The Definitive Guide to Underwriting Multifamily Acquisitions…

By Ken Lundin, Author, Operator and Investor

I’ve watched operators lose $4M in equity because they trusted a pro forma that looked perfect on paper. The model collapsed in the first six months of ownership. This is the definitive guide to underwriting multifamily acquisitions. I wrote it because most underwriting models fail before the first rent check clears. The math isn’t wrong. The assumptions were never stress-tested against the actual property.

The gap isn’t in the spreadsheet formulas. It’s in the operator who treats underwriting as a fundraising pitch deck. They should treat it as a decision tool that predicts whether cash flow will actually show up.

I’ve seen deals where the projected IRR was 18%. The actual return was 6%. The difference wasn’t market timing or bad luck. It was an operator who underwrote to a number he needed. He didn’t underwrite to the number the asset could deliver.

The best operators I know hit their projections within 8% variance. The worst miss by 40% or more. The gap shows up in month three. That’s when the renovation budget runs over. That’s when the lease-up stalls.

The spread between confident models and actual performance comes down to one thing. Did you do the work that predicts cash flow? Or did you just fill in the template?

Key Takeaway: Multifamily underwriting fails when operators treat the pro forma as a sales tool rather than a decision framework. The gap between projected and realized returns isn’t caused by bad spreadsheets. It’s caused by untested assumptions about renovation costs, lease-up velocity, and operating expenses. Top operators hit projections within 8% variance by stress-testing every input against property-specific data. Bottom performers miss by 40% or more. They underwrote to the IRR they needed to raise capital. They didn’t underwrite to the return the asset could deliver.

TL;DR

Underwriting starts at the unit level—rent comps, turn costs, and actual vacancy loss—not top-down assumptions borrowed from the broker package
Expense assumptions must be verified on-site: walk the property, pull utility bills, review payroll, and test the T12 against what’s actually happening in the building
CapEx reserves fail when they’re modeled on wishful timelines instead of the real condition of roofs, HVAC systems, and deferred maintenance you can see during due diligence
A deal is ready when every material assumption has a named source, a documented verification step, and a downside scenario that still clears your return threshold

What Actually Drives Cash Flow in Multifamily Underwriting

I’ve watched operators lose six figures in the first 18 months. They underwrote to a rent roll the seller’s broker sent over. They never pulled the actual lease files.

The unit showing $1,450 in monthly rent? That tenant signed at $1,295 two years ago. They locked in a renewal at 2% annually. You just underwrote $1,860 in phantom revenue per year on one unit. Most acquisitions have 50 to 200 of them.

The work that predicts cash flow happens at the unit level. I pull every lease. I compare what tenants actually pay today against what the model says they’ll pay in month one. If there’s a gap, I need to know whether it closes through expiration and re-lease. Or does it require a buyout I haven’t budgeted?

Most underwriting models aggregate to an average rent per unit. They call it conservative. That’s not conservative. It’s lazy.

Market rent surveys tell you what comparable buildings are asking. They don’t tell you what your building can get. Your building has different unit mixes. It has a different reputation. It has a different deferred maintenance profile.

I’ve seen operators underwrite to Yardi Matrix comps showing $1,650 for a two-bedroom. Their actual building hasn’t cleared $1,500 in 18 months. The rent assumption isn’t a market question. It’s a property-specific question.

I model three scenarios for every unit. First: in-place rent. Second: renewal probability based on lease expiration schedule. Third: the actual market rent I can defend with signed leases from the last six months at that property.

If I don’t have six months of turnover data, I assume my rent growth is zero. I hold that assumption until I prove otherwise with real leases.

Revenue isn’t what you hope to collect. It’s what the building has demonstrated it can collect. The building must collect from actual tenants who signed actual leases.

The operators who hit their returns don’t start with market assumptions. They start with the lease file. They adjust the model to match reality. They don’t adjust reality to match the model.

Once you know what drives revenue, the next gap shows up. That gap is in how you model what it costs to operate the asset.

Why Most Expense Models Are Wrong Before You Close

Most operators pull expense ratios from the broker’s package. Or they annualize the trailing twelve months. They call it diligence.

That works if the seller ran the property well. It works if nothing’s about to break. It fails the moment you inherit deferred maintenance. It fails when you inherit a property manager who hasn’t bid a contract in three years. It fails when you inherit a tax appeal that expires sixty days after close.

I’ve seen deals underwritten at 52% expense ratios close into 64% actuals. Nobody walked the roof. Nobody called the waste hauler. Nobody asked why repairs and maintenance was half the submarket average.

The TTM shows what the last guy spent. It doesn’t show what you’ll spend once you actually manage the asset.

Start with the rent roll and the vendor contracts. Walk every unit scheduled to turn in the next six months. The broker’s showing 8% turnover. You’re inheriting twenty units with original countertops. You’re inheriting HVAC from 2005. Your turn costs aren’t $1,200. They’re $4,500. They’re hitting in month two.

Pull the actual invoices for insurance, waste, landscaping, and snow removal. Don’t model what the market “should” cost. Call the incumbent vendors. Get a renewal quote. If they haven’t been rebid in five years, you’re underwriting a 20–30% step-up the day you take over.

Tax assessments get missed more than anything else. The current owner filed an appeal. Or they got an abatement. Find out when it expires. I’ve seen operators model the current tax bill into year three. They didn’t know the assessed value resets at sale. That’s not a projection error. It’s a $40,000 annual hole you didn’t budget.

Payroll’s the other one. You’re taking over a mismanaged property. You’re planning to add a porter. You’re planning to add a part-time leasing agent. You’re planning to add any headcount the seller didn’t carry. That’s an operating expense. It’s not a one-time cost. It goes in the pro forma. It doesn’t go in the capital budget.

Even accurate expense modeling doesn’t save you if your capital plan is fiction.

The Capital Reserve Trap That Kills Multifamily Returns

I’ve watched operators underwrite $15M acquisitions with a single CapEx line item. The line item reads: “3% of gross revenue.” That number didn’t come from a site visit. It came from a template someone downloaded in 2019.

The question isn’t whether you need reserves. The question is whether your reserves reflect the actual condition of the systems you just bought.

Walk the property with someone who replaces HVAC units for a living. Ask which compressors are loud. Ask which condenser coils show corrosion. Ask how many units are running on refrigerants that haven’t been legal to manufacture since 2020.

A 12-year-old heat pump in Florida isn’t the same asset as a 12-year-old heat pump in Colorado. The first one has maybe 18 months left. The second might run another four years.

I’ve seen operators reserve $250,000 for roofing. The inspection said “serviceable condition, 5–7 years remaining useful life.” Then they get into year two. They discover the ponding water they were told was cosmetic has rotted the decking in three sections. Now it’s a $340,000 tear-off. It requires a six-week tenant notice period. It includes lost rent they never modeled.

Your CapEx schedule should name systems, dates, and unit costs. Not percentages.

You’re underwriting a 1980s garden-style walkup. You need to know when the water heaters were last replaced. You need to know what it costs to replace them today in your market. You need to know whether the electrical panels can handle an upgrade to heat pump water heaters. You need to know if you want the tax credit.

The operators I know who hit their year-three cash flow targets didn’t guess at reserves. They got bids. They walked the site with the contractor who’d actually do the work. They built a replacement schedule based on observed condition and verified pricing. They didn’t use broker comps or national averages.

If your CapEx model fits on one line of the pro forma, you’re not underwriting. You’re inheriting someone else’s assumption about what things cost. You’re hoping it’s close enough.

Ready to Take the Next Step?

Book a Strategy Call

Multifamily Underwriting Comparison: Broker Package vs. Verified Model

Here’s what separates operators who hit their projections from those who miss by 40%:

Input	Broker Package Approach	Verified Underwriting Approach	Impact on Year 1 Cash Flow
Rent Assumptions	Market rent per CoStar ($1,650/unit)	Actual signed leases last 90 days ($1,485/unit)	-$99,000 on 50-unit property
Expense Ratio	TTM at 52%	Site-verified at 58% (tax reset + deferred contracts)	-$72,000 on $1.2M revenue
CapEx Reserves	3% of revenue ($36,000)	System-by-system schedule ($127,000)	-$91,000 first 18 months
Turn Costs	$1,200/unit (broker estimate)	$4,200/unit (contractor bid for actual condition)	-$60,000 on 20 turns

The gap between these two approaches on a 50-unit acquisition: $322,000 in year one. That’s the difference between a 14% IRR and an 8% IRR. It’s also the difference between raising a second fund and explaining to your LPs why the deal didn’t perform.

How to Know When Your Underwriting Is Actually Finished

I’ve watched operators walk away from deals they should have taken. I’ve watched them close on deals that were dead before the ink dried. The difference wasn’t the spreadsheet. The difference was whether they could name where every number came from. The difference was whether they knew what happens when the number is wrong.

A deal is ready when every material assumption has three things. First: a named source. Second: a documented verification step. Third: a downside scenario that still clears your return threshold.

Not broker numbers. Not trailing averages. Not “market rents per CoStar.”

Here’s what that looks like in practice. Your rent assumption isn’t “$1,450 for a two-bedroom.” It’s “$1,450 based on the twelve current leases for two-bedrooms signed in the last 90 days.” It’s “verified against the rent roll we pulled Tuesday.” It’s “with a 5% vacancy stress test that still delivers a 1.25 DSCR.”

Your property tax assumption isn’t last year’s bill. It’s last year’s bill plus the reassessment trigger from the sale. It’s confirmed with the county assessor. It’s modeled at 15% higher than their estimate. They’re always low.

Your CapEx reserve isn’t 5% of revenue. It’s $847,000 in year one. The HVAC systems are original to 2006. The parking lot needs a mill-and-overlay by month eighteen. You’ve got bids from two contractors you’ve used before. You know what fails and when. You walked every unit and documented it.

The verification step is what separates underwriting from storytelling. I’ve seen operators model a value-add renovation at $8,500 per unit. That’s what the last deal cost. But it was in a different submarket. It had a different scope. It was eighteen months ago.

The operator who wins that deal has three bids. They have a line-item scope. They have a construction timeline that accounts for permitting delays in that specific jurisdiction.

If your downside case breaks your return threshold, you don’t have a deal. Your downside case includes 10% lower rents. It includes 8% higher expenses. It includes a six-month lease-up delay. If that scenario breaks your threshold, you have a bet. You don’t have a deal.

Most operators can’t tell the difference. They can’t tell until the cash stops flowing.

FAQ

What is the definitive guide to underwriting multifamily acquisitions?

It’s the process of building a cash flow model from unit-level economics. It uses site-verified expenses. It uses capital reserves tied to actual system conditions. It doesn’t use inherited assumptions from the broker package.

I’ve seen operators treat underwriting like a checklist. They plug in market rents. They apply a 50% expense ratio. They reserve 5% for CapEx. They call it done.

The definitive approach works backward from every material assumption. It works backward to a named source. It works backward to a verification step you can defend in year three. That’s when the deal isn’t performing.

What are the most common mistakes in multifamily underwriting?

Using trailing twelve-month expenses without adjusting for deferred maintenance. Underwriting to market rents instead of actual in-place leases. Reserving for CapEx as a percentage of revenue instead of a line-item budget tied to system age.

The fourth mistake kills deals. It’s skipping the downside scenario because the base case already pencils. If your model only works when everything goes right, you’re not underwriting. You’re hoping.

How do you verify rent assumptions in a multifamily deal?

Pull the current rent roll. Compare every unit’s in-place rent to the lease-up rate for similar units in the last 90 days. Then walk five comparable units in buildings within a half-mile. Those buildings should match your post-renovation profile.

I don’t use market rent surveys. They reflect someone else’s building. They reflect someone else’s management. They reflect someone else’s concession strategy.

The gap between what a broker calls “market rent” and what you can actually collect in month four is where most deals die.

What expense ratio should I use for multifamily underwriting?

There is no ratio. I build expenses line by line. I start with the current property tax bill. I get verified insurance quotes for your entity and coverage limits. I review contracts you’ll actually inherit or renegotiate. I calculate payroll for the staffing model you’re going to run.

The 50% rule is a shortcut. It’s for people who don’t want to do the work. I’ve seen expense ratios range from 38% on a well-maintained Class A to 64% on a value-add deal. The value-add deal had deferred systems. It had an upcoming tax reassessment. Using one number for both is malpractice.

How much should I budget for CapEx reserves in a multifamily acquisition?

Budget what it costs to replace the systems that are failing now. Budget for systems that fail predictably in your hold period. Those systems include roofs, HVAC, plumbing, and parking lots. Price them at today’s contractor rates in your market. Don’t use a percentage of revenue.

I just underwrote a 150-unit deal in Phoenix. The HVAC systems averaged 18 years old. Replacement cost was $4,200 per unit. The replacement was due in year two. That’s $630,000. That’s 11% of year-one revenue. That’s three times the “3-5% reserve” the broker recommended.

The 3-5% rule assumes systems that were maintained. Most value-add deals inherit systems that weren’t.

What return threshold should trigger a pass on a multifamily deal?

Run your downside scenario. In that scenario, rent growth stalls. Expenses run 8% higher than modeled. CapEx hits in year two instead of year four. If that scenario doesn’t still clear your minimum IRR and equity multiple, you pass.

I use a 12% IRR floor. I use a 1.6x equity multiple floor in the downside case. The base case is marketing. The downside case is underwriting.

Most operators never run the downside. They’ve already decided they want the deal. They’re using the model to justify it. They’re not using the model to test it.

How do you underwrite deferred maintenance in a multifamily property?

Walk every building system with a contractor. That contractor will give you a replacement cost. They’ll give you a failure timeline. Then add 15% to the cost. Pull the timeline forward six months.

Deferred maintenance isn’t a line item. It’s a capital event with a price and a date. I’ve seen operators budget $8,000 per unit for interior renovations. They skip the $1.2 million roof replacement that fails in month eleven. It wasn’t on the broker’s CapEx schedule.

The inspection report will tell you “serviceable condition.” The roofer standing on the membrane will tell you it’s got two winters left. Maybe three if you’re lucky.

What’s the difference between underwriting for acquisition vs. underwriting for investors?

Underwriting for acquisition tells you whether to buy the deal. Underwriting for investors tells them whether to fund it.

The first one requires brutal honesty about downside scenarios. It requires honesty about hidden costs. It requires honesty about every assumption that could break.

The second one is a sales document. It shows your best case. Maybe it shows a sensitivity table.

I’ve seen operators confuse the two. They build one model that does both jobs. That means they either lie to themselves about the deal. Or they scare off their capital partners with worst-case thinking.

Build two models. One tells you the truth. One raises the money.

How do you account for rent growth in a multifamily underwriting model?

I don’t. Not in year one. Maybe 2-3% in year two. But only if I’ve got twelve months of actual performance. That performance must show I can push rents without killing occupancy.

The operators who get in trouble model 5% annual rent growth. They model it because “that’s the market average.” Or they model it because “that’s what we need to hit our return.”

Rent growth is something you earn by improving the asset. You earn it by proving demand. It’s not something you assume because your model needs it.

If your deal only works with aggressive rent growth, you’re buying the wrong deal. Or you’re paying the wrong price.

What due diligence items kill the most multifamily deals?

Environmental reports showing underground storage tanks nobody knew about. Title issues where the seller doesn’t actually own the parking lot. Structural engineers finding foundation problems the inspector missed.

But the one that kills deals most often isn’t a report. It’s walking the units. You realize the renovation budget you modeled at $12,000 per unit is actually $22,000. The plumbing is galvanized steel from 1974. Half the electrical panels are Federal Pacific. They won’t pass code.

You find that in week two of due diligence. Now you’re renegotiating price. Or you’re walking.

How do you validate contractor bids during underwriting?

Get three bids from contractors who’ve worked in that submarket in the last twelve months. Compare line items, not just total cost. Ask for references on similar-scope projects. Then add 15% to the lowest bid for your model.

According to industry research, construction cost overruns average 12-18% on multifamily renovations. The overruns are higher when you’re working with deferred maintenance. I’ve seen operators accept the first bid they get. They don’t verify scope. They don’t check references. Then they’re shocked when the contractor walks off the job in month four.

The bid you use in underwriting should come from someone you’d actually hire. It should reflect the real scope of work you documented during site visits. It should account for permitting timelines in that jurisdiction.

What role does market timing play in multifamily underwriting?

Market timing matters less than property-specific fundamentals. I’ve seen operators buy at the “wrong” time and still hit returns. They bought right. They underwrote conservatively. They executed the business plan.

I’ve also seen operators buy at the “right” time and lose money. They overpaid. They underwrote to aggressive assumptions. They couldn’t execute.

Enterprise deals now involve an average of 6-10 decision-makers spread across multiple departments, with each stakeholder bringing distinct success criteria and veto power to the buying process. In multifamily, your “decision-makers” are the market fundamentals. Those fundamentals include job growth, household formation, and supply pipeline. They also include your specific asset’s condition, location, and competitive position.

You can’t control interest rates. You can’t control cap rate compression. You can control whether you underwrote to the actual cash flow the property can deliver. You can control whether you stress-tested your assumptions. You can control whether you have capital reserves for what’s actually going to break.

Bottom Line

Most operators treat underwriting like a checkbox exercise. The ones who hit their 15%+ IRR targets do three things differently. First: they build rent assumptions from actual lease files, not broker comps. Second: they verify every expense line with invoices or site-confirmed vendor quotes. Third: they reserve capital based on equipment age and replacement cost, not a percentage someone pulled from a different market.

If your model doesn’t have unit-level data and named sources for every material assumption, you’re not underwriting. You’re guessing. Go back and verify the three numbers that matter most to your cash flow.