Credit and Covenant Modelling: What a Lender Actually Tests

Debt sizing: three constraints, and only one binds

Lenders do not size debt off a single rule. They apply several tests simultaneously and lend to whichever is most restrictive, which means the binding constraint is often not the one the borrower was focused on.

The leverage test caps total debt at a multiple of EBITDA — commonly in the 4.0x to 6.0x range for stable mid-market businesses, less for cyclicals and more for highly contracted cash flows. The interest coverage test requires EBITDA to exceed interest by a comfortable margin, typically 2.0x to 3.0x. The cash flow test, expressed as a debt service coverage ratio, requires cash available for debt service to exceed interest plus mandatory amortisation, typically at 1.20x to 1.35x.

Model all three and take the minimum. A business with strong EBITDA but heavy capex will be constrained by cash flow coverage long before it hits the leverage cap — and a model that sizes off leverage alone will propose a structure no lender will actually fund.

• Leverage: total net debt divided by adjusted EBITDA, against the agreed cap.
• Interest cover: EBITDA divided by net cash interest.
• Debt service cover: cash available for debt service divided by interest plus mandatory amortisation.
• Take the minimum of the three. That is the loan.

Building the covenant schedule

Covenants are tested at defined dates — usually quarterly, on a rolling twelve-month basis — against thresholds that typically tighten over the life of the facility as the business is expected to deleverage.

The schedule needs a row per covenant per test date: the calculated ratio, the threshold in force at that date, the headroom in percentage terms, and a pass/fail flag. Conditional formatting on the flag means a breach is visible without reading numbers.

Two mechanical details matter. Covenant EBITDA is defined in the credit agreement and is rarely identical to reported EBITDA — the permitted add-backs, and any cap on them, are negotiated and must be modelled as defined rather than as you would calculate EBITDA yourself. And the test is usually on a rolling last-twelve-months basis, so a quarterly model needs the trailing calculation built explicitly rather than annualising a single quarter.

The downside case is the real model

Lenders underwrite the downside. The base case establishes that the structure works when things go to plan, which is table stakes; the credit decision turns on what happens when they do not.

A credible downside is not the base case with every line reduced by a uniform percentage. It is a scenario with a stated cause and consistent consequences: a revenue decline of a specified magnitude, with the fixed and variable cost split determining how much of that drops through to EBITDA, working capital behaving as it actually does in a downturn (receivables slow, inventory builds), and capex reduced only to the extent maintenance requirements allow.

Then two further tests. The breakeven analysis: how far can EBITDA fall before the tightest covenant is breached? That single percentage is the most useful output in the whole credit model. And the liquidity test: does the revolver plus available cash cover the trough, and for how long?

Headroom, and what lenders consider adequate

Headroom is the gap between the projected ratio and the covenant threshold, expressed as the percentage EBITDA decline that would cause a breach. It is the number credit committees discuss.

Typical expectations run in the 25% to 35% range at the tightest test point for a stable business, and higher for anything cyclical. A structure showing 10% headroom in year two is not a structure a lender will approve, however strong the base case looks — because the base case is not what they are pricing.

The corollary for a sponsor: the maximum debt the model says is affordable and the maximum debt a lender will actually provide are different numbers, and the gap is headroom. Building the credit view alongside the equity view avoids proposing a structure that dies in credit committee.

Cash sweeps, baskets and the terms that change the model

Beyond the ratios, several credit agreement mechanics have direct modelling consequences and are frequently omitted.

The excess cash flow sweep requires a defined percentage of surplus cash — often stepping down as leverage falls — to prepay debt. It materially changes the deleveraging path and therefore equity returns.

Equity cure rights let the sponsor inject equity to fix a covenant breach, usually limited in number and frequency. Whether a cure counts as EBITDA or as debt reduction is negotiated and changes the arithmetic.

Permitted baskets for acquisitions, disposals and restricted payments determine what the borrower can do without consent, and any model assuming a bolt-on acquisition strategy needs to check the acquisition basket supports it.