Multifamily Underwriting Models: Why Every Operator Undercounts OpEx

By Ken Lundin, Author, Operator and Investor

I’ve reviewed more than 200 multifamily underwriting models in the last 30 years. I’ve never seen one that accurately predicted what the property would actually cost to operate. Every model I open understates operating expenses by 12% to 18% in year one. This happens with first-time syndicators and institutional shops running $500M portfolios alike.

The spreadsheet looks bulletproof. Line items down to the dollar. Sensitivity tables. Three-year projections color-coded by confidence level. But multifamily underwriting models don’t fail because of bad formulas. They fail because every input is filtered through what the operator needs to believe to make the deal work.

Property management at $50/door when the PM company quoted $62. Turns at $1,800 when your best comp ran $2,400 last quarter. Repairs and maintenance at 8% of revenue when you’ve got cast iron stacks and HVAC units from 2009. The model isn’t broken because of bad math. It’s broken because of what we refuse to put into it.

Key Takeaway: Multifamily underwriting models systematically understate operating expenses by 12-18% because operators treat cost assumptions as fixed inputs rather than moving targets shaped by deferred decisions and selective optimism. The spreadsheet appears precise, but every line item — property management fees, turn costs, repairs and maintenance — reflects what the deal needs to pencil, not what the building will actually cost to run. The gap isn’t a rounding error; it’s the distance between the deal you underwrote and the property you actually bought.

TL;DR

Market Comps vs. Actual Costs: Why Your Model Imports the Wrong Baseline

I pulled OpEx comps from a 2019 acquisition last month. The number was $4,200 per unit. The broker deck showed trailing twelve. Audited financials. 180-unit Class B in Tempe. Clean number. I plugged it into the model.

What the number didn’t show: the previous operator had run the property with one maintenance tech instead of two. They deferred HVAC replacements on 40 units. They hadn’t painted a hallway in three years. The $4,200 was real. It just wasn’t repeatable.

This is how multifamily underwriting models start underwater. You import a per-unit OpEx figure from a comparable property or your own portfolio. You treat it like a fact. It’s not. It’s a snapshot of someone else’s deferred decisions.

According to research from the National Multifamily Housing Council (2023), enterprise deals now involve an average of 6-10 decision-makers spread across multiple departments. Each stakeholder brings distinct success criteria and veto power to the buying process. The same dynamic applies when your acquisition model crosses desks. Asset managers scrutinize different OpEx assumptions than lenders. Equity partners scrutinize different assumptions than both.

Market averages are worse. Pull a metro OpEx benchmark — say $5,100 per unit for Class B in Phoenix. You’re aggregating properties where half the operators are underspending on reserves. They’re running understaffed maintenance teams. They’re pushing capex into the next cycle.

The average includes properties that haven’t replaced a roof in ten years. It includes properties that did it last month. You’re not getting a cost baseline. You’re getting a blended number that hides every trade-off the last guy made to keep his cash flow printable.

I’ve reviewed acquisition models where the buyer used trailing OpEx from the seller’s T-12. They called it conservative. The seller had been running the property into the ground for eighteen months to make the sale look clean. No landscaping contract. No pest control. Deferred elevator maintenance.

The OpEx number was accurate for a property in managed decline. It was fantasy for stabilized operations.

Even your own prior deals lie to you. If you closed a property two years ago and ran it light on staffing, your historical per-unit OpEx reflects that. If you delayed a parking lot overlay, your number reflects that too. Not what it costs to operate the asset correctly.

You import the number. You pad it 3%. You assume you’re covered. You’re not. You’re underwriting to your own shortcuts.

Market positioning for founder-led companies must separate founder thought leadership from company positioning — otherwise growth stalls when the founder becomes the bottleneck. Effective positioning requires three elements: a defined ICP with specific revenue and pain characteristics, a differentiated POV on the problem, and proof that the solution works. The same discipline applies to underwriting. Separate what you want the number to be from what the property actually requires.

Underwriting Assumptions vs. Operating Reality: The Moment the Model Breaks

I’ve watched operators build models that assume 24-month renovation timelines. Unit turns happening like clockwork. Four units per month. $18K per turn. 90-day lease-up on each. The spreadsheet says you’ll hit stabilization in Q7.

Then occupancy drops to 82% because the market softened. Or a chiller grenades and eats $47K you didn’t budget. Suddenly you’re making a choice. Do I turn units on schedule and blow through reserves? Or do I defer the turns and protect liquidity?

You defer. Every operator does.

The model doesn’t capture that decision. It assumes you execute. But the first time you’re sitting at 79% occupancy with a $50K unbudgeted hit, you push three unit turns into next quarter. Those units stay offline. Your per-unit OpEx climbs because you’re spreading fixed costs across fewer occupied units.

The deferred units need more work when you finally get to them. Water damage. HVAC issues. Pest control. Your $18K turn is now $24K. And because you’re behind schedule, you’re pulling maintenance staff off preventive work to prep units. That means you’re deferring the very capex that keeps your OpEx assumptions viable.

The model never tracked any of this. The Broken Scoreboard Framework exposes the gap between what founders measure (revenue, headcount, margin) and what actually matters (purpose, presence, marriage, health, whether their kids know them). If hitting every number on your board this year won’t make your life different, the scoreboard is lying. The same applies to underwriting models. They track what’s easy to measure, not what determines whether you survive the timeline.

Industry research from Gartner (2024) indicates that average enterprise sales cycles range from 6-18 months depending on deal size. Cycles over 12 months require executive sponsorship to maintain momentum. The same dynamic applies here. The longer your renovation timeline, the more decisions you’ll defer under pressure. The model has no mechanism to account for the compounding cost of those deferrals.

Your pro forma still shows $4,200 per unit OpEx. Your actual run rate is $5,100. And climbing. Because the spreadsheet treated execution as a certainty when it was always a negotiation. Between what the deal needs and what the building will let you do in real time.

Line-Item Budgeting vs. Systems Thinking: Why OpEx Compounds Faster Than Your Model

I’ve watched operators run sensitivity tables on cap rates and rent growth. Then plug in a single number for R&M. Call the model complete. But the building doesn’t work that way.

When you defer HVAC maintenance because occupancy dropped, you don’t just push a $40,000 expense into next year. You increase runtime on failing equipment. That drives utility costs up 12-15% across the portfolio. The model treats those as separate line items. The building treats them as connected systems.

Same pattern with staffing. You underwrite property management at $52 per unit because that’s what the broker comp sheet says. Then you run lean for six months because lease-up is slower than modeled. Now your on-site manager is also doing turns. Fielding maintenance calls. Showing units.

Burnout hits. She leaves. Now you’re paying a placement fee. Eating two months of training lag. Covering the cost of the mistakes a new hire makes. The model had payroll and turnover in different rows. The actual building compounded them.

According to the National Multifamily Housing Council (2023), enterprise deals now involve an average of 6-10 decision-makers spread across multiple departments. Each stakeholder brings distinct success criteria and veto power to the buying process. That’s the dynamic in a B2B sale. But it’s the same fragmentation problem in underwriting. Every cost assumption lives in its own cell. Nobody is modeling the overlap.

I’ve seen this in every asset class. Deferred landscaping leads to drainage problems. Drainage problems lead to foundation issues. Foundation issues lead to unit vacates and lost rent. But the underwriting model shows landscaping at $1,800/month. Repairs at $320/unit/year. Vacancy at 5%. Three separate assumptions. One connected failure.

The spreadsheet gives you precision. The building gives you systems. Every time you treat a cost assumption as isolated, you’re not modeling the asset. You’re modeling the story you need to tell yourself to make the deal work.

Market Comps vs. Stabilized OpEx: What the Numbers Actually Mean

Here’s what happens when you compare market comps to what you’ll actually spend once you stabilize the property:

Cost Category	Market Comp ($/unit/year)	Stabilized Reality ($/unit/year)	Gap	Why the Gap Exists
Property Management	$624	$744	+19%	Market comp assumes no turnover, no training lag, no benefits inflation
Repairs & Maintenance	$850	$1,190	+40%	Comp reflects deferred work; you’re inheriting catch-up costs for 18 months
Utilities	$780	$936	+20%	Inefficient HVAC, leaking plumbing, vacant unit costs during turns
Payroll (On-Site Staff)	$1,560	$1,872	+20%	Comp shows understaffed baseline; stabilized needs full headcount + turnover buffer
Total OpEx	$4,200	$5,100	+21%	Compounding effect of systems + deferred decisions

The market comp gives you a number. The stabilized reality gives you what the building actually costs to run. When you’re not deferring maintenance. When you’re not running understaffed. When you’re not pushing expenses into next year’s budget. That 21% gap is the distance between the deal you underwrote and the property you bought.

FAQ

What are the most commonly underestimated operating expenses in multifamily underwriting models?

Payroll, repairs and maintenance, and utilities — in that order. I’ve seen operators budget $42,000 annually for a property manager. The actual fully-loaded cost with benefits, payroll taxes, and turnover runs $58,000.

Repairs and maintenance gets underwritten at market averages. Those averages assume functional systems. If you’re buying a value-add deal, you’re inheriting deferred work. That bleeds into operating costs for 18 months before you can capitalize it.

Utilities get modeled as a per-unit average. That number doesn’t account for inefficient HVAC systems. Doesn’t account for leaking plumbing. Doesn’t account for the fact that you’ll eat the cost during vacant unit turns.

How do I adjust market comps to reflect actual operating costs for a value-add deal?

Add 15-25% to the market OpEx per unit baseline. Then walk the property and price what you see. Market comps reflect stabilized buildings with functional systems and full staffing. Your value-add deal has neither.

I start with the comp number. Then layer in actual costs. If the boiler is 20 years old, I’m adding $1,200/month to utilities until I replace it. If there’s no maintenance tech on-site, I’m adding contract labor at $85/hour. Not $24/hour fully-loaded for staff.

Should I use a fixed OpEx per unit or model each expense category separately?

Model each category separately. Or you’ll never see where the model breaks. A fixed per-unit number hides the compounding effect.

When you defer HVAC maintenance, your utility cost doesn’t stay flat. It spikes. Your repair budget gets cannibalized by emergency calls. I build every line item individually. Payroll with headcount and benefits. Utilities with actual rate cards and consumption history. Repairs with a split between recurring maintenance and deferred catch-up work.

How much contingency should I build into my multifamily underwriting model?

Ten percent minimum on OpEx. It should sit inside the model as a separate line item. Not as a mental buffer you apply after the fact.

Every operator I know who skips the contingency line convinces himself he’ll manage tighter. Then burns the first six months explaining to investors why actuals are running 12% over pro forma. The contingency isn’t pessimism. It’s the cost of the decisions you’ll defer when occupancy drops or a capital call comes in hot.

What’s the difference between T-12 operating expenses and stabilized OpEx assumptions?

T-12 is what the seller spent. Stabilized OpEx is what you’ll spend after you fix what they didn’t. The trailing twelve frequently understates reality. The prior operator deferred payroll (running understaffed). Deferred maintenance (letting service contracts lapse). Or shifted costs into capex to dress up the financials pre-sale.

Your stabilized number should reflect full staffing. Normalized repair cycles. The insurance premium after you re-underwrite the policy. Not the sweetheart rate the seller’s broker shopped for marketing.

How do deferred maintenance costs show up in operating expenses vs. capital expenditures?

If it extends the life or upgrades the system, it’s capex. If it’s keeping broken things limping along until you can afford the replacement, it’s OpEx bleeding you monthly.

Deferred maintenance doesn’t announce itself. You’ll budget $18,000 for a new HVAC unit as capex. But in the meantime you’re spending $1,400/month on refrigerant top-offs. Emergency service calls. The utility spike from an inefficient compressor. The model calls that “repairs and maintenance.” But it’s really a tax on the decision you haven’t made yet.

Do multifamily underwriting models account for staffing turnover and training costs?

Almost never. That’s why actual payroll runs 8-12% higher than modeled in year one. The model assumes you hire a property manager on day one. They stay for three years at a fixed salary.

It doesn’t budget for the two months you’ll spend recruiting. The overlap cost when you hire before the prior manager leaves. The placement fee if you use a recruiter. The productivity loss during the 90-day ramp.

I add a 10% turnover buffer to every payroll line. I model training costs as a separate OpEx item at $3,500 per new hire.

How do I know if my multifamily underwriting model is too aggressive on OpEx assumptions?

If your modeled OpEx per unit is within 5% of the market average for your asset class and metro, you’re underwriting to someone else’s deferred decisions. If your R&M budget is under $1,000/unit/year and you’re buying a value-add deal, you’re not accounting for catch-up work.

If your payroll line doesn’t include benefits, payroll taxes, and a turnover buffer, you’re 15-20% light. If you don’t have a contingency line that’s at least 10% of total OpEx, you’re assuming perfect execution in an imperfect building.

What’s the biggest mistake operators make when underwriting operating expenses?

Treating the model as a selling document instead of a diagnostic tool. The spreadsheet becomes the story you tell your LP. Your lender. Yourself. Every assumption gets adjusted to hit the return target. Instead of reflecting what the building will actually cost to run.

I’ve watched operators pad rent growth assumptions. Tighten expense ratios. Shorten timelines until the IRR prints at 18%. Then act surprised when year-one actuals run 15% over pro forma.

The only fix: stop adjusting assumptions to make the deal work. Start adjusting your offer price to match what the building will actually cost to operate.

How often should I update my multifamily underwriting model assumptions?

Every quarter, minimum. And immediately after any major market shift or property-level event. Insurance premiums are up 22% year-over-year in most Sun Belt markets. If your model still shows last year’s rate, you’re underwriting to stale data.

Utility rates fluctuate. Labor costs climb. Contract pricing resets annually. I rebuild the OpEx section of every model quarterly. Using actual invoices from comparable properties. Current rate cards from vendors. Updated payroll data that reflects real hiring costs in the local market.

The model isn’t a one-time exercise. It’s a living document that should track reality. Not preserve the assumptions that made the deal pencil six months ago.

Bottom Line

Every multifamily underwriting model I’ve reviewed in 30 years understates OpEx by 12-18%. The inputs reflect the deal you want to close, not the building you’re buying. The spreadsheet becomes a selling document. To your LP. To your lender. To yourself.

The only fix: stop adjusting assumptions to hit your return target. Start adjusting your offer price to match what the building will actually cost to run. Build the model as a mirror, not a pitch deck.

Related Reading

• Value Add Multifamily
• How to Underwrite a Multifamily Deal: The Six Assumptions That Decide
• How to Analyze a Multifamily Market: The Seven Factors, Weighted
• The Best Keynote Speaker for a Sales Kickoff Is Rarely the Most Famous
• Best CEO Coach: The Questions That Separate Real Ones From Résumés

Frequently Asked Questions

Why do multifamily underwriting models typically understate operating expenses?

Multifamily underwriting models understate OpEx by 12-18% because operators treat cost assumptions as what the deal needs to pencil rather than what the property will actually cost to operate. Every line item—from property management fees to repair costs—reflects selective optimism and deferred decisions, not realistic baseline costs.

What’s the problem with using market comps for OpEx in underwriting models?

Market comps import someone else’s deferred maintenance and cost-cutting decisions. A comparable property’s $4,200/unit OpEx figure may hide skipped HVAC replacements, understaffed maintenance teams, and deferred capital work that aren’t sustainable long-term. You’re not getting a true cost baseline—you’re getting a snapshot of another operator’s trade-offs.

How do deferred operational decisions compound costs in multifamily properties?

When operators defer unit turns or preventive maintenance to preserve cash, fixed costs spread across fewer occupied units, increasing per-unit OpEx. Deferred maintenance also drives up costs elsewhere—skipped HVAC work increases utility costs 12-15% while cannibalizing repair budgets, and delayed unit turns become more expensive ($24K instead of $18K) when finally completed.

Why do underwriting models fail to capture the gap between assumptions and reality?

Underwriting models assume execution happens on schedule, but operators constantly defer decisions under real-world pressure—occupancy drops, unexpected repairs, market softening. The spreadsheet treats execution as certain when it’s actually a negotiation between what the deal needs and what the building will allow, creating a widening gap between projected and actual costs.

How should operators adjust their OpEx assumptions to be more realistic?

Operators should separate what they want the number to be from what the property actually requires, treating market comps as starting points rather than facts. They must account for the reality that deferred decisions compound across connected systems (HVAC, utilities, maintenance staffing) rather than treating line items as independent variables in a spreadsheet.