Credit Analysis Ratios: The 24 Metrics That Actually Drive a Lending Decision

Twenty-four ratios carry a commercial credit decision: four leverage, three coverage, three liquidity, five efficiency, five profitability and four quality-of-earnings checks. Coverage decides whether the loan gets repaid. Leverage decides how much room there is when it does not. Everything else explains why those two numbers look the way they do.

Key facts

• YuSight cites 100% of the figures it computes, with one-click verification back to the source document and page — so a ratio in the memo can be traced to the line item, the statement and the page it came from rather than to a spreadsheet nobody can reconstruct.
• Only 42% of employer firms that applied for financing received the full amount sought, against 36% receiving some or most and 22% receiving none (Federal Reserve Banks, *2026 Report on Employer Firms*, 3 March 2026).
• Regulators do not set ratio thresholds. The OCC's Rating Credit Risk booklet says only that comparing a borrower's ratios "with prior periods and industry or peer group norms can identify potential weaknesses" (OCC, *Rating Credit Risk*, Comptroller's Handbook). Every threshold in this article is a lender policy convention, marked indicative.
• The best-known leverage threshold in commercial lending is no longer supervisory guidance. The OCC and FDIC withdrew from the 2013 Interagency Guidance on Leveraged Lending — the source of the 6x total-debt-to-EBITDA marker — on 5 December 2025 (OCC Bulletin 2025-44). The Federal Reserve did not withdraw.
• Peer benchmarks come from 644 NAICS industries in Statement Studies, published by ProSight Financial Association — the body created when the Risk Management Association merged with BAI in 2024 (ProSight, *Statement Studies*).

Why can't a ratio glossary win a credit decision?

Search "current ratio" and you get Investopedia, Corporate Finance Institute and Wall Street Prep, in some order, with definitions that are correct and complete. They own that query and they deserve to. There is nothing left to say about what the current ratio is.

That is not the analyst's problem. The analyst's problem is that Meridian Components, the borrower in this article, has a current ratio of 1.16x, a Debt/EBITDA of 2.98x, an interest cover of 3.71x and a return on capital employed of 17.7% — four numbers that read like a comfortable credit — and a DSCR of 1.11x, a quick ratio of 0.62x and a cash conversion of 57%, which read like a company that is one bad quarter from a covenant conversation.

Both sets are true. They are computed from the same audited statements. The decision lives in what you do about the contradiction.

So this article is organised around three things a definition page cannot give you: which ratios actually change the answer, how each one is commonly gamed or misread, and what to do when two of them disagree. Everything is computed from one borrower so you can watch them interact.

Which ratios actually move a lending decision?

Of the 24 below, three carry disproportionate weight in almost every commercial credit committee:

1. DSCR, because it is the only ratio that directly answers "does the cash flow pay this loan".
2. TOL/TNW (or its market's equivalent leverage gate), because it sets the loss cushion.
3. The cash conversion cycle, because it explains whether growth generates cash or consumes it — and in working-capital lending, that is the entire question.

A second tier — interest cover, current ratio, EBITDA margin, ROCE — is used mostly to rate the credit rather than to approve or decline it. They feed the internal rating grade, the pricing grid and the covenant package.

And a third tier exists mainly to catch lies. Cash conversion, capex-to-depreciation, other income as a share of PBT and contingent liabilities to net worth do not tell you whether the borrower is good. They tell you whether the first two tiers can be believed.

The ratios that make careers are not the ones with the tightest thresholds. They are the ones that disagree with each other.

The borrower: one set of financials, 24 ratios

Meridian Components is a mid-market auto-components manufacturer. Figures are currency-neutral and stated in millions; the arithmetic works identically in dollars, rupees or dirhams at the same scale.

Income statement, FY2026 (FY2025 in brackets)

Balance sheet at 31 March 2026 (FY2025 in brackets)

Derived figures used throughout

Total debt = 44.0 + 9.0 + 40.0 = 93.0 Net debt = 93.0 − 6.0 = 87.0 Total outside liabilities (TOL) = total liabilities = 138.0 Tangible net worth (TNW) = 78.0 − 8.0 intangibles = 70.0 Change in working capital = (52.0−44.0) + (46.0−40.0) − (34.0−30.0) = 10.0 Cash from operations = 31.2 − 10.0 − 3.45 = 17.75 Capex 14.0, of which maintenance capex 8.5 Operating lease payments 3.6 per year FY2027 debt service = principal 9.0 + interest 8.4 = 17.4 Contingent liabilities 18.0 (LCs 7.0, corporate guarantee to group company 11.0)

Two disclosures matter later: 6.0 of the long-term debt is a director's loan, and 11.0 of the 18.0 contingent liability is a guarantee given to a group company. Neither shows up in a ratio unless you decide it should.

Leverage: how much room is there when it goes wrong?

1. TOL/TNW — 138.0 ÷ 70.0 = 1.97x. Tests the cushion: how many units of somebody else's money sit on each unit of the owner's real money. It is the broadest leverage measure because TOL captures payables, provisions and accruals, not just borrowings — so a company stretching its suppliers cannot hide it here.

How it is gamed: by reclassifying the director's loan as quasi-equity. Do that and TOL falls to 132.0, TNW rises to 76.0, and the ratio improves to 1.74x — a 12% improvement with no cash movement. It is a legitimate treatment only with an executed, unconditional subordination and non-withdrawal undertaking on file. Without the document, it is a rounding error dressed as a credit strength. Revaluation reserves are the other classic: a plant revalued upward inflates TNW without adding a unit of loss-absorbing cash.

2. Total Debt/EBITDA — 93.0 ÷ 31.2 = 2.98x. Tests how many years of current earnings it would take to clear the debt. The 2.98x is the number that will get quoted in committee, and it will sound fine.

How it is misread: the well-known 6x marker came from the 2013 Interagency Guidance on Leveraged Lending, which the OCC and FDIC formally withdrew on 5 December 2025 (OCC Bulletin 2025-44). It was never a rule for mid-market lending anyway. The real gaming is on the denominator: add-backs. Every "one-time" restructuring cost, every pro-forma synergy, every management adjustment inflates EBITDA and deflates the ratio. Ask for the bridge from audited EBITDA to adjusted EBITDA, line by line, and price the difference.

The other trap: leases. Meridian pays 3.6 a year on operating leases. Capitalise them at a conventional 8x and debt becomes 121.8 against EBITDA of 34.8 — 3.50x, not 2.98x. Under IFRS 16 those leases are already on balance sheet; under a local GAAP that still keeps them off, they are not. Two borrowers with identical economics can differ by half a turn purely on reporting framework.

3. Net Debt/EBITDA — 87.0 ÷ 31.2 = 2.79x. Same test, giving credit for cash. Only use it when the cash is genuinely available: not margin money against LCs, not restricted deposits, not a balance that appears on 31 March and disappears on 3 April. Meridian's 6.0 is thin relative to a 44.0 revolver, which itself says the cash is a float, not a buffer.

4. Gearing — 93.0 ÷ 78.0 = 1.19x. The narrow leverage view, borrowings against book equity. It is the ratio most often written into loan documents because it is unambiguous. It is also the least informative of the four, because it ignores trade credit entirely. A borrower can have gearing of 0.8x and be funding itself on 120 days of supplier credit.

When leverage ratios disagree with each other: TOL/TNW at 1.97x and gearing at 1.19x are telling you that a meaningful part of Meridian's funding is non-debt — payables of 34.0. That is not automatically bad, but it is a claim that can be withdrawn in days, unlike a term loan.

Coverage: does the cash flow actually pay the loan?

5. DSCR — and the reason it has two answers. This is the single most consequential number in the file, and it is not one number.

Definition A (traditional, accrual): (PAT 10.35 + D&A 9.6 + interest 8.4) ÷ 17.4 = 28.35 ÷ 17.4 = 1.63x Definition B (cash available for debt service, after tax and maintenance capex): (EBITDA 31.2 − tax 3.45 − maintenance capex 8.5) ÷ 17.4 = 19.25 ÷ 17.4 = 1.11x

Same borrower, same year, same statements. 1.63x approves; 1.11x goes to committee with conditions. Definition A adds back all depreciation and ignores that Meridian must spend 8.5 a year keeping its presses running. Definition B does not.

How it is gamed: excluding interest from the denominator. Compute (EBITDA − tax − capex) ÷ principal alone and you get 2.14x. Some borrowers' CA-certified projections do exactly this. Read the denominator before you read the ratio.

Which definition is right depends on the facility. For an amortising term loan against equipment, Definition B is the honest one. For a full walkthrough of the cash-flow build — including household and affiliate layers — see how global DSCR is calculated for SBA 7(a) loans.

6. Interest Coverage — 31.2 ÷ 8.4 = 3.71x. Tests whether the borrower can carry the cost of debt before worrying about repaying it. On EBIT rather than EBITDA it is 21.6 ÷ 8.4 = 2.57x, and for a capital-intensive manufacturer the EBIT version is the more truthful one.

How it is misread: interest expense in the P&L is net of interest capitalised into fixed assets during a project. A borrower mid-expansion can show 8.4 of P&L interest while paying 11.0 in cash. Reconcile P&L interest to the bank's own interest debits before trusting the ratio. The implied rate here is 8.4 on average debt of 90.5, or 9.3% — plausible; if the implied rate comes out at 4% on a mid-market borrower, interest has been capitalised somewhere.

7. FCCR — (31.2 − 3.45 − 8.5 + 3.6) ÷ (17.4 + 3.6) = 22.85 ÷ 21.0 = 1.09x. The strictest of the three, because it treats lease rent as what it is: a fixed, non-negotiable claim on cash that ranks ahead of the lender. For an asset-light borrower with large leases, FCCR is the only coverage ratio that matters. Meridian at 1.09x has almost no headroom once leases are in.

When coverage and leverage disagree: Meridian is the textbook case. Debt/EBITDA of 2.98x says moderate leverage; FCCR of 1.09x says the amortisation schedule is too fast for the cash flow. The leverage is fine; the structure is wrong. That is a restructuring conversation — extend tenor, or convert part of the revolver to a term loan with a matched profile — not a decline. Leverage ratios are about the size of the debt. Coverage ratios are about its shape. When they disagree, change the shape.

Liquidity: can it survive the next 90 days?

8. Current ratio — 109.0 ÷ 94.0 = 1.16x. Tests whether short-term assets cover short-term claims. The 1.33x convention exists because it is the reciprocal of a 25% margin on current assets — the traditional lender's stake in the borrower's working capital.

How it is gamed: the placement of current maturities. Exclude the 9.0 current portion of long-term debt from current liabilities and the ratio becomes 109.0 ÷ 85.0 = 1.28x. Some spreading templates do this by default. Also watch the "other current assets" line — 5.0 here — which routinely hides advances to related parties and unrecoverable statutory deposits.

9. Quick ratio — (109.0 − 46.0 − 5.0) ÷ 94.0 = 58.0 ÷ 94.0 = 0.62x. Strips out inventory on the assumption that inventory does not become cash quickly. At 0.62x, Meridian cannot meet its current claims without selling stock, and its stock takes 100 days to turn.

10. Working capital cycle — 99.9 + 79.1 = 179 days. The gross operating cycle: how long a unit of cash spends inside the business between buying raw material and collecting from the customer, before considering supplier credit. It is the number that sizes the working-capital facility. Six months of operating cost tied up in the cycle is why Meridian's revolver is 44.0 against revenue of 240.0.

When liquidity and profitability disagree: ROCE of 17.7% with a quick ratio of 0.62x means the business earns well and holds no reserve. That combination is profitable and fragile, and it is exactly the profile that fails on a single customer default rather than on a downturn. The right response is a covenant on liquidity, not a decline on profitability.

Efficiency: where is the cash actually going?

11. DSO — (52.0 ÷ 240.0) × 365 = 79 days. Tests collection. Compare it to the credit terms the borrower says it gives. If the sales policy says 45 days and DSO is 79, either the policy is fiction or a large receivable has gone bad and has not been provided for.

How it is gamed: computing DSO on year-end receivables against full-year revenue when sales are seasonal. A borrower with a March-heavy order book shows an inflated DSO; one that bills in April shows a flattering one. Use quarterly averages where you have them, and cross-check against the ageing schedule — the >180-day bucket tells you more than the ratio does.

12. DPO — (34.0 ÷ 168.0) × 365 = 74 days. Tests how much the borrower is financing itself on suppliers. Rising DPO is one of the earliest and most reliable distress signals in commercial lending, and it precedes the bank statement going irregular by months.

How it is misread: a high DPO is not automatically strength. A borrower with a 74-day DPO and a stated supplier term of 45 days is 29 days late to its suppliers, which is a form of unsecured borrowing at an implied cost of the discount forgone.

13. Inventory days — (46.0 ÷ 168.0) × 365 = 100 days. Tests whether stock is turning. Compute it on COGS, not revenue: computing on revenue gives 70 days here and understates the problem by 30 days.

14. Cash conversion cycle — 100 + 79 − 74 = 105 days. The one efficiency number to keep if you keep only one. It says Meridian funds 105 days of operations itself. At a COGS run rate of 168.0, that is roughly 48.3 of permanent funding requirement — against a revolver of 44.0 and net working capital of only 15.0. The facility is fully drawn because the cycle demands it.

15. Asset turnover — 240.0 ÷ ((216.0 + 196.0) ÷ 2) = 240.0 ÷ 206.0 = 1.17x. Tests how much revenue the asset base produces. Low turnover with healthy margins usually means a recent capex cycle that has not yet earned out — which is consistent with capex of 14.0 against depreciation of 9.6.

Profitability: is the business worth lending to at all?

16. Gross margin — 72.0 ÷ 240.0 = 30.0%. For a lender the level is less interesting than the trend and the volatility. FY2025 was 61.5 ÷ 205.0 = 30.0% — flat, which for a components maker passing through steel prices is genuinely reassuring.

How it is gamed: by moving costs between COGS and operating expenses between periods. Check that the classification is consistent year on year before reading anything into a margin move.

17. EBITDA margin — 31.2 ÷ 240.0 = 13.0%, down from 27.9 ÷ 205.0 = 13.6%. Revenue grew 17.1% while EBITDA grew 11.8%. Meridian bought growth with margin. That is a fact worth one sentence in the memo and a question to management.

18. PAT margin — 10.35 ÷ 240.0 = 4.31%. The most distorted of the margins, because it sits below interest, tax and every accounting policy choice. Useful mainly as a check that the company is not loss-making after financing costs.

19. ROCE — and the second ratio with two answers.

Capital employed as total assets less current liabilities: 21.6 ÷ (216.0 − 94.0) = 21.6 ÷ 122.0 = 17.70% Capital employed as net worth plus total debt: 21.6 ÷ (78.0 + 93.0) = 21.6 ÷ 171.0 = 12.63%

The first definition removes short-term borrowings from capital employed because they sit inside current liabilities. For a borrower funding itself on a 44.0 revolver, that flatters ROCE by five percentage points. Meridian's implied cost of debt is 9.3%. At 17.7% ROCE the business comfortably out-earns its debt; at 12.6% the spread is three points and thinning. Pick the definition that includes all funded debt, and say which one you used.

20. ROE — 10.35 ÷ ((78.0 + 69.0) ÷ 2) = 10.35 ÷ 73.5 = 14.1%. On closing net worth it is 13.3%. ROE is the shareholder's ratio, not the lender's, and it improves when leverage rises — which is precisely the wrong incentive from a lender's seat. Read it alongside gearing or not at all.

Quality of earnings: can the first twenty be believed?

21. Cash conversion — 17.75 ÷ 31.2 = 56.9%. The most important number on this page after DSCR. It says that of every 100 of accounting profit before interest, tax and depreciation, only 57 arrived as cash. The other 43 went into receivables and inventory.

One weak year is a growth year. Three consecutive years below 70% is a company whose profit is an accounting event.

22. Capex ÷ depreciation — 14.0 ÷ 9.6 = 1.46x. Tests reinvestment. Above 1.0x means the asset base is growing; well below 1.0x for several years means the borrower is harvesting its plant, and the depreciation add-back in the DSCR is a fiction that will come due as a large replacement capex.

23. Other income ÷ PBT — 0.6 ÷ 13.8 = 4.35%. Tests whether profit comes from the business. At 4.35% it does. When this crosses 20-25% — typically forex gains, treasury income or a property sale — the operating business is not covering its own costs, and the DSCR built on it will not repeat.

24. Contingent liabilities ÷ TNW — 18.0 ÷ 70.0 = 25.7%. Tests off-balance-sheet exposure. And here is the detail: 11.0 of the 18.0 is a corporate guarantee to a group company. If that company falls over, 11.0 becomes a real liability, TNW drops to 59.0, and TOL/TNW moves from 1.97x to 2.53x overnight. The guarantee is the largest single risk on this balance sheet and it does not appear in twenty-three of these ratios.

What does an analyst do when two ratios disagree?

Four rules that resolve most conflicts:

1. Cash beats accrual. When cash conversion contradicts a margin, believe the cash. Meridian's 13.0% EBITDA margin and 56.9% cash conversion disagree; the cash is the fact.
2. Coverage beats leverage on structure; leverage beats coverage on size. A thin FCCR with moderate Debt/EBITDA is a tenor problem. A comfortable FCCR with high Debt/EBITDA is a refinancing problem.
3. Balance-sheet-date ratios lose to flow ratios. Current ratio and quick ratio are photographs taken on one day, and the day is chosen by the borrower. DSO, DPO and inventory days average across the year and are far harder to dress.
4. A ratio you cannot trace is not evidence. Which is the whole argument for financial spreading that carries the source of every figure with it.

Then write the disagreement into the memo. A credit assessment memo that reports 24 ratios and reconciles none of them has not analysed anything — see the structure a memo should follow and the 47-item approval checklist for where the reconciliation belongs.

How many ratios should a credit analysis actually include?

Fewer than you compute. Compute all 24; report the eight that changed the answer and the three that could have but did not. A memo carrying every ratio at equal weight is a data dump, and the committee will read the first table and stop.

The recommendation for Meridian, written from the 24 above: approve the working-capital limit at 48.0 against a demonstrated cycle requirement, restructure the term amortisation to lift FCCR above 1.20x, take a quarterly covenant on cash conversion at 70%, and obtain either the release or the quantification of the 11.0 group guarantee before disbursement. Four decisions, all traceable to a specific ratio.

FAQ

Which financial ratios do lenders look at most?

DSCR, TOL/TNW and the cash conversion cycle carry the most weight in commercial credit. Interest cover, current ratio and Debt/EBITDA usually decide the internal rating grade and the pricing rather than the approve-or-decline itself.

What ratios go into a credit memo?

All 24 in this article, grouped as leverage, coverage, liquidity, efficiency, profitability and quality-of-earnings checks, with three years of history and the peer benchmark alongside. Report the ones that moved the decision prominently and put the rest in an annexure.

How many ratios should a credit analysis include?

Compute twenty-plus, report about ten. The value is in which ones you highlight and how you reconcile the ones that contradict each other, not in the length of the table.

Is there a regulatory minimum for these ratios?

No. The OCC's Rating Credit Risk booklet asks examiners to compare ratios against prior periods and peer norms without prescribing thresholds. Every number in this article is a lender policy convention, which is why they are marked indicative.

Why does DSCR have more than one answer?

Because the numerator and denominator are both conventions. Meridian's DSCR is 1.63x on profit after tax plus depreciation plus interest, and 1.11x on EBITDA less tax less maintenance capex. Write your definition into policy and apply it to every file.

What is a good cash conversion ratio?

Sustained cash conversion above 75% of EBITDA is healthy. One year in the 50s during a growth phase is normal. Three consecutive years below 70% means the profit is being reported but not collected.

Should the director's loan count as equity?

Only against an executed, unconditional subordination and non-withdrawal undertaking. For Meridian it moves TOL/TNW from 1.97x to 1.74x, which is a real improvement in the ratio and no improvement at all in the borrower's cash position.

Do contingent liabilities belong in ratio analysis?

Yes, as their own metric. Meridian's 11.0 group guarantee would push TOL/TNW from 1.97x to 2.53x if called, and it does not appear anywhere in the other twenty-three ratios.

Is the 6x Debt/EBITDA leverage limit still a supervisory expectation?

Not for OCC- and FDIC-supervised banks. Both agencies withdrew from the 2013 Interagency Guidance on Leveraged Lending on 5 December 2025. The Federal Reserve did not withdraw, and in any case the marker was written for leveraged transactions, not mid-market lending.

How should ratios be benchmarked against peers?

Against a same-industry, same-size dataset rather than an all-industry average. Statement Studies, now published by ProSight Financial Association, covers 644 NAICS industries drawn from member banks' own borrowers.

Key takeaways

• Twenty-four ratios, four groups of judgement: leverage sizes the debt, coverage shapes it, efficiency explains the working-capital ask, and quality-of-earnings checks tell you whether to believe any of it.
• Meridian Components is not a marginal credit or a good one. It is a well-run business with a badly shaped debt structure, and only the disagreement between Debt/EBITDA at 2.98x and FCCR at 1.09x reveals that.
• Two of the 24 have more than one defensible answer. DSCR is 1.11x or 1.63x; ROCE is 12.6% or 17.7%. Fix your definitions in policy, or every analyst will pick the one that supports the recommendation they already wrote.
• The ratio that will actually cost money here — a corporate guarantee worth 25.7% of tangible net worth — is invisible in twenty-three of the twenty-four.

Every ratio in this article was computed from one set of statements, and every one of them should be traceable back to the line item, the page and the document it came from. YuSight's Financial Spreading module extracts and standardises the financials, computes leverage, coverage, liquidity, efficiency and profitability ratios plus your own custom ones, and cites 100% of the figures with one-click source verification — analyst-editable throughout, so the judgement stays where it belongs.

Watch YuSight spread a real balance sheet — [book a live demo](https://yuverse.ai/yusight).