What does a covenant actually buy a lender?

A bond or loan is a contract, and the interesting part is not the coupon — it is the list of things the borrower promises to do and not do. Those promises are covenants, and they are the main lever a lender has between signing and being repaid.

Two families

• Maintenance covenants. Tested every reporting period regardless of what the borrower does. Typically: "net leverage must remain below 4.0×" or "interest cover must exceed 2.5×." Characteristic of bank loans and private credit. A breach is a default under the agreement — which sounds dramatic but usually means lenders get a seat at the table: a waiver in exchange for a fee, tighter terms, extra security, or amortisation.
• Incurrence covenants. Tested only when the borrower takes a specific action — issuing more debt, paying a dividend, selling assets, making an acquisition. Characteristic of public high-yield bonds. Do nothing and you can never breach one.

The distinction is about timing of lender leverage. Maintenance covenants give an early trigger as performance deteriorates. Incurrence covenants only bite when the borrower reaches for something.

The common terms

• Limitation on indebtedness — usually a ratio test that must be satisfied before new debt is raised.
• Restricted payments — a capped basket governing dividends, buybacks and payments to affiliates. Stops value leaving the credit box.
• Negative pledge — limits granting liens to others, which would otherwise silently push you down the queue.
• Limitation on asset sales — often with a requirement that proceeds are used to repay debt.
• Change of control put — the holder can require repayment, conventionally at 101, if the company changes hands.
• Restricted and unrestricted subsidiaries — defines which entities are inside the covenant perimeter. Assets moved outside it are, in effect, moved outside your protection.

What covenants are worth

Here is the key insight, framed in the PD-and-LGD language of this course. Covenants do relatively little to reduce the probability that a business fails; a bad market or a broken product is not something an indenture can prevent. What covenants do is protect recovery and timing:

• They stop value leaking out of the credit box before you can claim it.
• They prevent new claims being layered ahead of yours.
• They force a conversation early, while options still exist.

So covenants are principally an LGD tool, and only secondarily a PD tool. That framing tells you what to look for when reading documentation.

The arithmetic of headroom

Covenants are ratios, so you can measure how much room a borrower has.

An illustrative issuer with EBITDA of $250m, net debt of $900m, and an incurrence test at 4.0× net leverage:

• Current leverage = 900 ÷ 250 = 3.6×
• Debt capacity at the test = 4.0 × 250 = $1,000m
• Headroom = 1,000 − 900 = $100m of additional debt before the test blocks it

Now consider that the denominator is a defined term. If the documentation permits $50m of add-backs to EBITDA — for cost savings expected but not yet realised, say — the calculation becomes 4.0 × 300 = $1,200m of capacity, and headroom jumps from $100m to $300m. A $50m definitional adjustment created $200m of borrowing room. This is why practitioners read the definitions section of an indenture as carefully as the covenants themselves.

Covenant-lite

Loans issued without maintenance covenants are called cov-lite, a structure that grew substantially in leveraged lending. The trade-off usually discussed is straightforward: without a periodic test, deterioration is detected later and lenders have less leverage to force early action, so defaults tend to occur later in a decline. Empirical findings on recoveries are mixed and depend heavily on what else sits in the capital structure, so the honest statement is that the timing effect is clear and the recovery effect is contested.

Try it now

Compute headroom for a real company:

1. Below are one large borrower's EBITDA, from its income statement, and its debt, from its balance sheet, both for the same fiscal year. Compute gross leverage as long-term debt divided by EBITDA.

2. Assume a 4.0× incurrence test and calculate how much more debt could be raised before the test binds. Then redo the leverage with "Debt, short and long" in place of long-term debt, and see what happens to the headroom when the definition of debt counts the short-term borrowing too.
3. Go back to long-term debt and redo it assuming EBITDA is 10% higher through add-backs. The difference is why covenant definitions matter as much as covenant levels.