What is the difference between AI underwriting for Class A, B, and C multifamily? The difference lies in the operating assumptions, expense ratios, rent growth trajectories, and exit cap rate ranges that AI must apply to each asset class. A prompt that works for a 2020-vintage Class A property in a major metro will produce nonsense numbers if applied unchanged to a 1965-vintage Class C property in a secondary market. Sponsors using AI well calibrate the underwriting assumptions to the asset class before running the model. For pillar context, see our AI multifamily underwriting guide; for value-add specifics, see our piece on AI multifamily value-add underwriting.
Key Takeaways
- Class A multifamily underwriting assumes lower operating expense ratios, higher exit cap rate stability, and lower CapEx reserves; AI should default to 35 to 40 percent expense ratios and $250 to $400 per unit replacement reserves.
- Class B underwriting sits in the middle: 40 to 50 percent operating expense ratios, $400 to $600 per unit reserves, and rent growth that's more sensitive to local economic cycles.
- Class C underwriting requires the most conservative assumptions: 50 to 60 percent expense ratios, $600 to $1,000 per unit reserves, and significantly more bad debt and tenant turnover.
- AI should adjust exit cap rate ranges by class: Class A trades at 5.0 to 5.75 percent in primary markets, Class B at 5.75 to 6.75 percent, Class C at 6.75 to 8.5 percent depending on submarket and condition.
- The biggest AI mistakes happen when investors apply one class's assumptions to another class without recalibrating, often producing pro formas that look reasonable but mis-state cash flow by 20 to 40 percent.
What Class Actually Means in Multifamily
Class designation in multifamily is not a regulatory category. It's an industry shorthand that combines age, condition, amenities, location, and tenant profile. The standard breakdown is roughly as follows.
- Class A: Built within the last 15 years (in most markets) or fully renovated luxury. Pool, fitness center, controlled access, in-unit washer/dryer, granite countertops. Located in core or near-core submarkets with strong demographics. Tenants earn 4 to 6 times rent and have stable employment.
- Class B: Built between roughly 1985 and 2010, or older buildings with substantial mid-life renovation. Basic amenities, may have shared laundry. Located in stable middle-class neighborhoods. Tenants are working-class professionals, families, and white-collar wage earners.
- Class C: Built before 1985 (in most markets), often with deferred maintenance. Limited or no amenities. Located in lower-income or transitional neighborhoods. Tenants are hourly wage earners, voucher recipients, and price-sensitive renters.
The boundaries blur in practice. A 1980s Class B property that's been gut renovated with modern finishes can underwrite like Class A in a hot submarket. A 2005 Class A property that's been undermaintained for a decade can underwrite like Class B by year 20 of its life. AI should not blindly accept the broker's class designation; it should triangulate from age, location, amenities, and rent per square foot relative to comps.
How Operating Expense Ratios Change by Class
The operating expense ratio (OpEx as a percentage of effective gross income) is the single most consequential underwriting assumption that varies by class. Industry benchmarks from NMHC and IREM suggest the following ranges.
- Class A: 35 to 42 percent expense ratio. Lower maintenance, newer systems, professionally managed, in-unit washer/dryer reduces common-area utility cost.
- Class B: 42 to 50 percent expense ratio. Mid-life systems require more maintenance, mixed metering, more leasing turnover than Class A.
- Class C: 48 to 60 percent expense ratio. Higher repairs and maintenance, more vacancy, higher bad debt, higher legal and eviction costs, often master-metered utilities.
When AI underwrites a deal, the first sanity check should be whether the T12 operating expense ratio falls in the expected range for the asset class. If a property marketed as Class A is showing a 55 percent expense ratio in the T12, something is wrong: either the property is misclassified, the operating expenses include non-recurring items, or the income line has been overstated. AI should flag the mismatch and require a rebuild.
Rent Growth Assumptions by Class
Rent growth is not uniform across asset classes. Class A rents are more sensitive to new supply (because new construction is also Class A), but more resilient through recessions because the tenant base is higher-income. Class C rents are less affected by new supply but more affected by local employment shocks and policy interventions like eviction moratoriums or rent control.
A reasonable AI underwriting prompt by class:
- Class A: 2.5 to 3.5 percent annual rent growth in stable markets, with potential negative growth in oversupplied submarkets. Stress test years 1 and 2 with flat or modestly negative rent assumption.
- Class B: 3 to 4 percent annual rent growth in stable markets, 4 to 6 percent in growth markets. More resilient through cycles because the tenant base trades down from Class A in downturns.
- Class C: 2 to 3 percent annual rent growth, but with significantly more volatility year to year. Bad debt and concessions absorb meaningful portions of nominal rent growth.
CapEx and Reserve Assumptions
The reserve for replacement (RfR) line in the pro forma is a function of building age and system condition. AI should size this differently by class.
- Class A: $250 to $400 per unit per year. Systems are new, roofs have 15+ years of useful life, HVAC and appliances are under warranty for the first 5 to 10 years of hold.
- Class B: $400 to $600 per unit per year. Mid-life systems require active maintenance and selective replacement. Roofs, parking lot resurfacing, and HVAC compressor replacements hit during typical 5 to 7 year holds.
- Class C: $600 to $1,000 per unit per year, sometimes higher for properties with significant deferred maintenance. Plumbing reroutes, electrical upgrades, foundation work, and roof replacements are real and recurring.
For Class B and C value-add deals, the pro forma should separate the recurring RfR reserve from the upfront business plan CapEx. AI should treat these as two distinct line items. The business plan CapEx hits in years 1 to 2 (renovating unit interiors, upgrading common areas, addressing deferred maintenance), then the recurring reserve continues throughout the hold. Conflating these understates first-year capital requirements and overstates ongoing cash flow.
Exit Cap Rate Calibration
Exit cap rate is the largest single driver of long-term IRR and the assumption most often gamed in pro formas. AI should anchor exit cap to the going-in cap rate plus a class-specific adjustment for hold period and assumed market conditions.
- Class A: Exit cap typically equals going-in cap plus 25 to 50 basis points for a 5 to 7 year hold. In compressed-cap-rate environments, conservative underwriting adds 50 to 75 basis points.
- Class B: Exit cap typically equals going-in cap plus 25 to 50 basis points, with a 50 basis point conservatism layer if the going-in cap is below the 10-year average for the market.
- Class C: Exit cap typically equals going-in cap plus 50 to 100 basis points. Class C is most sensitive to liquidity contraction in downturns because the buyer pool is smaller.
According to recent industry analysis from JLL and CBRE, multifamily cap rates have stabilized in the high 5s to low 6s for stabilized Class A and B assets in primary markets, with Class C in secondary markets trading 100 to 200 basis points wider. AI should reflect this stratification, not apply a single national assumption.
Tenant Profile and Bad Debt Adjustments
Vacancy and credit loss assumptions also vary by class. Class A typically underwrites at 5 to 7 percent total vacancy and credit loss (with most being physical vacancy). Class B underwrites at 6 to 9 percent (with bad debt becoming meaningful). Class C can run 10 to 15 percent total economic vacancy in markets with high turnover, eviction backlogs, or significant voucher concentration.
AI should always cross-check the T12 actual vacancy against the assumed pro forma vacancy. If the T12 shows 12 percent vacancy on a property the broker calls Class B with 6 percent assumed pro forma vacancy, either the property is mismanaged, the class designation is wrong, or the pro forma is aggressive. Any of those conclusions changes the underwriting.
If you're ready to transform your multifamily underwriting process across asset classes, The AI Consulting Network specializes in exactly this kind of class-calibrated AI workflow.
The Right AI Prompt Template by Class
A class-aware AI underwriting prompt looks like: "This is a [year] vintage, [units] unit Class [A/B/C] multifamily property in [submarket]. Apply the following underwriting parameters: operating expense ratio in the [X to Y] percent range, rent growth of [X] percent, exit cap rate of [X] percent equal to going-in plus [X] basis points, reserve for replacement of $[X] per unit, vacancy and credit loss of [X] percent. Build a 5-year pro forma with these assumptions and stress test the exit cap rate at plus 50 and plus 100 basis points."
This kind of structured prompt produces underwriting output that's defensible at IC and reasonable to LPs. For cost benchmarks on the AI underwriting stack that supports class-calibrated workflows, see how much AI underwriting software costs for multifamily investors. For background on how AI is reshaping multifamily transactions, see NMHC Research.
Frequently Asked Questions
Q: How do I know whether a property is really Class A, B, or C?
A: Triangulate from age, location, amenities, rent per square foot vs market median, and tenant income relative to rent. Don't trust the broker classification without verification. AI can help by comparing the subject property to known comps in the submarket.
Q: Should AI use different cap rate assumptions for primary vs secondary markets within the same class?
A: Yes. Class B in a primary market trades 100 to 200 basis points tighter than Class B in a tertiary market. Geography matters as much as class.
Q: Why does the operating expense ratio matter so much?
A: Because NOI flows directly into both cap rate valuation and debt service coverage. A 5 percentage point error in expense ratio assumption can move NOI by 10 to 15 percent and valuation by the same amount.
Q: How does AI handle the transition from Class C to Class B through a value-add renovation?
A: The underwriting model should run two parallel sets of assumptions: pre-renovation (Class C expense ratios, lower rents, higher vacancy) for years 1 and 2, and post-renovation (Class B expense ratios, higher rents, lower vacancy) for years 3 onward.
Q: What's the single biggest mistake in cross-class AI underwriting?
A: Applying Class B assumptions to a Class C deal because the seller's broker pitched it as Class B. Always validate the class designation against the actual property condition and tenant profile.