The Coverage-Constrained Takeout: When the Permanent Loan Sizes to Coverage, Not Value

The permanent loan that retires a construction facility used to size to value. A lender took the stabilized net operating income, capitalized it into a value, and lent 70 to 75 percent of that number. For the whole low-rate decade, that loan-to-value ceiling was almost always the binding constraint, and the pro forma was built on it: the sponsor modeled the LTV loan, sized the equity to the balance, and moved on. At today's rates, that model is quietly wrong. A 1.25x coverage floor now sizes a smaller loan than the LTV ceiling would allow, so the takeout sizes to coverage, not value, and the difference is real equity the sponsor never budgeted for.

This is not a small technical footnote to a financing plan. It is the mechanism reshaping the capital stack of every ground-up deal underwritten before rates moved, and it behaves in a way that inverts the usual intuition about which assets are most exposed. What follows is the arithmetic of the coverage-constrained takeout, the point at which it takes over from value, the asset classes it hits first, and the discipline a 2026 feasibility study has to adopt so the equity gap is on the page before the lender finds it.

The three lenses a takeout lender sizes through

Every takeout lender sizes a loan through three lenses at once and funds the smallest of the three. The first is coverage. Debt service coverage is net operating income divided by annual debt service, and the lender sets a floor, commonly 1.25x. Rearranged, the largest loan coverage will support equals net operating income divided by the product of the coverage floor and the debt constant, where the debt constant is the annualized debt service per dollar borrowed. The second is value. The maximum loan by loan-to-value is stabilized value times the LTV ceiling, and because value is net operating income divided by the cap rate, that loan equals net operating income over the cap rate, times the LTV. Unlike the coverage loan, it does not move with the coupon; it moves with the cap rate and the value. The third is debt yield: net operating income divided by the loan amount, a rate- and amortization-proof figure the lender floors at, say, 9 or 10 percent. The lender computes all three and lends the least. Whichever produces the fewest proceeds is the binding constraint.

The three lenses are linked by an identity worth holding onto: the debt yield of a coverage-constrained loan is exactly the coverage floor times the debt constant. At a 1.25x floor and a 7.73 percent constant, that is 1.25 × 0.0773 = 9.66 percent. When the coverage floor and the debt constant multiply to a number above the lender's debt-yield floor, coverage is doing the binding; when they fall below it, debt yield takes over. That single identity is why, on high-cap-rate assets, the debt-yield floor is the constraint that actually bites, a point we return to below.

Why flat income now supports a smaller loan

The mechanism that rewired the math is the debt constant. For a fully amortizing loan paid monthly, the annual constant is 12 × [ (i / 12) ÷ (1 − (1 + i / 12)⁻ⁿ) ], where i is the annual note rate and n is the number of monthly payments — 300 on a 25-year amortization, 360 on a 30-year. The constant rises with the rate, and it is the reason a lender's coverage loan shrinks even when net operating income has not moved a single dollar.

Read the ladder against recent history. A pandemic-era coupon near 4 percent carried a 25-year constant of roughly 6.3 percent; a 7 percent takeout carries a constant of 8.48 percent. That is an increase of roughly 35 percent in debt service per dollar borrowed, on flat income. A property whose net operating income comes in at exactly what the low-rate pro forma projected still supports about 35 percent less coverage-based debt, purely because each borrowed dollar now costs more to service. That single fact is the engine of the rewired math.

The crossover between value and coverage

There is a clean way to know, for any given deal, which lens binds. Set the value loan equal to the coverage loan and solve. The equation (NOI ÷ cap) × LTV = NOI ÷ (DSCR × constant) reduces to a single threshold: the crossover constant equals the cap rate divided by the product of the LTV and the coverage floor. If the actual debt constant sits above that crossover, coverage binds and the loan is smaller than value would allow. If it sits below, value still binds. To read the crossover as a rate rather than a constant, find the note rate whose constant equals it on the relevant amortization ladder.

The implication is the analytical heart of this piece: the crossover constant rises with the cap rate. A high cap rate produces a high crossover, so coverage does not bind until rates are very high and value stays the constraint. A low cap rate produces a low crossover, so coverage binds even at modest rates. Low-cap-rate property, in other words, hits the coverage wall first — the opposite of where most underwriters instinctively look for rate stress.

A worked example: the gap that was never in the model

Put numbers on it. Take a stabilized net operating income of $1,000,000 and a 6.0 percent cap rate, for a value of $16,666,667. A 75 percent LTV ceiling supports a $12,500,000 loan. Apply a 1.25x coverage floor on a 25-year amortization. The crossover constant is 0.06 ÷ (0.75 × 1.25) = 6.40 percent, and every constant on the 25-year ladder, starting at 7.02 percent for a 5 percent rate, sits above it. This deal is coverage-constrained across the entire current rate ladder.

At a 7 percent takeout, the coverage-supported loan is $9,433,000, which is 25 percent below the $12,500,000 the LTV ceiling would allow. The developer has to fund the $3,067,000 difference in equity, a gap a low-rate-era pro forma, which assumed the 75 percent loan would fund, would never have flagged. And the gap widens with every increment of rate: a full point higher, at 8 percent, the required extra equity is $3,862,000, and the coverage loan is barely two-thirds of the value loan.

The reversal: low cap rates take the hardest hit

This is where the conventional intuition inverts. The instinct is that high-cap, higher-risk assets, the hotels and older office, are the ones squeezed hardest by the coverage constraint. The algebra says the opposite. Hold LTV at 75 percent and the coverage floor at 1.25x, and the crossover constant tracks the cap rate directly: a 5.5 percent cap yields a 5.87 percent crossover, below a 5 percent rate on either ladder; a 6.5 percent cap yields 6.93 percent, about a 5 percent rate on a 30-year schedule; a 7.0 percent cap yields 7.47 percent, near a 5.7 percent rate on a 25-year; an 8.0 percent cap yields 8.53 percent, near a 7.1 percent rate. The higher the cap rate, the further rates have to climb before coverage binds at all.

Quantify it on a common $1,000,000 of net operating income. Multifamily at a 5.5 percent cap, 75 percent LTV, 1.25x coverage, and a roughly 6 percent constant on a 30-year amortization capitalizes to $18.18 million of value and a $13.64 million LTV loan, but a coverage loan of only about $10.35 million, a gap of roughly $3.29 million, 24 percent below the value loan. A hotel at an 8.5 percent cap, 65 percent LTV, 1.45x coverage, and a 9.26 percent constant at an 8 percent rate capitalizes to $11.76 million of value and a $7.65 million LTV loan, against a coverage loan of about $7.45 million, a gap of just $0.20 million, or 2.6 percent. And because the hotel's 10 to 11 percent debt-yield floor caps its loan near $9.1 to $10.0 million, that floor never binds either; the high cap rate has already shrunk the value loan so far that LTV and coverage effectively co-bind. The coverage-driven equity gap is roughly an order of magnitude larger for low-cap multifamily than for high-cap hotel.

An asset-class map of the binding constraint

Generalize the pattern across property types, drawing cap rates from the CBRE H2 2025 U.S. Cap Rate Survey, CRED iQ's CMBS data, Matthews, and Cushman & Wakefield, and loan terms from the Freddie Mac and Fannie Mae multifamily guides, the PeerSense and Janover CMBS references, and Select Commercial.

The current readings anchor the map. CBRE's H2 2025 survey, published in February 2026, found the all-property cap rate holding steady and believed to be at its cyclical peak, with 2025 transaction volume up about 19 percent and debt becoming more available at higher LTVs. Multifamily sits around 5.6 to 5.9 percent; apartment transactions averaged 5.7 percent across 2025, the longest such plateau in 25 years. CRED iQ's CMBS deals in early 2025 show office at 7.4 percent, multifamily 5.9, retail 6.7, industrial 6.4, self-storage 6.2, and hotel 7.3. Cushman & Wakefield and Matthews put first-quarter 2026 hotels near 8.2 percent on average, from 6.48 percent for luxury metro product to 8.60 for flagged economy, and office single-tenant NNN at 7.9 percent, Class A at 7.6, Class B at 8.0, and Class C between 8.7 and 9.4. Senior housing averaged about 6.8 percent in mid-2025, with assisted living near 7 percent and memory care near 9.6, which PGIM notes prices roughly 150 basis points wide of apartments.

The pattern is unambiguous. Multifamily, industrial, and self-storage carry crossover constants around 5.9 to 6.9 percent, below every point on the current constant ladder, so they are firmly coverage-constrained and carry the largest equity gaps. Agency multifamily is the sharpest case: a roughly 5.5 to 5.9 percent cap and a 1.25x floor, with Freddie Mac at 1.20x on select 10-year fixed-rate debt and 1.15x on floating debt tested at the capped rate. Office and hotels sit at the other end, with crossover constants near 9.8 to 9.9 percent, close to the top of the ladder, and 10 to 12 percent debt-yield floors that typically bind first; their high cap rates have already shrunk the value loan, so coverage rarely gets the chance to bind. Retail lands in between, coverage- or LTV-bound depending on the tenant, with credit-tenant NNN leases sizing to floors as low as 1.05 to 1.10x. Senior housing would sit with the coverage-bound low-cap classes on its roughly 6.0 to 7.5 percent cap, except that HUD 232's 35- to 40-year amortization lowers the debt constant and softens the bind, even at a 1.45x floor. The rule the map draws is simple: coverage binds first where cap rates are low; value and debt yield bind first where cap rates are high.

The rate shock behind the arithmetic

None of this would matter if rates had not moved. They have. As of early July 2026 the 10-year Treasury sits around 4.38 percent on June 26 and 4.48 percent on July 7, with an end-June weekly average near 4.44 percent per Advisor Perspectives, down from a peak near 4.8 percent in January 2025 and about 4.5 percent in July 2025. One-month Term SOFR and 30-day average SOFR both run about 4.14 percent, per Pensford in early July 2026. The Federal Open Market Committee has held the funds rate at 3.50 to 3.75 percent; its April 29, 2026 decision was the fourth consecutive hold, on an unusually divided 8-to-4 vote, and the lowest level since November 2022. Permanent commercial mortgage rates, per Select Commercial on July 9, 2026, start around 5.57 percent for multifamily and 6.48 percent for CMBS, with PeerSense putting CMBS spreads over the roughly 4.31 percent 10-year Treasury at 190 to 215 basis points for multifamily, 200 to 225 for commercial, 320 to 345 for hotel, and 200 to 225 for self-storage. Construction loans, which float, average about 8.4 percent in 2026 per Buildermuse, priced at SOFR plus a spread or prime plus 1.00 to 1.50 percent at community banks, with private and bridge capital at 9.75 to 12 percent.

Measured against the trough, the move is violent. The 10-year Treasury bottomed at a record 0.55 percent in August 2020; multifamily cap rates bottomed near 4.2 percent in late 2021; pandemic-era commercial coupons ran 3 to 4 percent. New commercial originations now average about 6.24 percent against roughly 4.76 percent on the debt they are refinancing, per CRE Daily, a 150 basis-point refinancing shock against maturing loans and a 300-plus basis-point shock against the 2020–21 trough in which many construction deals were first conceived.

The hedging squeeze on floating-rate construction debt

Because construction loans float, the borrower typically has to buy an interest-rate cap, and the cost of that protection exploded alongside the rate move — compounding the takeout problem by eroding the net operating income available for debt service during the construction and bridge period and by front-loading a large cash outlay just as the equity gap opens. The scale is easiest to see through the CRE CLO market. CRED iQ, via Commercial Observer on March 27, 2023, analyzed interest-rate-cap agreements on nearly 700 floating-rate loans securitized in commercial real estate collateralized loan obligations, covering more than $30 billion in aggregate notional balance. The price moves underneath that figure were extreme. CRED iQ cited a $337.5 million agreement expiring in February 2024 at a 4 percent strike that would have cost approximately $780,000 a year earlier and had risen to $3.5 million by mid-March 2023, a 348 percent year-over-year increase. On a more typical deal, Chatham Financial's Matt Moore told Commercial Observer that a two-year cap on a $25 million loan at a 4 percent strike had risen to $569,000, from $97,000 a year earlier. Chatham, the largest independent commercial-real-estate derivatives advisor, places more than 6,000 caps a year and is a fair gauge of a market in which floating-rate debt is roughly one-third of all commercial real estate debt, per the Mortgage Bankers Association.

The maturity wall and the construction-to-perm gap

The coverage constraint is arriving into a refinancing environment already under strain. The Mortgage Bankers Association's February 2026 read has about $875 billion of commercial and multifamily debt, 17 percent of roughly $5.0 trillion outstanding, maturing in 2026, down 9 percent from about $957 billion in 2025, with another $652 billion in 2027. S&P Global Market Intelligence, on a broader definition, counts about $998 billion maturing in 2025, $1.148 trillion in 2026, and $1.257 trillion in 2027 — cumulatively well over $1.5 trillion across the three years. The cushion that carried the market this far is thinning: only about 50 to 55 percent of 2025 maturities are estimated to have actually paid off, the rest rolled forward, and lenders have signaled that the extend-and-pretend practice is largely over. Industry estimates put the resulting refinancing shortfall at $300 to $500 billion, per C2R Capital.

Office is the epicenter. About $148 billion matures in 2026, and per Morningstar DBRS in January 2026, more than half of the roughly $100 billion in securitized commercial mortgages coming due this year are unlikely to pay off at maturity, against payoff rates that topped 80 percent in 2023 and ran about 75 percent in 2024 and 2025. Trepp reported office CMBS delinquency rising 103 basis points to an all-time high of 12.34 percent in January 2026, past the prior peak of 11.76 percent in October, with overall CMBS delinquency up 17 basis points to 7.47 percent; first-quarter 2025 distressed volume reached $116 billion, up 31 percent year over year.

For a ground-up deal, the same dynamic plays out at takeout. Construction loans size to cost, and loan-to-cost ceilings have tightened, to roughly 60 to 65 percent for multifamily and 55 to 60 percent for commercial, down from 70 to 75 and 65 to 70, at floating rates. When the coverage-constrained permanent loan comes in smaller than the outstanding construction balance, the developer faces a takeout gap that has to be filled with fresh equity or subordinate capital: the construction-loan analogue of the refinancing wall.

Filling the gap: mezzanine and preferred equity

When the coverage loan lands below both the LTV ceiling and the outstanding construction balance, the capital stack has to be rebuilt from the middle. As of mid-2026, mezzanine debt prices all-in around 11 to 16 percent, with construction mezzanine at 15 to 20 percent and higher; preferred equity runs about 12 to 18 percent, or 11 to 16 for more common structures. Senior debt stays near 60 to 65 percent of value or cost, and mezzanine or preferred capital lifts the combined stack to roughly 75 to 80 percent on stabilized assets and 85 to 90 percent of cost in development, but never to 100 percent, and total leverage above about 85 percent of cost generally exceeds what either product will support. Private credit is now the dominant provider of this gap capital. One structural detail governs relative pricing: mezzanine prices about 100 to 200 basis points inside comparable preferred equity, because a UCC Article 9 foreclosure on the pledged equity clears in 30 to 60 days while preferred equity's contractual remedies can take 6 to 18 months or more.

Underwriting to a stressed takeout, not going-in value

The conventions are tightening in the same direction the math points. Office CMBS debt-yield minimums have risen from a historical 8 to 9 percent to 10 to 12 percent today, and office LTV ceilings have fallen from 70 to 75 percent to 55 to 65, per Janover and PeerSense; CBRE reported debt yields surging 90 basis points to 10.3 percent in the first quarter of 2025. Trepp notes that 2026 underwriting recalibrates the debt-service test to current coupons and rejects future rent-growth assumptions. Debt yield — net operating income over loan amount, immune to the rate and amortization games that can flatter a coverage calculation — is increasingly the quiet binding constraint on high-cap-rate assets, precisely where coverage does not reach.

For the analyst, the discipline follows directly. A pro forma underwritten in the low-rate era assumed the LTV ceiling would bind at going-in value; a rigorous 2026 feasibility study has to model the coverage-constrained takeout explicitly. Run the coverage loan at the projected takeout rate and amortization — the amortizing constant, whatever the actual payment structure — and set it beside both the LTV loan and the outstanding construction balance. Quantify the equity gap, and price the mezzanine or preferred capital required to fill it. For low-cap assets (multifamily, industrial, self-storage) the coverage test is now the decisive sizing constraint; for high-cap assets (office, hotel) the debt-yield floor has to be modeled alongside LTV. The firm's differentiator is running all three lenses — LTV, coverage, and debt yield — at stressed takeout rates, not the LTV loan at going-in value.

One caution belongs on every such model. The rate levels above are mid-2026 readings, the FOMC's last decision split 8 to 4, and markets price one or two possible cuts before year-end 2026. A forward takeout rate is a scenario, not a fact, and the coverage loan should be sized across a band of them rather than a single point. That is the whole purpose of testing the takeout before the lender does: to know where the equity gap opens, and how large it grows, while there is still time to fund it.