities and distance?

You can sanity-check the rent level independently. HUD publishes Fair Market Rents by metropolitan area and bedroom count, which is not a market rent but is a useful floor and an indicator of where subsidized demand sits. HUD also publishes Comprehensive Housing Market Analyses for individual metros, which are detailed independent reports on demand, supply and rent conditions, and are among the most underused free documents in this field.

For the demand side, BLS Economy at a Glance gives metro employment and unemployment, and data.census.gov gives household income, renter share and household formation. A rent growth assumption materially above local income growth is a bet that renters will spend a rising share of income, which can be true for a while and is rarely true for five years.

Supply, the assumption most often left out

A rent forecast is a statement about supply as much as demand, and models frequently address demand at length and supply in a sentence.

The Census Bureau publishes the Building Permits Survey, which gives permitted units by structure size for metropolitan areas and individual jurisdictions, monthly. Multifamily permits lead deliveries by roughly one to two years depending on the market and the product, so today's permit count is a reasonable leading indicator of the competitive supply arriving during your hold.

The Housing Vacancies and Homeownership series gives rental vacancy rates nationally and by region, which is a check on whether a submarket vacancy assumption is plausible against the broader trend.

The question to ask a sponsor is specific: how many multifamily units have been permitted within the submarket over the last twenty-four months, and what share of them competes directly with this asset on price and product? A sponsor who has looked will have a number. A sponsor who has not will describe the submarket as supply-constrained.

This is the assumption we most often decline a market over, and the market guides state where and why, including the metros we would not buy in at all.

The two expense lines that break models

Operating expenses are usually grown at a modest assumed rate, and two lines routinely refuse to cooperate.

Property taxes. In many jurisdictions a sale triggers reassessment, so the tax bill the seller paid is not the tax bill the buyer will pay. A model carrying forward the seller's taxes understates expenses permanently and the error compounds into the exit value, because it inflates NOI at sale as well as during the hold. Ask directly whether the model assumes reassessment, at what basis, and what the assessor's practice actually is in that county. This is checkable: county assessors publish their rolls and their methods.

Insurance. Premiums have moved sharply in several markets, particularly coastal and severe-convective-storm exposed ones, and a renewal can arrive at a multiple of the expiring premium. Ask what the current premium is, when it renews, what the last two renewals did, and whether the quote in the model is a real quote or an assumption.

A third worth checking is payroll, which is often modeled on the seller's staffing rather than the plan's. A renovation program and an intensive leasing push usually need more people, not fewer.

A useful discipline is to ask for the model's year one expenses against the trailing twelve months actuals, line by line, with an explanation for every line that goes down. Expenses that fall on a value-add plan deserve a specific reason.

The capital stack

The financing assumptions have caused more loss in recent years than the operating assumptions, and unlike the operating assumptions they are facts rather than forecasts at the point you invest.

Establish: fixed or floating; if floating, the index and spread; the term and the exact maturity date; whether there is an interest-only period and for how long; the amortization schedule after it; the rate cap strike, its expiry, and the current cost of replacing it; the debt service coverage and loan-to-value covenants; and whether a cash management or lockbox trigger exists and at what level it springs.

Then ask the question that connects them: what combination of circumstances breaches a covenant or requires new equity, and how far away is that from the base case? A sponsor who has run that analysis answers in numbers.

Also establish who signs the loan and what else they guarantee, because lenders assess guarantors on a global basis. That matters more than investors expect and is set out in lender workouts and forbearance.

For market context on multifamily financing conditions, Freddie Mac multifamily research publishes regular outlooks, and the FHFA house price index gives an independent series on price movement.

The exit cap rate does more work than anything else

The exit value in almost every model is stabilized net operating income divided by an assumed exit capitalization rate. That single assumed number frequently drives more of the projected profit than the entire operating plan.

The discipline is simple and rarely applied. A model assuming an exit cap rate lower than the entry cap rate is assuming the buyer pays more per dollar of income than the sponsor did, which is a forecast about the market rather than about the property. It should be stated, justified, and treated with suspicion.

Conservative practice is to assume the exit cap rate is modestly higher than entry, expanding by a stated amount per year of hold, on the reasoning that the building is older at exit and the rate environment is unknown. Ask what the model assumes and ask to see the return with the exit cap rate expanded by a further fifty and a hundred basis points.

That second calculation is the most valuable twenty seconds in the whole review. If a fifty basis point expansion removes most of the projected profit, the deal is a bet on cap rates rather than a business plan, whatever the narrative says.

Stressing a model in twenty minutes

You do not need the spreadsheet. Ask for the return under five specific changes, each in isolation, and each of them exists already in the sponsor's own model.

Exit cap rate expanded by fifty and by one hundred basis points. Rent premiums achieved at half the assumed level. Renovation costs twenty percent over budget and the schedule six months late. Property taxes reassessed at the purchase price. Insurance up fifty percent at renewal.

Then one combined case, because risks arrive together rather than singly: half the premium, six months late, and fifty basis points of cap rate expansion. That combination is not pessimistic, it is ordinary.

Two things are being tested. The first is the deal's resilience. The second, and more useful, is whether the sponsor has already run these. A sponsor who produces them immediately has stress-tested their own work. One who treats the request as adversarial has told you something about how the next five years of questions will go.

The wider diligence sequence is in how to vet a sponsor and the condensed version is nine questions before wiring.

What we will show you

We will give you the model's assumptions rather than only its outputs: the rent premium and how many units have actually achieved it, the expense assumptions against trailing actuals line by line, whether taxes are modeled on reassessment, the current insurance premium and renewal date, the full loan terms, and the entry and exit cap rates with the sensitivity around them.

We will also tell you which assumption we are least confident about, because every model has one and a sponsor who claims otherwise has either not looked or is not saying.

What we will not do is present a projection as a forecast. Returns are not promised, distributions can be suspended, more capital can be required, and you can lose the entire amount. What a projection is worth depends entirely on the assumptions underneath it, which is why this page describes how to check them rather than asking you to trust ours. Your share of the result arrives on a Schedule K-1, and how to read it is in reading your Schedule K-1.

What to test in the model

1. How many units have been renovated and re-leased at the assumed premium, over what period?
2. How many multifamily units have been permitted in this submarket in the last 24 months, and how many compete directly?
3. Does the model assume property tax reassessment on sale, and at what basis?
4. What is the current insurance premium, when does it renew, and what did the last two renewals do?
5. May I see year one modeled expenses against trailing twelve months actuals, line by line?
6. What are the entry and exit cap rate assumptions, and what justifies the difference?
7. May I see the return with the exit cap rate expanded by 50 and 100 basis points?
8. What combination of events breaches a loan covenant or requires new equity?

What this is built on

1. U.S. Department of Housing and Urban Development, HUD User, Fair Market Rents
2. U.S. Department of Housing and Urban Development, HUD User, Comprehensive Housing Market Analyses
3. U.S. Census Bureau, Building Permits Survey
4. U.S. Census Bureau, Housing Vacancies and Homeownership
5. U.S. Census Bureau, data.census.gov
6. U.S. Bureau of Labor Statistics, Economy at a Glance
7. Federal Housing Finance Agency, House Price Index
8. Freddie Mac, Multifamily research and outlooks
9. Internal Revenue Service, Schedule K-1 (Form 1065), Partner's Share of Income, Deductions, Credits