Debt Yield: The Constraint Institutional Lenders Prefer (and Why CMBS Pivoted Post-GFC)

Key Takeaways

Why Lenders Prefer Debt Yield

Ask what sets the proceeds on a quote for a low-cap-rate asset and the answer is usually debt yield, not loan-to-value or debt service coverage (the break-even math below shows why). One property sets debt yield apart from the other constraints: it does not move with debt structure. Lower the coupon by 100 basis points and the DSCR rises mechanically; switch the loan to interest-only and the DSCR rises another 15–25%; extend the amortization from 25 to 30 years and the DSCR rises again. The underlying property hasn't generated a single extra dollar of NOI in any of these moves. Debt yield, computed as NOI divided by the loan amount, doesn't move at all. The denominator is the loan dollar; the numerator is the property's unlevered cash flow; the structure of the debt is irrelevant to the ratio.

That property, the structural insensitivity to debt engineering, is what makes debt yielda core institutional discipline. A CMBS investor, a life co underwriter, a bridge lender, and a bank credit committee can all look at a debt yield number and know exactly what cash return they're earning on the loan dollar before considering principal recovery. They can't extract the same information from DSCR without first asking what coupon the loan carries, what amortization schedule applies, and whether there's an IO period, and even then, the DSCR computed off a 5-year IO period reverts to a much lower number once amortization kicks in. Debt yield is the metric you can compare across term sheets, across vintages, across asset classes, on the same basis.

This article walks the institutional case for debt yield: the formula and what it measures, the post-GFC case for debt yield in CMBS, the illustrative 2026 minimums by lender type, the break-even cap rate framing that translates debt yield into a market-cap-rate cushion, and the 2026 cycle context where compressed cap rates are pushing debt yield ahead of LTV as the binding test on premium-priced deals. The comparator to DSCR runs throughout, but the deeper DSCR mechanics (amortizing versus IO proceeds lifts, stressed-rate testing) live in theDSCR Sizing Constraintsarticle in this cluster.

The Formula and Its Definitions

The formula is short and stated cleanly:

Debt Yield = Net Operating Income ÷ Loan Amount

The output is expressed as a percentage. A property with $1.0M of NOI and a $10.0M loan has a 10.0% debt yield. The same property with a $12.5M loan has an 8.0% debt yield. The same property with a $20.0M loan has a 5.0% debt yield. The number falls as leverage rises: mechanically, monotonically, with no degree of freedom for the lender or borrower to engineer it.

The numerator is NOI on the same institutional basis as DSCR: gross revenue less vacancy and credit loss less normalized operating expenses, with capital reserves typically held below the line. The convention is in-place or underwritten NOI for permanent lenders (CMBS, life co, bank, agency) and stabilized-at-conversion NOI for transitional lenders (bridge, private debt). The lender does not size debt yield against the borrower's stabilized projection on a permanent loan; the lender rebuilds NOI on its own basis at quote stage.

The denominator is loan amount, and the convention here is the total loan amount, the face value of the note, not net proceeds after holdbacks, reserves, or financing costs. CMBS sizing in particular runs debt yield on the loan amount as disbursed to the borrower including any future-funded portion (capex reserves, leasing reserves, interest reserves) committed by the lender. Bridge debt sizing is consistent: the lender's debt yield is computed on the maximum commitment, not the initial advance. The institutional reader should always verify which loan-amount convention the term sheet uses; broker-pitched debt yield numbers occasionally compute on initial advance to flatter the ratio.

Debt yield is a lender's metric in the same way DSCR is. The borrower works backward from the lender's minimum: if the lender posts an 8.0% minimum and the property's underwritten NOI is $5.0M, the maximum loan at the debt yield constraint is $5.0M ÷ 8.0% = $62.5M. The borrower's sizing question is not "what's my debt yield?" but "what's the loan amount at the lender's minimum, and which of the four constraints binds tightest on this deal?"

The Post-GFC Pivot to Debt Yield

Legacy CMBS conduits were sized on LTV and DSCR, and 44 of the 50 conduit deals with the largest losses are 2006–2007 vintages. Trepp'sJune 2019 study of conduit loss severitiesfound that the 50 deals with the greatest losses had suffered $23.0 billion of write-downs, 42 percent of all conduit write-downs, and carried a weighted-average LTV of 70.5%, against 65% for the 50 deals with the smallest losses. A representative bad deal, GSMS 2007-GG10, was securitized at a 73.6% LTV, a 1.30x DSCR and a 7.96% weighted-average debt yield, and has realized cumulative losses of 16.34%. A pool-level debt yield below 8% leaves little income cushion under the loan when values fall.

The lesson is that LTV and DSCR can both move without any change in the property's income. LTV moves with the appraisal: the appraiser uses comp set transactions that already embed the prior cycle's exuberance, so value rises with the market and falls with it. DSCR moves with structure: lower the coupon by sliding down the curve, extend the amortization, switch to IO. As an illustration, a loan sized to a 1.25x DSCR on an interest-only basis at a 5.5% coupon would show about 1.01x on a 30-year amortizing schedule at the same coupon (a mortgage constant of 6.81% versus 5.50%). The IO option is the proceeds-maximizing election.

Debt yield is the constraint that does not depend on either. The numerator is the property's NOI, the same NOI the lender's credit committee normalizes from rent roll and T-12. The denominator is the loan amount, a number the lender controls directly. There is no appraisal between them, no coupon election, no amortization-schedule degree of freedom. It is not immune to judgment: NOI normalization and the loan-amount convention (initial advance versus total commitment) still move the ratio, which is why the formula section defines both carefully. Trepp's data shows the pattern lenders care about: conduit pools securitized at debt yields of 10–12% averaged 4.03% losses, pools at 8–10% averaged 8.07%, and pools at 6–8% averaged 10.26%. Trepp also notes that more recent CMBS 2.0/3.0 loans generally have much higher recovery rates because of more conservative underwriting and tighter transparency and disclosure requirements.

A higher debt yield floor bites hardest at refinancing. As an illustration, a loan sized at a 7.5% debt yield on $5.0M of NOI is $66.7M. If NOI grows 15% to $5.75M, a refinance at an 8.0% minimum sizes to $71.9M (1.08x the original loan) and one at a 9.0% minimum sizes to $63.9M (0.96x). The new loan comes in below the original only when the new minimum exceeds (1 + NOI growth) × the original debt yield, here 1.15 × 7.5%, or about 8.6%.

Institutional Minimums by Lender Type

Illustrative 2026 debt yield ranges vary by lender type, by stabilized versus value-add posture, by asset class, and by market. The matrix below is a practitioner baseline set out as illustrative assumptions, not a survey of current term sheets. As with DSCR minimums, these are floors; the practical binding number tiers up for weaker property types (office), weaker markets (tertiary), weaker sponsors, and higher LTV.

Illustrative debt yield ranges by lender type, 2026 (practitioner assumptions, not a survey). Bridge is shown on its binding basis, stabilized NOI at conversion; its acquisition debt yield is far lower. The agency range is implied from LTV and cap rate assumptions, not a published minimum.

A few lender-specific calibrations behind the matrix:

• CMBS conduit (8.0–10.0%). Debt yield is a lead constraint in CMBS conduit underwriting. It binds first whenever the market cap rate sits below the break-even described in the next section (debt yield × LTV cap). The pool-level weighted-average debt yield is a standard credit statistic for conduit deals; Trepp reports it alongside weighted-average LTV and DSCR for each deal. Office sits above the illustrative range, at 11.0%+ on stabilized assets. Multifamily and industrial sit at the floor of the range; retail varies meaningfully with location and anchor profile.

• Commercial banks (9.0–11.0%). Banks treat debt yield as a covenant ratio more than a sizing constraint; DSCR with a stressed-rate overlay is often the tighter test on bank deals. Where debt yield matters most in bank lending is the loan covenant tier: some bank loans include a debt-yield covenant tested periodically, and what a breach triggers (cash management, a cure period, or a default) depends on the loan documents.

• Life insurance companies (8.5–10.5%). Life co debt yield discipline is conservative, alongside stressed-DSCR. Life co underwriting can test debt yield, DSCR at the headline rate, and DSCR at a stressed rate simultaneously, taking the binding proceeds at the minimum of the three. A life co quote is often the lowest-proceeds permanent option on a given deal, in exchange for the lowest spread.

• Agency multifamily (implied 6.9–8.7%). Fannie Mae and Freddie Mac multifamily loans are sized on DSCR and LTV tiers, so debt yield is an implied output rather than a stated minimum. As an illustration, an 80% LTV at a 5.5% cap rate implies a 6.9% debt yield (5.5% ÷ 80%), and a 75% LTV at a 6.5% cap rate implies 8.7% (6.5% ÷ 75%). Whether DSCR, LTV or the implied debt yield binds depends on the asset's NOI yield versus its cap rate.

• Bridge / private debt (9.0–11.0% stabilized). Bridge lenders run debt yield on three bases: in-place NOI at acquisition, underwritten stabilized NOI at conversion, and a permanent take-out test (detailed in the bridge section below). The acquisition figure is structurally low because the property is in transition: lease-up, repositioning, capex deployment, often 4.0–6.5% on heavy-transitional deals. The lender accepts the thin acquisition debt yield in exchange for the stabilized projection, typically 9.0–11.0% at conversion, and a higher coupon to compensate for the structural risk. Coverage (DSCR) is often below 1.0x at acquisition on bridge debt, which makes debt yield on stabilized NOI the sizing constraint that informs the most until stabilization. Proceeds are the minimum across the debt yield, LTV or LTC, and take-out tests.

Break-Even Cap Rate Logic

The single most important institutional translation of debt yield is the break-even cap rate framing. The math is short, but the implication is the institutional reader's framework for reading a term sheet at a glance. Start from the definitions:

Debt Yield = NOI ÷ Loan
Cap Rate = NOI ÷ Value
Loan = LTV × Value

Substitute the third equation into the first:

Debt Yield = NOI ÷ (LTV × Value) = (NOI ÷ Value) ÷ LTV = Cap Rate ÷ LTV

Rearrange:

Break-Even Cap Rate = Debt Yield × LTV

This is the framework. If a lender quotes an 8.0% debt yield minimum and caps LTV at 65%, the implied market-cap-rate floor below which debt yield (rather than LTV) binds is 8.0% × 65% = 5.2%. Above a 5.2% cap rate, LTV is the binding constraint and debt yield is loose. Below a 5.2% cap rate, the lender's proceeds-at-LTV implies a debt yield below 8.0%, so debt yield kicks in and overrides LTV. The same logic runs for every combination of debt yield minimum and LTV cap.

The institutional reader uses this framework to sanity-check term sheets without running the full model. A life co quote at a 9.5% debt yield minimum and a 60% LTV cap implies a 5.7% cap rate break-even; below that, debt yield governs. A CMBS conduit quote at an 8.5% debt yield and a 70% LTV cap implies a 5.95% cap rate break-even; below that, debt yield governs. On bridge quotes sized to cost, the same product (debt yield × LTC cap) gives a yield-on-cost break-even, which equals a market cap rate break-even only when cost equals value. Knowing the break-even cap rate before opening the model tells the practitioner whether debt yield is in play.

A worked example. A trophy stabilized multifamily asset in a gateway market (Manhattan, San Francisco, Boston) underwrites at a 4.5% market cap rate on $5.0M of NOI. Implied value = $5.0M ÷ 4.5% = $111.1M. The borrower's pro forma at 65% LTV implies a $72.2M loan. Test the lender's debt yield: $5.0M ÷ $72.2M = 6.9%. The CMBS quote at 8.0% debt yield minimum sizes to $5.0M ÷ 8.0% = $62.5M, a $9.7M proceeds gap versus the borrower's pro forma. Debt yield is the binding constraint; LTV is loose. In an institutional debt broker's mental model, this read happens in seconds: "4.5% cap rate against an 8.0% debt yield means the break-even is 5.2%; the deal is 70 bps inside the threshold, so debt yield governs."

Why Debt Yield Ignores Debt Structure

Walk the comparison directly. Take a single stabilized multifamily property: $5.0M of NOI, current market cap rate 5.5%, implied value $90.9M. Three lender quotes, all sized to the same loan amount of $50.0M for comparability:

Three lender quotes, identical property and loan amount, manipulating coupon and amortization. The DSCR ranges from 1.23x to 1.82x, a 48% spread driven entirely by debt structure, not by property fundamentals. Debt yield is 10.0% in all three cases. The metric is structurally insensitive to debt engineering, which is precisely why institutional lenders prefer it.

Three different DSCR values: 1.23x, 1.47x, 1.82x. The property hasn't changed. The NOI hasn't changed. The loan amount hasn't changed. The only thing that's moved is the structure of the debt: coupon, amortization schedule, IO toggle. DSCR moves with all three. Debt yield moves with none of them: it's pinned at NOI ($5.0M) divided by loan ($50.0M) = 10.0% in every case.

This is the punchline. A lender who sizes against DSCR can be tactically pushed into more proceeds by sliding down the curve, extending the amortization, or electing IO. A lender who sizes against debt yield cannot. The same property at the same NOI sizes to the same loan balance at the same debt yield minimum regardless of which coupon environment, which amortization schedule, or which IO posture the borrower negotiates. Debt yield is the constraint that pins the lender's underlying risk position to the property's actual cash-generation capacity, independent of the financial engineering applied on top.

The institutional consequence: a deal that clears a 1.30x DSCR on a 5-year IO at a 5.5% coupon carries a 7.15% debt yield (1.30 × 5.5%), below most of the illustrative minimums above. The DSCR looks "good" relative to the headline; the debt yield is the discipline the lender actually cares about. When the IO period rolls off and the loan begins amortizing, the DSCR collapses (because debt service jumps), but the debt yield doesn't move. The lender's risk position hasn't changed; only the headline coverage ratio has.

Bridge Lending: Why DY Dominates

Bridge lending is the segment where debt yield's dominance over DSCR is most absolute. The reason is structural: bridge loans are sized against transitional properties: under-leased, mid-renovation, mid-lease-up, mid-stabilization. The in-place NOI at acquisition is often a fraction of the stabilized projection, sometimes near zero in the case of heavy-renovation deals. DSCR at acquisition is frequently below 1.0x and meaningless as a sizing test. A 0.6x DSCR on an acquisition basis tells the lender nothing about whether the deal will stabilize at 1.25x in year three; what tells the lender something is the debt yield on underwritten stabilized NOI at the maximum commitment.

Bridge debt yield mechanics follow a different convention from permanent debt yield. The lender computes (illustrative ranges):

• Acquisition debt yield. In-place NOI ÷ initial advance, computed at acquisition. Often 4.0–6.5% on heavy-transitional deals. Disclosed but not the binding test.

• Stabilized debt yield at conversion. Underwritten stabilized NOI (year three or year four depending on business plan) ÷ maximum commitment. Typically 9.0–11.0% on the lender's underwritten stabilized NOI. This is the binding test: the lender will not size the maximum commitment above the level that delivers the stabilized debt yield minimum.

• Refinance debt yield at conversion test. The bridge lender often requires the property's underwritten NOI at the conversion date to support a refinance to permanent at a permanent-market debt yield minimum (typically 8.0–9.0%). This is the "permanent take-out" test and it constrains the bridge sizing from the back end: if the property can't refinance into a permanent loan at conversion, the bridge lender won't extend or convert, and the borrower faces a maturity event.

The institutional bridge lender thinks about debt yield in three time slices: where it is today (acquisition), where it lands at stabilization (conversion), and whether the property's stabilized cash flow will support a permanent take-out (refinance). The acquisition number is mostly disclosure; the stabilized and refinance numbers are the binding constraints, alongside the LTV or LTC cap on the initial advance and total commitment. DSCR plays almost no role in this stack; it's a derived consequence of the coupon, the amortization, and the debt yield, not a primary sizing test.

This is the structural reason the AQ-141 opportunistic-with-bridge modelin the Apers Marketplace runs all three debt yield slices on its bridge debt tab and references DSCR only as a consequence of the sizing rather than as a separate constraint. The institutional discipline that the model enforces is the discipline bridge lenders actually use: debt yield governs, DSCR follows.

2026 Cycle: When DY Governs vs When LTV Does

Whether debt yield or LTV governs depends on the cap-rate-versus-debt-yield spread: in compressed-cap-rate markets, the lender's break-even cap rate (debt yield × LTV) sits above the market cap rate, so the lender's max-LTV proceeds imply a debt yield below the minimum, and debt yield becomes the binding constraint. In wider-cap-rate markets, LTV (or DSCR) binds first.

Work two parallel examples on the same lender quote: a CMBS conduit minimum of 8.5% debt yield and 65% LTV. Asset A is a trophy gateway-market multifamily property at a 4.5% market cap rate. Asset B is a sun-belt secondary-market multifamily property at a 6.0% market cap rate. Same lender, same minimums.

Same lender, same minimums (8.5% debt yield, 65% LTV) applied to two assets at different cap rates. On the trophy gateway deal (cap rate inside the 5.5% break-even), debt yield is the binding constraint. On the sun-belt secondary deal (cap rate outside the 5.5% break-even), LTV is the binding constraint. The break-even cap rate framework predicts which constraint binds without running the model.

The institutional reader should walk this analysis on every term sheet. For assets at compressed cap rates, such as primary-market multifamily and industrial, debt yield is the more likely binding constraint on a CMBS conduit quote. For secondary-market deals and for asset classes with wider cap rates (retail, hotels, office), LTV tends to bind first. The actual answer is asset-by-asset, lender-by-lender, but the framework for getting there is the break-even cap rate.

The 2026 cycle wrinkle that matters most is the refinance gap on maturing loans originated at lower debt yields. The illustration in the post-GFC section shows the arithmetic: the new loan equals the original loan times (1 + NOI growth) times the original debt yield divided by the new minimum. With 15% NOI growth, a 7.5% original debt yield and an 8.0% new minimum, the loan grows 8%; at a 9.0% new minimum, it shrinks 4%. If coupons have also risen since origination, DSCR may bind before debt yield does, and any shortfall is funded with new equity, an extension, a discounted payoff, or a workout. This dynamic is covered in detail inRefinancing Risk: Maturity Wave and Default Rate Sensitivity; the relevant point here is that the proceeds gap depends on which of debt yield, DSCR and LTV binds on the new loan.

How to Model It

Modeling debt yield correctly is procedurally short but conceptually adjacent to two other things: the four-constraint sizing cross-check (DSCR / LTV / debt yield / LTC) and the term-sheet comparison workflow (CMBS versus life co versus bank versus bridge). A useful debt yield model does five things:

• 1. Normalize NOI. Rebuild from rent roll and T-12. Strip one-time items. Benchmark OpEx to market expense ratios for the asset class. The lender's underwritten NOI is the relevant numerator, not the broker's pro forma, not the borrower's stabilized projection on a permanent loan, not the in-place plus rent-bumps figure that bridges into year-two.

• 2. Compute debt yield at proposed loan amount. Simple: NOI ÷ loan. Use total loan commitment, not net proceeds. The output number anchors the term-sheet read.

• 3. Solve for max loan at the lender's debt yield minimum. The other direction: NOI ÷ debt yield minimum = maximum loan at the constraint. Compare to the borrower's requested loan amount and to the proceeds at the other three constraints (DSCR, LTV, LTC).

• 4. Compute the break-even cap rate. Debt yield × LTV cap = break-even cap rate. Compare to the market cap rate on the asset. If market cap is inside (below) the break-even, debt yield binds; if outside (above), LTV binds. This is the diagnostic that tells the practitioner which constraint will govern before running the full sizing.

• 5. Compare term sheets on the same basis. Across CMBS, life co, bank, agency, and bridge: each lender's debt yield minimum, the max loan at that minimum, the implied break-even cap rate, the binding-constraint identification. The institutional debt-broker output is a single sheet comparing all lender quotes side by side on the same NOI basis.

The Apers Marketplace pocket models implement this workflow on a single tab. AQ-110 (multifamily acquisition)runs the four-constraint sizing with debt yield as a first-class output alongside DSCR, LTV, and LTC, and presents the break-even cap rate inline. AQ-141 (opportunistic with bridge)extends the pattern to the bridge-specific stabilized-and-refinance-debt-yield logic, with conversion testing against permanent-market minimums. AQ-301 (anchored retail)runs the asset-class-tier overlay for retail debt yield minimums, which run higher than multifamily and where anchor-tenant credit quality affects the lender's discipline materially.

Common Mistakes

Seven errors that show up routinely in IC memos and term-sheet analysis on debt yield:

• Treating debt yield as a borrower's metric. Debt yield is the lender's cash-on-cash return on the loan dollar. The borrower's question is the inverse: at the lender's minimum, what's the maximum loan amount, and which of the four constraints is tightest? Sponsors who pitch "our debt yield is 9.2%" are quoting the lender's number, not their own; the relevant borrower metric is leveraged cash-on-cash return on equity, not debt yield on debt.

• Using net loan proceeds instead of total loan amount in the denominator. Total loan amount includes future-funded portions (capex reserves, interest reserves, leasing reserves) and any holdbacks the lender has committed to. Computing debt yield on net proceeds flatters the ratio: a 5% holdback on a 9.0% debt yield, for example, lifts it by about 47 bps (9.0% ÷ 0.95 = 9.47%), and produces a number a lender's credit committee will restate on the committed amount.

• Using stabilized NOI on a permanent loan. Permanent lenders (CMBS, life co, bank, agency) size debt yield off in-place or normalized underwritten NOI, never the borrower's year-three stabilized projection. Bridge lenders size off stabilized NOI at conversion, but the conversion test runs against its own stabilized debt yield minimum (9.0–11.0% in the illustrative range), and the conversion is contingent on the borrower actually executing the business plan. Sponsors who model permanent debt yield off stabilized NOI structurally overstate proceeds.

• Comparing debt yield across asset classes without the tier adjustment. A 9.0% debt yield on a Class A primary-market multifamily property and a 9.0% debt yield on a single-tenant office property in a secondary market are not the same number. Asset-class tiering matters: in the illustrative ranges above, office sits at 11%+ because of tenant credit and demand uncertainty, retail at 9.0–10.5% with anchor-tenant sensitivity, and multifamily and industrial at 8.0–9.5%. A debt yield comparison across asset classes without normalizing for the lender's tier-by-asset-class minimum is structurally misleading.

• Ignoring the break-even cap rate. The single most useful diagnostic for which constraint binds (debt yield vs LTV) is the break-even cap rate (debt yield × LTV). Sponsors who jump into full four-constraint sizing without first computing the break-even cap rate spend iteration time on the model that a 10-second back-of-the-envelope could have saved. The break-even is the institutional shortcut.

• Confusing debt yield with cost of debt. Debt yield is NOI ÷ loan, the property's unlevered yield as a percentage of the loan dollar. Cost of debt is the loan's all-in coupon, the rate the borrower pays. These are unrelated numbers. A loan at a 5.5% coupon on a property with a 10.0% debt yield is structurally different from a loan at a 5.5% coupon on a property with a 7.0% debt yield, even though the cost of debt is identical. Debt yield measures the asset's cash-generation capacity relative to the loan; cost of debt measures the rate the borrower pays. Conflating the two produces analysis that doesn't survive an institutional credit committee.

• Quoting "debt yield" without specifying lender type and asset class. A 9.0% debt yield means very different things at a CMBS conduit (probably comfortable on multifamily, fails on office), a life company (likely the headline floor), a regional bank (probably with a covenant attached), an agency multifamily deal (comfortable against the implied range), and a bridge lender at acquisition (extraordinarily strong, suggests the property is barely transitional). Always qualify the debt yield you're quoting with the lender category, the asset class, and the in-place-versus-stabilized basis.

Related Articles

Debt yield is one of four sizing constraints in the institutional debt-analysis cluster. Sibling deep-dives:

• DSCR Sizing Constraints, Amortizing vs IO: the comparator. DSCR moves with coupon, amortization, and IO; debt yield doesn't. The institutional minimums and the stressed-rate overlay.
• LTV vs LTC: When Each Governs: the value and cost-basis constraints. The break-even cap rate framework ties debt yield to LTV; LTC binds on construction and value-add.
• Interest Rate Risk: Caps, Swaps, and Hedging: the rate-volatility response that affects debt yield indirectly through the stressed-rate overlay on the DSCR side.
• Prepayment Analysis: Defeasance and Yield Maintenance: CMBS prepayment economics.
• Refinancing Risk: Maturity Wave and Default Rate Sensitivity: the 2026 context. How refinancings size under higher debt yield minimums and higher coupons.
• Cap Rate Calculator and Formula: cap rate compression below the debt yield break-even threshold is the condition under which debt yield governs. The cross-cluster connection.

FAQ

Sources

• Trepp, Credit Matters: CMBS Conduit Loss Severities (June 2019).