Mixed-use underwriting is not complicated because the individual components are hard to analyze. Office, retail, and multifamily each have well-understood valuation frameworks. The complication is that a single mixed-use asset contains all three at once — and the cap rate, income stability, and financing assumptions that work for one component actively contradict the assumptions you need for another. In our work with acquisition teams underwriting multi-component assets, we've found this collision point is where most models break.
The Component Stacking Problem
A prototypical mixed-use building might have ground-floor retail on NNN leases, office floors 2-6 on full-service gross leases, and residential units on floors 7-12. Each component has its own income trajectory, lease structure, vacancy profile, and market cap rate. The retail might trade at a 6.5% cap, the office at 7.2%, and the multifamily at 4.8%. If you blend those three into a single overall cap rate using a simple income-weighted average, you get a number that is technically derived from real data and almost completely useless as a valuation guide.
Why? Because a blended cap rate hides the risk segmentation. The multifamily component might represent 60% of NOI but carry the tightest financing terms and lowest exit cap — meaning any underperformance in the retail or office segment creates disproportionate debt service pressure without being visible in the blended figure. We've seen models where a 15% retail vacancy swing reduced overall NOI by only 4% — until you ran the financing, at which point the DSCR dropped from 1.28x to 1.11x because the debt was sized against the full blended NOI without a retail stress haircut built in.
Lease Structure Conflicts Across Components
Each component type carries different baseline lease structures, and those structures create conflicting underwriting requirements within a single asset.
| Component | Typical Lease Type | Expense Treatment | Vacancy Assumption (typical) |
|---|---|---|---|
| Ground-Floor Retail | NNN / Modified Gross | Tenant pays most expenses | 8-12% (market-dependent) |
| Office | Full-Service Gross | Landlord absorbs all OpEx | 12-20% (Class B/C) |
| Multifamily | Gross (annual term) | Landlord pays building OpEx | 4-7% (stabilized) |
The issue isn't that any single row is hard to model. It's that your waterfall calculations — management fees, insurance allocations, shared HVAC, common area maintenance — need to be split across components with different expense treatment rules. If you apply a single expense ratio to total gross revenue, you will almost certainly over-expense the NNN retail (which doesn't carry landlord OpEx) and under-expense the office (which does). In practice, we see this error producing NOI variances of 6-10% on mixed-use assets with meaningful retail or office components.
Capital Structure Complications
Lenders treat mixed-use collateral differently depending on how they categorize the dominant use, and their answer can change your available loan-to-value by 5-10 percentage points. A building that is 55% multifamily by square footage but 60% office by income might qualify for agency debt financing if underwritten as multifamily — or conventional commercial debt if the lender considers it primarily office. Those two outcomes carry meaningfully different amortization schedules, IO period availability, and prepayment flexibility.
The financing path matters to the underwriting model because the debt assumptions flow directly into cash-on-cash return and equity multiple projections. An acquisition team that runs one model assuming agency debt and presents to IC without flagging that agency eligibility depends on how the asset is classified is taking a risk that may not surface until the lender's term sheet arrives. In our experience, financing assumption conflicts on mixed-use deals are one of the top three reasons IC approvals require a second-round revision.
Tenant Improvement and Leasing Commission Allocation
TI (tenant improvement) and LC (leasing commission) assumptions also diverge sharply by component. Office re-leasing TI in a post-2022 market typically runs $50-$90/SF for Class B space in major metros, depending on tenant fit-out requirements and lease term. Retail TI is often lower on NNN deals but can spike for food and beverage or experiential tenants. Multifamily has essentially no TI but does have turn costs.
When an analyst builds a single per-unit or per-SF TI and LC assumption for a mixed-use asset without segmenting by component, the model either understates the office leasing cost or overstates retail — or both simultaneously. The IRR impact depends on the lease expiry schedule. A building with 40% of its office space rolling in years 2-4 of the hold period will have a very different returns profile than one with long-term office WALTs, and that difference needs to be visible in the component-by-component capital expenditure waterfall, not smoothed into a single line.
Market Intelligence Requirements Are Multiplied
A standard single-use acquisition requires submarket comps for one asset class. Mixed-use requires three separate comps analyses running in parallel — retail leasing activity in the immediate trade area, office submarket absorption and net effective rents, and multifamily vacancy and effective rent trends for that specific submarket and unit mix.
In our data, the retail and office comp research alone accounts for 60-70% of the total research time on mixed-use deals because both markets are segmented by micro-location, floor plate, and tenant mix in ways that multifamily isn't. A restaurant tenant's willingness to pay bears almost no relationship to what a financial services firm will pay for the same square footage on the same block. Running credible comps for three different tenant pools simultaneously is the point at which most deal teams either short-cut the analysis or slow down significantly on deal velocity.
The practical issue we hear most often from deal teams: mixed-use deals take 1.5 to 2x longer to underwrite than single-use assets of equivalent size. That's not a skill gap — it's a structural complexity problem that compounds under pipeline pressure.
Risk Correlation Between Components
One of the less-discussed complexities in mixed-use underwriting is that the components are not risk-independent. Ground-floor retail vacancy in a building with struggling office tenants tends to rise — because foot traffic from office workers is part of what supports retail viability. The office market stress that created the vacancy created the retail stress. You can't model them as independent Bernoulli trials with separate probability distributions. They're correlated, and your stress scenarios need to reflect that correlation.
A bear case that applies a 20% retail vacancy assumption and a 15% office vacancy assumption independently is likely more optimistic than the data supports, because they're not independent events. In reality, a submarket with 20% retail vacancy probably also has 20%+ office availability, not 15%. The correct bear scenario runs them together with a correlation coefficient derived from comparable submarket histories — a quantitative step that falls outside what most Excel models can handle cleanly.
Practical Recommendations for Mixed-Use Underwriting
Based on what we've observed across mixed-use deal pipelines, the following structure produces more defensible underwriting than a blended approach:
- Build three separate income modules. Retail, office, and residential each get their own vacancy, rent growth, and lease-up assumptions — then sum to total NOI at the property level.
- Apply component-specific cap rates for exit valuation. Weight the exit by income contribution from each component rather than using a single blended exit cap.
- Run separate TI and LC schedules per component. Link them to the individual lease expiry schedules in each module.
- Model financing eligibility explicitly. Determine lender classification of the asset before sizing debt. If agency eligibility is at risk, run a commercial debt scenario in parallel.
- Correlate stress scenarios. Don't apply independent vacancy stresses to retail and office. Use a shared downside scenario where both rise together.
None of these steps are conceptually difficult. The barrier is time and data assembly. Having submarket comps for three asset classes available simultaneously, along with pre-built component modules that don't require rebuilding from scratch per deal, is what separates teams that underwrite mixed-use well from teams that rush it.
The Bottom Line
Mixed-use properties trade at premiums to single-use assets in many markets precisely because they're harder to underwrite. That complexity discount is real. The acquisition teams that consistently win on mixed-use deals are the ones who have systematized the three-component analysis to a degree where they can run defensible numbers as fast as their competitors run simplified ones. The models that break aren't wrong in their formulas — they're wrong in their structure. Fixing the structure is where the value is.
