Two numbers sit in every multifamily model: projected rent growth and projected insurance expense. Both are assumptions, but they are not the same kind of assumption. Rent growth may or may not occur. Insurance has to be purchased, on whatever terms the market is willing to offer. The practical rule that follows is straightforward: do not use hoped for revenue growth to absorb an expense estimate that has not been independently validated.
Revenue is optional in the forecast, expenses are contractual in the operation
Market rent growth depends on supply, demand, concessions, resident affordability, product quality, and competing inventory, none of which the owner controls directly. Insurance, taxes, utilities, payroll, repairs, and servicing contracts are obligations required simply to operate the property. A missed rent growth forecast reduces upside. An understated expense can impair debt service and distributions starting in year one.
Expense growth is also asymmetric in a way rent growth is not. Premiums move at renewal, often without warning. Deductibles and exclusions can shift more risk onto ownership even when the stated premium looks stable. A tax reassessment can create a step up that never reverses. A quoted premium backed by a complete policy analysis matters more than a broker's directional estimate, because the estimate is a guess and the quote is a number the carrier will actually honor.
Insurance expense is more than the annual premium
A full underwriting of insurance covers the property premium, general liability, excess liability, flood, wind, named storm and equipment breakdown coverage, business interruption limits and waiting periods, ordinance and law coverage, replacement cost assumptions, deductibles, and exclusions and sublimits. The insurer's valuation assumptions need to be checked against current replacement cost, and the lender's coverage requirements need to be compared against whatever policy the current owner is carrying, since the two rarely match exactly. Loss runs matter as much as the quote itself: open claims and prior loss history tell you what the next renewal is likely to look like. A property can be technically insurable and still carry a deductible large enough to create a real liquidity obligation. A lower premium paired with a materially higher named storm deductible is not a lower economic cost. It is a different distribution of the same cost, shifted toward the owner's balance sheet instead of the carrier's.
Texas and Florida require different expense diagnostics
In Texas, wind, hail, convective storm, Gulf Coast, flood, freeze, and roof age exposure vary materially by location and construction type, and property tax has to be modeled from a post closing value estimate and the applicable local tax rate rather than carried forward from the seller's number. Texas appraisal districts value taxable property at market value as of January 1 and must reappraise at least once every three years, so an acquisition resets the clock. The state's temporary non-homestead circuit breaker, which limits annual appraisal increases, applies for 2026 only to real property valued at $5.32 million or less and is scheduled to expire December 31, 2026, which means it generally will not protect an institutional scale multifamily asset.
In Florida, coastal location, wind zone, flood zone, roof condition, building age, construction type, and loss history all affect coverage terms, and improving headlines about personal lines reinsurance should not be assumed to apply to commercial apartment policies. Florida non-homestead property can carry a 10% annual assessment increase limitation, but a sale, foreclosure, transfer of beneficial title, or qualifying change of control resets the property to just value, so acquisition underwriting has to estimate the post transfer tax base rather than extend the seller's capped assessment. In both states, a statewide headline is not a substitute for the property's own address, construction, loss experience, and policy terms.
A small expense miss requires a large amount of additional value to offset
The Federal Reserve found that average inflation adjusted multifamily insurance expense rose from roughly $39 per unit per month in 2019 to $68 in 2024, an increase of more than 75%. Yardi Matrix reported average operating expenses near $8,950 per unit as of January 2024, up 7.1% year over year, with insurance the fastest growing major category, and a related Yardi analysis put apartment insurance growth at 27.7% year over year through the same period. At income restricted Houston properties, Yardi reported a 41.5% year over year increase in insurance expense through September 2024, a figure specific to that affordable housing subset rather than the broader commercial market.
The math on a miss is unforgiving. A $250 per unit annual insurance underestimate on a 250 unit property reduces NOI by $62,500. At a 5.25% capitalization rate, that recurring difference corresponds to roughly $1.19 million of value. A $1,000 per unit annual tax underestimate on the same 250 units reduces NOI by $250,000, worth roughly $4.76 million at the same cap rate. Compare that against the incremental NOI from rent growth, which has to first survive vacancy, concessions, and bad debt before it reaches the bottom line, while an expense increase reduces NOI dollar for dollar with no offsetting friction. That asymmetry is the entire argument for testing expense shocks before adding revenue upside to a model.
Conservative underwriting is an evidence standard, not a lower number
Requiring a current bound quote instead of relying on a historical premium, normalizing taxes to the expected post acquisition assessment, and pulling utility bills, payroll registers, contracts, and multi year trailing history are what separate a defensible expense base from a seller's budget. Controllable and uncontrollable expenses get modeled separately, each run through a base case built on the documented quote and current tax estimate, a conservative case built on an adverse but plausible renewal and reassessment, and a stress case built around a major deductible or delayed claim reimbursement. An insurance refund, a successful tax protest, or a premium decline does not enter the model until the result is actually supportable, not anticipated.
The Emeth standard
A generic sponsor might describe conservative underwriting as simply using a lower rent growth rate. That is not the same thing as conservative underwriting. A 2% rent growth assumption is not conservative if the insurance line is an unverified seller budget, the tax line still reflects the seller's capped assessment, and the model never accounts for the named storm deductible. Real conservatism starts with the quality of the evidence behind each number, not the size of the number itself.
Rent growth can improve an investment. It should never be asked to rescue an expense line that was never properly measured. Emeth verifies the recurring costs first, funds the nonrecurring exposures second, and only then evaluates what revenue growth would add on top.