What Is DSCR? The Debt Service Coverage Ratio Explained for Commercial Lenders

The debt service coverage ratio is the cash a borrower generates in a period divided by the debt service falling due in the same period. A DSCR of 1.25x means 25% more cash than the loan demands. What it is really measuring is not repayment capacity — it is margin of safety: how much of the borrower's cash flow is already spoken for, and how far things can go wrong before they are not.

That second reading is the one that changes a decision. This page is about what the ratio buys, where it goes blind, and when to override it. YuSight's Financial Spreading module builds the ratio from the source documents at 95.2% extraction accuracy against a manual benchmark, so the argument in committee is about the judgement, not the data entry.

Key facts

• DSCR is the most common single reason a file gets escalated. In the FDIC's Small Business Lending Survey, 67% of banks said an insufficient debt service coverage ratio increases the number of approval levels a loan must pass through — more than any other credit factor in that figure (FDIC, 2022 Small Business Lending Survey, Section 3). The survey drew responses from approximately 1,300 of 2,000 sampled banks, a 68% response rate (FDIC, Executive Summary).
• No regulator prescribes a formula or a floor for general commercial lending. The OCC's Rating Credit Risk booklet defines the underlying idea — "Business cash flow is the operating revenue derived from ordinary business activities less operating costs paid (not simply incurred), plus noncash expenses such as depreciation" — and stops there (Comptroller's Handbook, *Rating Credit Risk*). Every numeric threshold in this article is lender policy.
• Coverage is the early-warning indicator, by name. The same booklet: "Shortfalls in cash flow or debt service coverage are usually the most obvious indications of a problem credit."
• A 1.25x DSCR on a business with 35% contribution margin survives a 5.4% revenue decline. That is the whole cushion, and it is worked line by line below.
• DSCR is silent on the thing currently worrying supervisors. The OCC's Semiannual Risk Perspective, Spring 2026 notes that "a substantial volume of CRE loans originated in a lower-interest environment will mature in the next several years and will need to be refinanced at prevailing rates" (OCC). A perfectly healthy amortising DSCR says nothing about that.

What is DSCR actually testing?

Written out, it is unremarkable:

DSCR = Cash available for debt service in a period ÷ Debt service falling due in that period

Three things follow from that sentence, and they are the whole ratio.

It is a period measure, not a solvency measure. DSCR compares a flow to a flow. It has no opinion on whether the borrower's assets exceed its liabilities, whether it could survive a year with no sales, or whether it will exist in five years. A business can be balance-sheet insolvent and post a 1.8x DSCR; a business can be cash-rich and post 0.9x.

It is a claim on cash the lender is competing for. Every unit of the numerator is also being claimed by suppliers, tax authorities, employees, capital expenditure and shareholders. A DSCR of 1.25x says the lender's claim is covered and that the residual after the lender's claim is 20% of the cash flow. That residual is what pays for everything else that goes wrong.

It is a ratio the lender partly authors. The denominator is a function of the structure the lender chose — tenor, amortisation profile, interest rate, whether there is a balloon. Stretch the amortisation and the DSCR improves without the borrower doing anything. This is the single most under-appreciated fact about the ratio, and it is quantified below.

The general form is settled. The inputs are not: the numerator can be EBITDA, cash flow after tax and capex, net operating income, post-tax cash profit, or a group figure including a guarantor's household. The full catalogue of definitions, with one borrower run through ten of them, is in the DSCR formula and every variant lenders use. This page assumes you have picked one and asks what it tells you.

Why is the numerator the contested part?

The denominator is close to arithmetic. Interest is on the loan schedule. Scheduled principal is on the loan schedule. The only genuine questions are whether accrued or cash interest is used, whether a balloon inside the test window belongs in it, and whether a revolver's fluctuating balance counts. Those are answerable from documents.

The numerator is a judgement about which cash is real and which is available, and every one of those judgements moves the answer:

• Is it accrual or cash? Reported EBITDA includes revenue invoiced and not collected. A borrower with 95 days of receivables reports EBITDA it has not received.
• Is it recurring? A one-off insurance recovery, a government grant, a gain on disposal, a contract termination fee. Each one is real cash and none of it will happen again.
• Is it available? Cash inside a subsidiary in a jurisdiction with exchange controls, or inside a minority-owned affiliate, counts on the consolidation and cannot be used to pay the loan.
• Is it discretionary? Owner compensation, distributions, discretionary bonuses and deferred maintenance are all cash the borrower could stop spending. Adding them back assumes the borrower will, and that assumption belongs in the memo as a condition, not in the numerator as a fact.
• Is it before or after the money the business needs to stay in business? Maintenance capex, cash tax, and the working capital a growing business absorbs.

Two competent analysts starting from the same audited accounts routinely produce numerators 20–30% apart on the strength of those five questions alone. That is why the discipline is to fix the definition first, in writing, and to show the derivation from the filed statement to the numerator on the face of the spread rather than in a footnote.

A worked DSCR, line by line

Meridian Fabrication is a structural steel fabricator, owner-managed, financing a plant upgrade. Year ended 31 December 2025. Figures in thousands of currency units — the arithmetic works in any currency.

Debt service and the ratio:

Debt service = Term loan interest 1,320 + Revolver interest 280 + Scheduled principal 2,000 = 3,600 DSCR = EBITDA 4,500 ÷ 3,600 = 1.25x

A clean 1.25x. Most credit policies would call that compliant and move on. The rest of this page is about what that number does and does not entitle you to conclude.

What does a 1.25x DSCR actually buy you?

Start with the algebra, because it is simple and almost never stated in a memo. Coverage breaks — reaches 1.00x — when cash flow has fallen to equal debt service. So:

Maximum fall in cash flow before DSCR = 1.00x = 1 − (1 ÷ DSCR) At 1.25x: 1 − (1 ÷ 1.25) = 1 − 0.80 = 20.0%

Meridian's EBITDA can fall from 4,500 to 3,600 — a fall of 900, or exactly 20% — before it stops covering. That is the cushion in EBITDA terms, and it is the same for every borrower at 1.25x.

Now translate it into the language the credit committee actually argues in, which is revenue. Meridian's contribution margin is 35%, so every unit of lost revenue costs 0.35 units of EBITDA, and the fixed cost base of 12,300 does not move:

Revenue fall that destroys 900 of EBITDA = 900 ÷ 0.35 = 2,571 As a percentage of revenue = 2,571 ÷ 48,000 = 5.4%

A 1.25x DSCR at Meridian survives a 5.4% revenue decline. Not a 20% decline. Not a recession. One lost customer in a book where the top account is 12% of sales, and this borrower is at 1.00x.

Run the same arithmetic across coverage levels, holding revenue at 48,000 and contribution margin at 35%:

The gap between the third and fourth columns is operating leverage, and it means the same DSCR is not the same protection in two different businesses. Take a distributor with the same revenue and the same 1.25x, but a contribution margin of 12% and a correspondingly smaller fixed cost base:

Revenue fall that destroys 900 of EBITDA = 900 ÷ (0.12 × 48,000 ÷ 48,000) = 900 ÷ 0.12 = 7,500 = 15.6% of revenue

The distributor's 1.25x tolerates a 15.6% revenue decline. The fabricator's identical 1.25x tolerates 5.4%. A single portfolio-wide DSCR floor treats them as equivalent credits. They are not, and the difference is roughly threefold.

The practical instruction: never quote a DSCR in a memo without the sentence that follows it. "DSCR 1.25x" is a number. "DSCR 1.25x, which is a 5.4% revenue decline given a 35% contribution margin and a fixed cost base of 12,300" is a credit opinion.

How sensitive is DSCR to things the borrower does not control?

Two inputs sit in the denominator and belong to the lender and the market, not the borrower.

Interest rates. Meridian has 19,200 of debt at a blended 8.33% (1,600 ÷ 19,200). Holding EBITDA constant:

Coverage breaks at +469 basis points on unhedged floating debt. If the facility is floating and unhedged, that number belongs in the credit memo next to the DSCR, and the hedging condition belongs in the term sheet.

Amortisation. The lender wrote this input:

Same borrower, same debt, same EBITDA. DSCR from 0.94x to 1.53x. A credit committee that declines at 8 years and approves at 12 has not found a better borrower; it has found a longer runway and a bigger balance outstanding when the market turns. If your structuring team can hit any DSCR the policy demands by moving the tenor, the DSCR floor is not a risk control — it is a drafting exercise. The counterweight is a tenor rule tied to asset life, and a hard look at the balance outstanding at maturity.

How is a DSCR at underwriting different from a DSCR as a covenant?

They share a name and almost nothing else. Treating them as one number is the most common structural error in a credit file.

Two consequences matter.

The covenant number is almost always the more generous of the two, because it is computed from a negotiated definition rather than an analyst's view. A borrower can pass a 1.25x covenant while the lender's own analytical DSCR is below 1.10x, and nothing in the loan documentation will flag it. Both numbers belong in the memo, with the gap decomposed. The drafting mechanics — add-back caps, cash netting, accounting-basis freezes, whether a balloon is debt service — are set out in covenant testing for DSCR and leverage, and the operational side in covenant monitoring in commercial lending.

A covenant DSCR only exists if someone tests it. A ratio computed once at sanction and never recomputed is not a control. That is a workflow question rather than a formula question, and it is where most of the value in the ratio is lost.

What DSCR does not tell you

Seven blind spots. Each one has produced losses on files where the coverage ratio was fine.

1. It says nothing about the balance sheet

Meridian carries 19,200 of debt on 4,500 of EBITDA — 4.27x leverage at a 1.25x DSCR. Now consider a second borrower with the same 4,500 EBITDA, 30,000 of debt at the same blended 8.33%, amortising over 30 years:

Interest = 30,000 × 8.33% = 2,500 Principal = 30,000 ÷ 30 = 1,000 Debt service = 3,500 DSCR = 4,500 ÷ 3,500 = 1.29x Leverage = 30,000 ÷ 4,500 = 6.67x

The second borrower has 56% more debt and a better DSCR. Coverage and leverage are different questions and a policy that covenants only the first is covenanting half the credit. The full ratio set that surrounds DSCR is in the 24 credit analysis ratios that drive a lending decision.

2. It ignores working capital, and growth is where that bites

An EBITDA-based DSCR treats a growing receivables book as free. It is not. Suppose Meridian grows revenue 25% to 60,000 next year, holding 75 days of receivables, 60 days of inventory on direct cost and 45 days of payables, with EBITDA rising to 5,400:

Receivables 48,000 × 75/365 = 9,863 → 60,000 × 75/365 = 12,329 +2,466 Inventory 31,200 × 60/365 = 5,129 → 39,000 × 60/365 = 6,411 +1,282 Payables 31,200 × 45/365 = 3,847 → 39,000 × 45/365 = 4,808 (961) Net working capital absorbed 2,787 Reported DSCR = 5,400 ÷ 3,600 = 1.50x Cash after working capital = 5,400 − 2,787 = 2,613 Actual cash coverage = 2,613 ÷ 3,600 = 0.73x

The reported DSCR improves from 1.25x to 1.50x in the year the borrower runs out of money. Fast-growing, working-capital-hungry borrowers are the population where an EBITDA DSCR is most flattering and most dangerous. The fix is to show a cash-flow-based numerator beside the EBITDA one, and to read the movement in the working capital cycle directly — see bank statement analysis for lenders.

3. It is blind to refinancing risk

A DSCR built on scheduled amortisation says nothing about the balance outstanding at maturity. Meridian's 8-year profile on an 8-year loan repays the debt; a 5-year facility on a 20-year profile leaves 12,000 outstanding at maturity, and the borrower's 1.25x DSCR tells you precisely nothing about whether anyone will refinance it. The OCC's Spring 2026 Semiannual Risk Perspective frames exactly this exposure across the CRE book. Test it separately: model the balloon at a stressed take-out rate and see what DSCR the refinanced structure would need.

4. It cannot see concentration or quality of cash flow

Two borrowers at 1.25x: one with 400 customers and no account over 3% of revenue, one with a single anchor customer at 45% on a contract with 90 days' notice. Identical ratio, incomparable credits. DSCR is an arithmetic mean over a portfolio of revenue it does not inspect.

5. An annual DSCR hides the trough inside the year

A seasonal business at 1.25x annually can be at 0.6x for four consecutive months. If debt service falls monthly and cash arrives in two quarters, the annual ratio is a description of the year and not of the borrower's liquidity in March. For seasonal borrowers, run coverage monthly and size the revolver against the trough.

6. It can be manufactured by not spending

Deferring maintenance capex raises EBITDA and therefore DSCR, in the year the deferral happens. So does letting the fleet age, cutting the training budget and stretching payables. A DSCR trending upward while capex trends toward zero is a warning, not a comfort. Compare capex to depreciation year on year; a business spending materially below depreciation for three consecutive years is consuming the asset base to service the loan.

7. It is measured looking backwards

The DSCR in a monitoring pack describes a period that ended weeks or months ago, computed from statements prepared afterwards. In the OCC's own phrasing, a coverage shortfall is "usually the most obvious indication of a problem credit" — obvious, and late. The early indicators sit in the bank statements and the bureau file, not in the annual accounts: see bureau versus bank statement reconciliation and repayment track record analysis.

When should a lender override the DSCR?

Overriding a ratio is not the same as ignoring one. An override is a documented decision that the ratio is measuring the wrong thing for this borrower, with a stated reason and a compensating control. It runs in both directions.

Decline, or restructure, despite a passing DSCR, when:

• The numerator depends on add-backs that are discretionary spend the borrower has not agreed to stop. Convert them into a documented condition or take them out.
• Coverage is achieved only through an amortisation profile longer than the economic life of the asset being financed.
• The passing ratio is a single strong year in a volatile series. Test the worst of the last five, not the last one.
• Working capital absorption at the projected growth rate exceeds the coverage cushion, as in the 0.73x example above.
• Debt is floating and unhedged and the break-even rate move is inside plausible market range.
• The cash flow is concentrated in one contract, one customer or one geography with a shorter life than the loan.

Approve, with conditions, despite a failing DSCR, when:

• The shortfall is timing, not capacity — a construction or ramp-up period where a moratorium and a properly sized interest reserve carry the gap. Size the reserve to the modelled shortfall plus a margin, and covenant coverage from the first full operating year rather than from drawdown.
• The failing year is one identifiable non-recurring event that is documented, quantified and demonstrably behind the borrower.
• A guarantor's household or an affiliate's cash flow genuinely covers the gap and can be counted without double-counting — the method is in global cash flow analysis in commercial lending, and the SBA-specific mechanics in how global DSCR is calculated for SBA 7(a) loans.
• The structure removes the shortfall: a cash sweep, a debt service reserve account, a shorter tenor with a smaller balance, or additional equity ahead of drawdown.

What is never an override is "the relationship is strong" or "the collateral covers it". On the second, the OCC is explicit: "Almost all credit transactions are expected to have secondary or even tertiary sources of repayment (collateral, guarantor support, third-party refinancing, etc.)," but "the rating assessment, until default has occurred or is highly probable, is generally based on the expected strength of the primary repayment source." Collateral changes loss given default. It does not change coverage.

How should DSCR appear in the credit memo?

Four lines, in this order, and no fewer:

1. The ratio, with its definition stated. "DSCR 1.25x, computed as EBITDA before exceptional items divided by cash interest plus scheduled principal." Not "DSCR 1.25x".
2. The covenant ratio on the agreement's own definition, where a facility exists, and the gap to line 1 explained.
3. The cushion, in the borrower's own operating terms. "Equivalent to a 5.4% revenue decline at the current contribution margin."
4. The two or three sensitivities that could plausibly break it — rate move, loss of the largest customer, working capital absorption at plan growth — with the resulting ratio.

That is four sentences and it is the difference between a memo that records a ratio and one that makes an argument. The structure it belongs in is set out in what a credit assessment memo must contain, and getting from a filed PDF to a defensible numerator is a spreading problem before it is a ratio problem — see what financial spreading is and how it works.

FAQ

How do you calculate the debt service coverage ratio?

Divide the cash the borrower generated in a period by the debt service that fell due in the same period. Our worked borrower has EBITDA of 4,500 and debt service of 3,600 — interest of 1,600 plus scheduled principal of 2,000 — so the DSCR is 1.25x.

What is a good DSCR ratio for a business loan?

Most commercial lenders look for 1.20x to 1.35x on an operating-cash-flow basis, but the honest answer is that it depends on how volatile the borrower's cash flow is. A 1.25x at a high-fixed-cost manufacturer is thinner protection than a 1.15x at a stable distributor, and the ratio alone will not tell you that.

What is the minimum DSCR lenders require?

There is no regulatory minimum for general commercial lending. Common policy floors sit around 1.15x to 1.25x, with project and infrastructure structures often covenanting a minimum-year DSCR near 1.20x and specific programmes setting their own numbers. Whatever your policy says, you are being held to your policy, not to a standard.

Is a DSCR below 1.0 always a decline?

No, but it is always a question. Below 1.0x the borrower did not generate enough cash to pay the loan in that period, so the file has to explain what did pay it — a moratorium, a reserve, an equity injection, a guarantor — and why the next period is different.

How far can profits fall before a 1.25x DSCR breaks?

Cash flow can fall by 20%, because 1 minus 1 divided by 1.25 is 0.20. What that means in revenue terms depends on the contribution margin: at 35% it is a 5.4% revenue decline; at 12% it is 15.6%. Always convert the cushion into revenue before quoting it.

Does DSCR use EBITDA or actual cash flow?

Both are used, and they give different answers. EBITDA is faster and is what most committee papers quote; cash flow after tax, maintenance capex and working capital is what the borrower actually has. Show both, because the gap between them is a credit fact.

Is DSCR the same as the interest coverage ratio?

No. Interest coverage tests only whether the borrower can pay interest; DSCR adds scheduled principal to the denominator. A borrower can look comfortable on interest cover and fail badly on DSCR the moment amortisation begins, which is exactly what happens at the end of a moratorium.

Why does a borrower with a strong DSCR still default?

Usually because the ratio was measuring the wrong risk. Leverage, a balloon nobody would refinance, a single customer walking, or working capital absorbed by growth — none of those show up in a coverage ratio, and all of them cause defaults.

How often should DSCR be recalculated after drawdown?

At every covenant test date at minimum, and more often for seasonal or fast-growing borrowers. A ratio computed once at sanction and never revisited is documentation, not monitoring.

Can a lender approve a loan below its own DSCR floor?

Yes, if the exception is documented, approved at the right level, and supported by a compensating control — a reserve, a sweep, additional equity, a shorter tenor. What examiners look for is not the absence of exceptions but a consistent, evidenced reason for each one.

Key takeaways

• DSCR is a margin-of-safety measure, not a repayment-capacity measure. The useful number is the residual after the lender's claim, not the claim itself.
• The cushion has a formula: cash flow can fall by 1 minus 1 divided by the DSCR. At 1.25x that is 20%.
• Convert that cushion into revenue before you quote it. Our fabricator at 1.25x survives a 5.4% revenue decline; a distributor at the same 1.25x survives 15.6%.
• The lender authors part of the denominator. Moving the amortisation from five years to twelve took the same borrower from 0.94x to 1.53x.
• The DSCR you underwrite and the DSCR the agreement tests are different instruments with different definitions, different periods and different consequences. Put both in the memo.
• The ratio is blind to leverage, working capital, refinancing risk, customer concentration, intra-year seasonality and deferred capex. Each of those needs its own test.
• Overriding a DSCR is legitimate when the ratio measures the wrong thing for this borrower — provided the reason is written down and a compensating control replaces it.

Watch YuSight spread a real balance sheet — bring one borrower's accounts and see the DSCR, its derivation and its sensitivities computed from the source documents, every figure traceable to the page it came from.

Next: every DSCR formula and when each applies, DSCR calculation for term loans in India, and global cash flow analysis in commercial lending.