Multifamily Lease-Up Underwriting: The Coverage Date

Published October 25, 2026

About 320,000 apartment units were still in active lease-up at the end of the third quarter of 2026 — down roughly 12% in the quarter and about 55% from a year earlier, per ALN Apartment Data's October 12 report. The pool of competing new product is shrinking, which reads as good news for anyone buying a recent delivery. What has not shrunk is the discount attached to the leases already signed inside that pool: about 26% of conventional properties were offering a concession to new residents at quarter-end, at an average of just over four weeks off an annual lease (per ALN).

Those two facts together are what makes a partially leased 2025 delivery one of the harder assets to price right now. The usual failure is not the stabilization assumption — most buyers model absorption reasonably well. The failure is that agency lease-up executions do not test stabilization. They test coverage on a specific date, and that date can arrive well before the concessions roll off the rent roll.

Why multifamily lease-up underwriting hinges on a coverage date, not a stabilization date

The two agency lease-up paths are structured around different timing, and both are tighter than they look.

Fannie Mae's Near-Stabilization Execution term sheet shows a minimum 75% physical occupancy at rate lock and certificates of occupancy for 100% of residential units as of the commitment date. The critical line is how coverage is defined: the underwritten Debt Service Coverage Ratio (DSCR) is the DSCR achievable within four months after rate lock, at Fannie Mae's discretion — 1.25x at Tier 2, or 1.15x for Multifamily Affordable Housing — with Loan-to-Value (LTV) capped at 75% of as-stabilized value, a minimum loan size of $10M, and monthly interest-only payments for the first 12 months.

Freddie Mac's Optigo Lease-Up Loan term sheet starts earlier in the lease-up — roughly 50% occupied, 60% leased, and 60% or more of certificates of occupancy issued at rate lock, with 100% at closing — and expects stabilization within 12 months of closing. Closing coverage can go as low as 1.05x, but the market-based minimum sits in the 1.25x–1.35x range depending on market, and an additional three-month debt service escrow may be required based on actual operations at underwriting. Manufactured housing communities are ineligible under that offering. Both sets of terms are subject to change — verify current parameters directly with the lender.

So multifamily lease-up underwriting has two deadlines, not one. Fannie's is a four-month coverage projection; Freddie's is a 12-month stabilization covenant backed by a reserve. Neither waits for the rent roll to clean itself up.

The concession tail does not clear inside four months

A free-rent concession is not a one-time marketing expense. It is written into a lease that typically runs 12 months, which means the discount sits in the effective-rent line until that lease renews at a clean rate. A property can take physical occupancy from the mid-70s into the high 80s over a quarter and still collect well below asking across nearly the whole rent roll.

The scale of the discount is the part models tend to understate. Per RealPage, about 15.4% of stabilized apartments were offering concessions nationally in August 2026 — up roughly 1.1 points year over year — at an average discount near 11% of asking rent, or about 5.7 weeks free on an annual lease. RealPage's September 2026 reading eased to about 14.5% of units, but the average discount held near 11.1%, roughly 5.8 weeks free: fewer properties are discounting, not shallower discounts. Colliers put first-quarter 2026 concessions at roughly 7.2% of asking rent across about a quarter of units. Zillow's spring 2026 data showed concessions on nearly 40% of listings on its platform, concentrated in metros including Denver (roughly 68%), Charlotte (roughly 67%), Dallas (roughly 64%), and Austin (roughly 64%). These are platform- and methodology-specific measures that vary by submarket, so they are useful as a range check rather than as a deal input.

For a lease-up asset the relevant number is worse than any of those averages, because every lease on the property was signed inside the concession window. There is no legacy cohort of clean leases diluting the discount.

A 280-unit 2025 delivery, two ways

Consider an illustrative deal: a 280-unit 2025 delivery in a high-supply Sun Belt submarket, 78% physically occupied and about 84% leased, asking $1,750 per unit per month, with eight weeks free on new leases. Other income is modeled at $95 per unit per month, and the contract price is $57M — below the as-stabilized value, reflecting lease-up risk. All figures below are illustrative, not comps.

Two line items do the damage. The concession line moves from a 1.5% stabilized allowance to roughly 13.5% of gross potential rent, because occupancy climbing to 88% means more concession-bearing leases, not fewer. And the expense line runs above stabilized levels during lease-up: full marketing spend, full onsite payroll, and real estate taxes already assessed on a completed building. That second item is the one most often copied straight from the stabilized column.

The result is an NOI about 35% below the as-stabilized number on the exact date the lender tests coverage.

The amortizing coverage test is where proceeds actually get set

Assume an illustrative 6.25% fixed coupon — plausible against a 10-year Treasury par yield near 5.24% on October 9, 2026, per the U.S. Treasury daily yield curve, though spreads move daily and should be quoted, not assumed. Interest-only service on that coupon is a 6.25% constant; a 30-year amortizing schedule at the same rate carries an approximately 7.39% constant. Running the four sizing tests against an as-stabilized value near $60.7M (the $3.49M NOI at a 5.75% cap, itself an assumption):

The headline is $45.5M. The amortizing test on fully stabilized income is roughly $37.8M — about $7.7M, or some 17%, lower. On the income the property actually produces at the coverage date, the same test indicates roughly $24.4M.

This is why the interest-only period is not a convenience. It is the only reason the deal sizes at all, and the resize risk is simply deferred to the month the interest-only window closes. Anyone sizing equity off the LTV line is understating the check by millions. The same discipline applies to DSCR and NOI on any stabilized acquisition — it just bites harder here, and the agency sizing framework is worth reading alongside the term sheets.

The lease-up credit enhancement is a deferred equity call

The Freddie path makes the resize risk explicit rather than leaving it to the refinance. Per the Optigo Lease-Up Loan term sheet, a lease-up credit enhancement may be required at Freddie Mac's discretion, sized at a minimum of 5% of the unpaid principal balance, or 10% when structured as a guaranty. It is released only once the property meets the required amortizing debt coverage ratio on a three-month average basis, with the most recent month's net rental income also at that level. If the required coverage is not reached within roughly 12 months, the enhancement is applied to resize the loan and recast payments.

On a $37.8M loan, a 5% enhancement is roughly $1.9M — capital that is committed at closing, earns nothing, and converts to a principal paydown if the absorption schedule slips. Against the $57M price, this illustrative deal needs roughly $11.5M of equity if the 75% LTV test governs, $19.2M if the stabilized amortizing test governs, and $32.6M if the month-4 amortizing test governs (before closing costs). The enhancement scales with the loan, so it is a larger share of a thinner equity check: about $2.3M, or 20% of equity, on the $45.5M loan; about $1.9M, or 10%, on the $37.8M loan; and about $1.2M, or 4%, on the $24.4M loan. That is a meaningful slice of the equity stack exposed to a timing miss, not a pricing miss — and largest precisely where the sponsor has taken the most proceeds. Modeling it as a contingent capital call with a date on it, rather than as a reserve, changes the return profile materially. The enhancement's terms are set deal by deal, so confirm the actual sizing and release conditions with the lender.

A related trap: the required release coverage is the amortizing ratio, which on this deal works out to roughly 1.06x at the interest-only-sized loan amount ($3.49M NOI against $44.7M × 7.39% ≈ $3.30M of amortizing debt service). A property can stabilize fully, hit its pro forma, and still fail the release test — because the test was never about stabilization.

Where this breaks

The honest counter-case is real, and it runs in both directions.

Absorption may beat the model. Net absorption still ran about 107,000 units in the third quarter of 2026 — down about 52% from the third quarter of 2025, but against deliveries that fell about 32% over the same period (per ALN) — and with lease-up stock down roughly 55% year over year, a 2025 delivery in a submarket with thin 2027 supply could stabilize faster than a four-month projection implies. That is a submarket-level question, not a national one — the same split described in 2026's two-speed market applies here.

The supply thinning may also be temporary. About 731,000 units were under construction at the end of the quarter, up roughly 17% year over year, with deliveries of approximately 383,000 units expected in 2027 against about 313,000 in 2026 (per ALN). A lease-up bought on today's competitive set could face a different one at renewal.

The rate assumption matters, but less than the timing. A 50 bps move in the coupon, from 6.25% to 6.75%, lifts the 30-year amortizing constant from about 7.39% to about 7.78% and cuts stabilized amortizing proceeds by roughly 5%, from about $37.8M to about $35.9M. Testing coverage on month-4 income instead of stabilized income cuts the same proceeds by roughly 35%, to about $24.4M. A buyer who negotiates hard on spread but copies the stabilized column into the coverage test has worked on the smaller of the two numbers. And bridge debt is a legitimate alternative to an agency lease-up execution for a buyer who would rather carry rate risk than covenant risk — with its own costs, as the rate cap math shows. Agency capacity is not the binding constraint in 2026 either, given the $176B combined cap.

Finally, sizing to an interest-only test is not automatically wrong. A sponsor with a genuine lease-up track record and a defensible absorption schedule may rationally take the higher proceeds. The point is pricing the resize, not avoiding it.

Lease-up deals are priced on absorption and sized on coverage, and those are different questions with different deadlines. The concession tail, the lease-up expense load, and the amortizing constant are the three inputs that decide whether the equity check in the model resembles the one at closing. Underwriting the as-stabilized column alone answers none of them. The occupancy-first revenue model is a useful starting frame; the coverage date is what it has to survive.

UWmatic is an AI-powered underwriting platform built for multifamily and mobile home park investors. Build the as-stabilized and coverage-date income statements side by side, size proceeds against interest-only and amortizing constants, and see where the equity check lands before rate lock. Try the free underwriting calculator →

This analysis reflects current market interpretations as of the publication date and may evolve as new data becomes available. Figures cited are drawn from public sources including ALN Apartment Data third-quarter 2026 occupancy, absorption and concession reporting, Fannie Mae and Freddie Mac multifamily product term sheets, U.S. Treasury daily par yield curve data, RealPage monthly concession data for August and September 2026, and Colliers and Zillow concession measures reported by Multifamily Dive, and are subject to revision. Nothing in this post is investment advice; readers should conduct their own diligence and consult qualified professionals before making investment decisions.