Interest Coverage and the Ratios a Lender Actually Tests

Covenants: how a solvent company defaults

A default does not require a missed payment. Most corporate loan agreements contain financial covenants: ratios the borrower promises to keep above or below a stated level, tested on fixed dates, usually quarterly. Break one and you are in default even though every instalment was paid on the day it fell due, the order book is full and the bank account is in credit. That is the whole point of the mechanism — the lender wants a tripwire that fires before the money runs out, not after.

A standard package for a mid-market borrower contains two or three of them, and they measure genuinely different things. A coverage test asks whether current earnings can service current obligations. A leverage test asks whether the total stock of debt is proportionate to the earnings that support it. You can pass one and fail the other, and this article works exactly that case. Everything below uses one company: EBITDA of $8,000,000, depreciation and amortisation of $3,000,000 giving EBIT of $5,000,000, interest of $1,500,000, gross debt of $30,000,000 against cash of $2,000,000 for net debt of $28,000,000.

Interest coverage and times interest earned are the same ratio

Settle this one first, because it wastes more meeting time than it deserves. The interest coverage ratio is EBIT ÷ interest expense. Times interest earned is EBIT ÷ interest expense. They are the same calculation under two names — "times interest earned" is the older phrasing, common in textbooks and in North American credit documentation, and "interest cover" is the shorter one, common in Europe. If someone tells you the two differ, ask them which definition of the numerator they mean, because that is the only place a real difference can hide.

On our company, EBIT of $5,000,000 against interest of $1,500,000 gives 3.33×. Read it plainly: operating profit is three and a third times the annual interest bill, so earnings could fall by roughly 70% before interest stopped being covered at all. A covenant set at 3.00× therefore has very little slack in it despite sounding generous — earnings only have to fall 6.25% for EBIT to hit $4,500,000 and the test to fail.

DSCR: the one that adds principal, and therefore the one that bites

The debt service coverage ratio measures earnings against the whole of debt service, not just its interest half: DSCR = EBITDA ÷ (interest + scheduled principal repayment). Some agreements refine the numerator — EBITDA less cash taxes, less maintenance capital expenditure, less dividends — and every one of those subtractions makes the test harder. The denominator is what matters here, because principal repayment is a real cash obligation that interest coverage simply does not see.

Build the case. Our company's baseline structure repays $4,000,000 of principal a year, so debt service is $1,500,000 + $4,000,000 = $5,500,000 and DSCR is 8,000,000 ÷ 5,500,000 = 1.45×, comfortably above a 1.25× covenant. Now refinance the identical business into a five-year fully amortising loan repaying $6,000,000 a year. Debt service becomes $7,500,000 and DSCR falls to 8,000,000 ÷ 7,500,000 = 1.07× — a breach. Interest coverage did not move at all: EBIT and interest are unchanged, so it is still 3.33×. To pass the 1.25× test under the new schedule the company would need EBITDA of 1.25 × $7,500,000 = $9,375,000, which is $1,375,000 more than it earns.

Push the same logic to the other extreme and it becomes obvious. Put the identical business on a bullet loan that repays nothing until maturity and DSCR is 8,000,000 ÷ 1,500,000 = 5.33×. One company, three DSCRs — 5.33×, 1.45×, 1.07× — and one unchanged interest coverage of 3.33× across all three. DSCR is not a measure of how profitable you are. It is a measure of how fast you have promised to repay.

Working a bad quarter through the covenant test

Covenants are almost always tested on trailing twelve months, which is what makes them slow to break and slow to heal. Start from four quarters of $2,000,000 of EBITDA each. Replace them one at a time with $1,700,000, then $1,400,000, then $900,000 — a business sliding, not collapsing. LTM EBITDA goes $8.0m, $7.7m, $7.1m, $6.0m, a fall of 25% over the year. Because D&A of $3,000,000 is fixed, LTM EBIT goes $5.0m, $4.7m, $4.1m, $3.0m, which falls proportionally faster: 40% against EBITDA's 25%.

Now read the three tests quarter by quarter. Interest coverage: 3.33×, 3.13×, 2.73×, 2.00× — it breaches a 3.00× covenant at the third test. DSCR on the baseline $4,000,000 amortisation: 1.45×, 1.40×, 1.29×, 1.09× — it survives the third test by four hundredths and breaches at the fourth. Net debt to EBITDA: 3.50×, 3.64×, 3.94×, 4.67× — it also breaches at the fourth. The covenant that looked loosest on day one is the one that broke first, and that is the general lesson: the binding constraint is not the ratio with the biggest headroom in multiples, it is the ratio whose numerator falls fastest.

Quantify the headroom directly and the ordering stops being a surprise. On the baseline structure, the interest cover test binds when EBITDA reaches 3.00 × $1,500,000 + $3,000,000 = $7,500,000, a fall of 6.25%. The leverage test binds at $28,000,000 ÷ 4.00 = $7,000,000, a fall of 12.5%. The DSCR test binds at 1.25 × $5,500,000 = $6,875,000, a fall of 14.06%. Those three percentages, not the three multiples, are the numbers to put in front of a board.

The leverage covenant, and why the definitions never match the accounts

The other half of a standard package is a leverage test: net debt ÷ EBITDA, capped at some multiple. It answers a different question from any coverage ratio. Coverage asks whether this year's earnings service this year's obligations; leverage asks how many years of earnings the whole debt represents, which is a proxy for whether the loan can ever be refinanced. Our company at 3.50× against a 4.00× cap has 12.5% of EBITDA headroom, and note that this test tightens automatically as earnings fall even if the company repays not one unit of debt.

Then comes the part that catches people out: none of these terms means what the accounts mean. A facility agreement defines its own EBITDA, usually capitalised as Consolidated EBITDA, and that definition can add back exceptional items, restructuring costs, share-based payment, and in leveraged deals the estimated run-rate effect of synergies not yet realised. Net Debt is likewise defined, and whether it includes lease liabilities, shareholder loans, factoring, earn-outs or trapped cash is a negotiated point, not an accounting fact. Two people computing "net debt to EBITDA" from the same audited accounts can honestly land several turns apart.

What a breach actually triggers

A breach is an event of default. In principle that lets the lender accelerate — declare the whole loan immediately repayable — cancel undrawn facilities and enforce security. In practice acceleration is rare, because a lender who accelerates a viable business usually gets less back than one who does not. What happens far more often is a negotiation: a waiver or an amendment, priced with a fee and usually a higher margin, tighter reporting, new restrictions on capital expenditure and dividends, and sometimes an equity cure — the shareholders injecting cash that the agreement allows to be treated as EBITDA for the purposes of the failed test.

There are two consequences worth planning for even when the lender is accommodating. The first is cross-default: a breach under one agreement typically constitutes a default under the others, so a single covenant failure can put every facility in the group into play at once. The second is accounting. Under IAS 1, a liability is classified as current unless the entity has a right at the reporting date to defer settlement for at least twelve months, and a covenant tested on or before that date that has been breached removes that right. A long-term loan can therefore jump to current liabilities in the balance sheet, taking the current ratio with it, unless a waiver is obtained in time. Talk to the lender before the test date, not after.

Frequently asked questions

Is interest coverage computed on EBIT or EBITDA?

Both are used, and your agreement decides which one binds you. The textbook interest coverage ratio and times interest earned use EBIT. Many credit agreements use EBITDA instead, on the argument that depreciation is not a cash cost and interest is paid in cash. The gap is not small: on our figures, EBIT gives 5,000,000 ÷ 1,500,000 = 3.33× while EBITDA gives 8,000,000 ÷ 1,500,000 = 5.33×. Never compare your ratio to anyone else's without checking which numerator each of you used.

Can a profitable company breach a covenant?

Routinely, and that is what the covenants are for. Our worked company remains profitable at every test date — LTM EBIT never goes below $3,000,000 — and still breaches all three tests within a year. Profit is a level; covenants measure ratios, and a ratio can fail because the numerator fell, because the denominator rose, or because a refinancing changed the repayment schedule while the trading did not change at all.

Why does DSCR use EBITDA rather than net income?

Because debt service is paid in cash, and net income is after two charges that distort the cash picture in opposite directions. Depreciation and amortisation reduce net income without consuming cash, so subtracting them understates capacity to pay. Interest is already inside the denominator, so leaving it in the numerator too would double-count it. Starting from EBITDA fixes both. Careful agreements then subtract the cash items EBITDA ignores — cash taxes, maintenance capital expenditure — because those genuinely compete with the lender for the same money.

What is a covenant headroom analysis?

It converts each covenant into the fall in earnings that would break it, so the tests become comparable. On our baseline structure, the interest cover test breaks at $7,500,000 of EBITDA (a 6.25% fall), the leverage test at $7,000,000 (12.5%) and the DSCR test at $6,875,000 (14.06%). Expressed that way, the ordering is obvious and the board can see that a routine 7% miss is already a default. Multiples alone hide this entirely, because 3.33× against 3.00× and 3.50× against 4.00× look similar and are not.

Do the covenant definitions really differ from the accounts?

Yes, systematically. EBITDA is not defined by IFRS or US GAAP at all, so a facility agreement has to define its own, and it will — with add-backs for exceptional items, restructuring, share-based payment and sometimes unrealised synergies. Net debt is defined too, and whether lease liabilities, shareholder loans or restricted cash count is negotiated line by line. The practical rule is to compute your covenant ratios from the agreement's definitions clause every quarter and reconcile them to the statutory accounts separately, rather than assuming the two will agree.