The first DSCR loan I underwrote wasn’t called a DSCR loan. It was just a CMBS loan on a suburban office building, and the only number the lender’s underwriter cared about was whether net operating income, as they calculated it, would cover debt service by at least 1.30 times.
Their NOI was meaningfully lower than the broker’s NOI. The deal still penciled, but the loan sized smaller than we’d modeled. We had to bring extra equity to closing.
That gap — between the NOI a seller quotes and the NOI a lender uses to qualify a loan — is the entire story of DSCR underwriting. Every “DSCR loan” in commercial real estate, whether it’s a $480,000 single-family rental from Kiavi or a $185 million CMBS execution from Wells Fargo, runs the same test: property cash flow divided by annual debt service, against a minimum threshold the lender will not move off.
This guide explains what a DSCR loan actually is, how lenders calculate the coverage ratio, the thresholds you’ll see across product types, what counts as NOI for DSCR purposes, and the most common reasons institutional deals fail the test.
I spent 10 years on the buy side at Stockbridge Capital, the last five as Director of Research. Over that stretch we closed more than $2B in CRE. Almost every deal involved a DSCR test we had to clear, and the ones that broke down all broke down the same way. The rest of this article is the version of DSCR I wish I’d had handed to me on day one.
What a DSCR loan actually is
A DSCR loan is a mortgage that qualifies the borrower on the property’s debt service coverage ratio — net operating income divided by annual debt service — rather than on personal income, employment history, or tax returns.
In CRE this isn’t a separate product category. Every commercial mortgage is, in effect, a DSCR loan. The lender’s underwriter computes NOI, sizes debt service against a coverage threshold, and sets the loan amount accordingly. The borrower’s personal balance sheet matters for guaranties, recourse carve-outs, and net worth and liquidity covenants — but it doesn’t drive the loan size. The property does.
In the residential investor market, “DSCR loan” is a specific non-QM product sold by specialty lenders like Kiavi, Visio Financial, RCN Capital, Lima One, CoreVest, Easy Street Capital, and Truss Financial Group. These loans cover single-family rentals, 2-4 unit properties, and small multifamily up to 10 units. They were built for buy-and-hold investors who don’t want to document personal income, who already own multiple properties (which complicates conventional qualification), or who hold rentals in LLCs.
Both ends of the market use the same formula. They underwrite very different deals.
How DSCR is calculated for loan qualification
The formula:
DSCR = Net Operating Income / Annual Debt Service
A 1.25x DSCR means NOI covers debt service 1.25 times — there’s a 25% cushion above breakeven. A 1.00x DSCR means the property exactly breaks even at the debt service level. A 0.95x means the property doesn’t cover debt service from operations and the borrower has to fund the gap.
The math is trivial. The two inputs aren’t.
Net operating income
For DSCR purposes, NOI is annual revenue minus annual operating expenses, before debt service, before depreciation, before capital expenditures, and before income taxes. Lenders almost never use the seller’s NOI. They compute their own — and on stabilized commercial deals it’s usually 5-15% lower than the offering memorandum NOI.
The lender adjustments are predictable. I’ll cover them below.
Annual debt service
For fixed-rate loans, this is principal plus interest payments over a 12-month period at the proposed rate and amortization. For interest-only loans, it’s just interest (which makes DSCR look better — lenders know this, which is why most also test against a fully-amortizing constant or a debt yield). For floating-rate loans, lenders stress-test the rate, typically using the strike rate of a required interest-rate cap plus a cushion.
Most CMBS and balance-sheet lenders use a 30-year amortization to compute “constant” payment levels even on shorter-term loans, because the goal is to test whether the property could service a typical takeout loan. The constant is the all-in periodic payment per dollar of principal, annualized.
Typical DSCR thresholds by lender type
Coverage requirements vary widely. Here’s what I see in 2026:
Non-bank residential DSCR lenders (Kiavi, Visio, RCN, Lima One, CoreVest): 1.00x to 1.25x on most products. Some will lend down to 0.75x with rate adjustments and lower LTV. SFR and small-MF investors. Loan sizes $75K to $3M typically, with portfolio products up to $30M.
Agency multifamily (Fannie Mae DUS, Freddie Mac Optigo, HUD): 1.20x to 1.30x on conventional. Affordable-housing tranches often go down to 1.15x with additional credit enhancement. Run by Berkadia, Walker & Dunlop, Greystone, JLL, CBRE, and PGIM among others.
Bank balance-sheet on commercial (JPMorgan, Wells Fargo, Bank of America, M&T, Synovus): 1.20x to 1.30x on stabilized assets, sometimes 1.15x on multifamily. Tighter on office and retail in 2026 given the rate cycle.
Life companies (MetLife, Prudential, New York Life, Northwestern Mutual, MassMutual, TIAA): 1.25x to 1.35x on core, sometimes higher on transitional credits. Often the most conservative on NOI inputs but the most competitive on rate.
CMBS (Goldman Sachs, Morgan Stanley, Citi, Deutsche Bank, Wells Fargo securitization): 1.20x to 1.25x typical, 1.30x on retail and office, 1.40x+ on hotels. Standardized underwriting per the rating agency presale process.
Bridge and value-add (Madison Realty Capital, Mesa West Capital, BRT Apartments, Square Mile Capital, Slate Asset Management): Usually structured around a stabilized DSCR (1.20x-1.30x) at takeout rather than a going-in DSCR. Going-in coverage can be sub-1.00x as long as the business plan gets the property to coverage by stabilization.
The threshold isn’t the interesting part. The interesting part is which NOI they’re testing it against.
What counts as NOI for DSCR purposes
Here’s where seller NOI and lender NOI diverge. Pull any offering memorandum and run through the lender’s adjustments line by line. The pattern repeats on every deal.
Vacancy and credit loss: The seller’s NOI typically uses actual vacancy from the trailing 12 months, which on a fully-leased asset is close to zero. Lenders apply a market vacancy regardless — usually 5-7% for stabilized multifamily, 7-12% for office, 5-8% for industrial, and 6-10% for retail. On a 100%-leased trophy office building, the lender’s vacancy adjustment alone can wipe out 6-8% of NOI.
Management fees: Most institutional sellers self-manage at zero cost in the financials. Lenders impute a market management fee — 3-4% of effective gross income for office and industrial, 3-5% for multifamily, 4-5% for retail. The seller saved $250K. The lender’s NOI is $250K lower.
Replacement reserves: Almost no seller NOI includes them. Lenders require them: $250-400 per unit for multifamily, $0.15-0.30 per SF for industrial, $0.30-0.50 per SF for office, $0.20-0.40 per SF for retail. This isn’t a recovery of capex. It’s a non-cash expense applied for sizing purposes.
Real estate taxes: Sellers show trailing taxes. Lenders reassess at the purchase price, which on an asset that’s been held for 10+ years can mean a tax increase of 20-60%. On California Prop 13 protected basis, the reassessment alone can drop NOI 8-15%.
Free rent burn-off: If a tenant signed recently and is still in their abatement period, the lender doesn’t credit the future rent yet — they wait for the abatement to burn off. On Year-1 underwriting this can be a meaningful haircut.
Tenant credit haircuts: On office and retail, lenders apply a credit haircut to weaker tenants. A 5,000 SF lease from a non-investment-grade local restaurant counts at 70-80% of contractual rent in the lender’s analysis. The seller’s NOI counts it at 100%.
Step rent treatment: Some lenders use average rent over the lease term. Some use Year-1 contractual rent. Some use marked-to-market rent for short-dated leases. Each method produces a different NOI.
Stack these adjustments and the lender’s NOI lands 5-15% below the OM number on a typical deal. On a value-add or transitional asset, the gap can be 20-30%. This is why deals that look like they pencil on a broker’s spreadsheet fail the DSCR test at the lender’s underwriter.
Why deals fail the DSCR test
In a decade of underwriting, I’ve seen the same four failure modes repeat:
1. Inflated seller NOI. The lender normalizes, NOI drops 12%, debt service eats the cushion, the loan doesn’t size. Solution: lower price, more equity, or a different lender with looser inputs.
2. Rate movement. You modeled at LOI rates, the lender quoted at current rates 50 bps higher, debt service is 8% higher, coverage drops below threshold. Solution: forward rate-lock if available, hedging with a cap (lender will accept the strike rate), or restructured debt.
3. Lease rollover. A major tenant has 18 months left on their lease. The lender stress-tests NOI assuming the tenant doesn’t renew or renews at a lower rate. Sized DSCR falls below threshold. Solution: pre-leasing, lease extension before closing, structured holdback, or mezzanine debt to bridge the gap.
4. Reassessment shock. The asset is in a jurisdiction with full reassessment on transfer (Florida, Texas, most of the Northeast). Property taxes jump 35% post-close. Lender NOI reflects the higher tax. DSCR fails. Solution: tax appeal modeling, abatement programs if available, or higher equity.
A fifth, more recent failure: lenders applying a stressed exit cap rate when they size the loan based on a refi assumption at maturity. If today’s market cap is 5.75% but the lender uses 6.50% as their stressed exit assumption, your DSCR at refi is materially worse than your DSCR at closing, and they may size the loan to the stressed scenario rather than today’s.
DSCR for SFR investors vs. institutional CRE
The vast majority of search volume for “DSCR loan” comes from single-family rental and small-multifamily investors. These are real customers buying real properties, and the non-bank DSCR lender market — Kiavi, Visio, RCN, Lima One, CoreVest, Easy Street, A&D Mortgage, Truss Financial — exists specifically to serve them. Honest answer: these readers aren’t DDee.ai’s customer.
We built DDee.ai for institutional CRE acquisitions teams. The DSCR conversation in that world is different in three ways:
First, the test runs on a property cash flow that’s been through dozens of normalizations. A $145M industrial portfolio acquisition involves hundreds of pages of leases, schedules, rent rolls, and TI/LC obligations. Computing the NOI the lender will actually use means reconciling every lease, normalizing every expense line, and reassessing every assumption. That’s where teams spend weeks, and it’s where most of the diligence errors I’ve seen happen.
Second, the DSCR test interacts with debt yield and LTV simultaneously. Bank and life-co lenders size to the most restrictive of the three. Knowing which constraint is binding tells you where to push — on debt structure, on tenant credit, or on basis.
Third, the IC memo has to show the DSCR sensitivity table. Not just the going-in number, but DSCR at +50 bps, +100 bps, at exit cap, at stressed rollover. The output isn’t a loan approval — it’s a recommendation to a committee.
This is the version of the DSCR conversation we built around. Every lease term gets extracted with a citation back to the source PDF. Every NOI normalization is documented. Every tenant gets a credit score and a default-probability estimate. The IC-ready package shows the going-in coverage, the stressed scenarios, and the rollover-adjusted NOI side by side.
For institutional teams looking to compress the diligence-to-IC timeline on the DSCR analysis specifically: that’s the workflow we replace.