Debt Service Coverage Ratio (DSCR)

Every coverage ratio has the same shape: the cash a business can use to pay its lenders, over what it owes them in the period. The denominator is rarely argued about — interest plus the principal due on schedule. The numerator is where banks, policies and facility agreements differ, and it is written into the documents for that reason.

Take Highview Industries' FY25, the clean file in Issue 04. Interest was US$0.64m and scheduled principal US$0.80m: the new line's term loan at its US$0.5m a year, plus the last of an older equipment loan, repaid through the same year. Debt service of US$1.44m. Divide three numerators by it:

The same company in the same year covers its debt service 3.27×, 2.72× or 1.47×, depending on what the reader decided to count. None of the three is wrong; each answers a different question. The first asks what the business earns; the second what is left after the tax authority; the third what is left after the business has kept itself running and paid its owners. A covenant set on one and tested on another is a dispute waiting for a bad year — which is why the credit paper states its definition beside every ratio.

A DSCR is most useful looking forward, on the debt that is being asked for. Meridian Logistics, in Issue 05, has no debt today and wants US$6.0m: trucks that can be repaid over five years and a distribution-centre fit-out that should be gone inside three. At an assumed 7.5%, the first year's debt service is about US$1.85m — US$1.40m of principal and US$0.45m of interest. FY25 EBITDA less tax was US$3.63m:

Comfortable, on the year just closed. But 58% of that revenue comes from one customer whose contract can end on ninety days' notice. If it did, the US$21.8m of revenue that remains, at the same 8.2% margin, earns about US$1.79m of EBITDA before tax — 0.97× of the new debt service, and that is before the dedicated costs that do not leave on day ninety. The ratio describes the contract, not the business. That is why the senior read in Issue 05 is about structure rather than coverage: the generic trucks funded normally, the fit-out that only this customer needs paid off inside the period the break clause puts at risk.

Kowhai Retail Group, in Issue 06, owes the bank almost nothing at year-end: its seasonal line is cleaned down every February. On bank interest alone its cover looks enormous. But fourteen stores are leased, and the rent — US$2.60m in FY25 — has to be paid whether Christmas sells or not. Fixed charge cover puts it on both sides:

Interest cover says the bank is barely exposed. Fixed charge cover says the business had 1.40× of earnings for every dollar of commitments it could not walk away from, after a season that did not clear. Give back the same 5 points of margin in a second season and it falls to 0.84×: below the line where earnings cover the commitments at all. The seasonal line sits on top of those commitments, which is why the senior read in Issue 06 funds last year's true peak and leaves the rest of the uplift to the shareholders or to evidence.

For an owner-managed business the company's cash and the owner's are drawn from the same well. Global cash flow analysis adds them up: the business's cash available, the owners' other income, less their drawings, personal borrowing and anything they have guaranteed elsewhere, against every debt the group owes. It is the test of whether the people standing behind a loan can actually stand behind it.

Sometimes the useful move runs the other way. Issue 03 is a dairy family splitting two farms between two successors. Combined, the business carried its NZ$13.0m of term debt at 2.4× interest cover, with the patriarch's guarantee over all of it. The senior read declines to rely on that combined picture: it underwrites each successor's entity on its own, stressed at a NZ$6.50 payout, and releases the guarantee only as each one proves it can stand without the other. A global number can hide the part that cannot pay; taking the group apart is how a desk finds it.

A coverage ratio is an annual average of cash against an annual total of payments. It does not see the month the season peaks, the customer who can leave, or whether the earnings on top of the fraction are real. A covenant tested on reported EBITDA inherits every choice made in producing it — the provision released, the costs capitalised, the income pulled forward — which is exactly the trap in Issue 07, where the covenants were clean because the profit was flattered.