Covenant Design and Monitoring in Private Credit: A Practitioner’s Framework for Structuring Maintenance Tests That Actually Work

The Underwriting Challenge: When Covenants Fail Before the Default Does

Consider a $10M EBITDA specialty manufacturing company that secured a $40M senior secured term loan in early 2024. The deal closed with a 4.0x total leverage covenant, set with what the lender called a “comfortable” cushion to the sponsor model. Eighteen months later, EBITDA had declined 22% due to raw material inflation and a lost customer contract. Leverage had drifted to 5.1x, but the covenant didn’t trigger until 5.25x. By the time the lender had contractual intervention rights, the company had already drawn its revolver to fund operations and was negotiating with suppliers on extended payment terms.

This scenario plays out with regularity across middle-market direct lending. The problem is rarely the covenant structure in isolation, it’s the calibration. As sidley.com notes in its 2026 analysis of private credit transactions, covenants set too loosely allow borrowers to exhaust liquidity well before a breach gives lenders the ability to intervene proactively. Getting that balance right is one of the most consequential decisions an underwriter makes at the closing table.

The Analytical Framework: Calibrating Covenants to Business Reality

Effective covenant design is not about setting a number, it’s about mapping a contractual tripwire to a borrower’s actual financial trajectory under stress. The framework involves four sequential steps:

Step 1: Underwrite to Normalized EBITDA, Not Sponsor EBITDA

Before any covenant level is proposed, the underwriter must establish a credible, independently normalized EBITDA baseline. Sponsor models routinely include add-backs for one-time costs, synergies, and run-rate savings that may never materialize. A disciplined lender strips these out and underwrites to a conservative LTM or forward EBITDA figure. fitchratings.com identifies EBITDA normalization (including adjustments for non-cash provisions, asset write-downs, and one-time charges) as central to its corporate rating methodology. The same discipline belongs in private credit underwriting.

Step 2: Set the Covenant Level with a 25–35% Cushion to the Lender’s Model

Industry practice, documented by sidley.com, places the appropriate cushion between a borrower’s projected EBITDA and the covenant trigger at 25–35%. This means if normalized forward EBITDA is $10M and the leverage covenant is 4.0x, that covenant level should reflect an EBITDA assumption of roughly $7.0–$7.5M before triggering, not the sponsor’s optimistic $11M forward projection. Cushions above 35% risk creating “cov-loose” structures that render the covenant practically meaningless.

Step 3: Define the Covenant Universe: Maintenance vs. Incurrence

Maintenance covenants are tested at regular intervals, typically quarterly, regardless of borrower action. Incurrence covenants are triggered only by specific events such as additional debt issuance or acquisitions. Direct lending transactions historically favored full maintenance covenant packages; the rise of cov-lite structures imported from the syndicated loan market has eroded this in competitive upper-middle-market deals. Lenders must be clear-eyed about which type they are accepting and what monitoring obligations accompany each.

Step 4: Define Testing Mechanics, Lookback Periods, and Cure Rights

LTM EBITDA calculations can be manipulated through the selection of lookback periods, particularly around acquisitions. Lenders should specify precisely which entities are included (addressing the unrestricted subsidiary risk noted by sidley.com), how pro forma adjustments are permitted, and the mechanics of any equity cure rights that allow sponsors to inject capital to restore compliance.

Walking Through the Numbers: A $10M EBITDA Manufacturing Company

Return to our earlier borrower. At close, the lender’s normalized EBITDA was $9.5M (after removing $500K in sponsor add-backs from the $10M presented figure). The loan amount was $40M, implying a closing leverage of 4.2x on the lender’s model. Applying a 30% cushion methodology:

• Lender’s normalized EBITDA at close: $9.5M
• 30% stress scenario EBITDA: $6.65M
• Debt outstanding: $40M
• Implied stressed leverage: 6.0x
• Appropriate covenant ceiling: 5.5x (set between closing leverage and stressed scenario)

Had the lender set the covenant at 5.5x rather than 5.25x, a 22% EBITDA decline to $7.41M would have produced leverage of 5.4x, still inside the covenant. But that is precisely the point: the covenant should be calibrated to trigger before the borrower reaches distress, not after. With a tighter 5.0x covenant, the same EBITDA decline triggers the breach at $8.0M EBITDA, giving the lender intervention rights while the company still has operational flexibility to implement a remediation plan.

Stress testing should model at minimum three scenarios: base case, 15–20% EBITDA decline, and a severe 30–35% EBITDA compression. As dlpcapital.com emphasizes, disciplined lenders ask not whether a deal works in the best case, but whether they can survive delays and market disruption, then structure around that reality.

Industry Standards: What Institutional Lenders Require, and What Gets Missed

Institutional credit risk rating frameworks, including those outlined by the occ.gov Comptroller’s Handbook on Rating Credit Risk, establish that sound underwriting must coordinate repayment with cash flow, covenants, and collateral simultaneously. Covenants cannot be evaluated in isolation from coverage ratios and collateral recovery assumptions.

Common institutional lender requirements include:

• Minimum Interest Coverage Ratio (ICR): Typically 1.5x–2.0x in direct lending, tested quarterly alongside leverage
• Maximum Total Leverage: Set at closing leverage plus a defined headroom band, step-down schedules aligned to amortization or acquisition integration
• Minimum Liquidity: Absolute cash floor or available revolver covenant, particularly in cyclical sectors
• Capex Limitations: Incurrence-based caps to prevent capital allocation that crowds out debt service

The most frequently cited red flags in covenant monitoring include:

• Consistent use of maximum permitted add-backs in quarterly compliance certificates (a sign the borrower is managing to the covenant rather than to the business)
• Unrestricted subsidiary designations being used to move assets outside the covenant perimeter
• Repeated equity cure injections, more than one cure per credit cycle typically signals structural underperformance, not a one-time event
• Ratchet provisions that reset leverage ceilings upward after incremental borrowings, progressively weakening the lender’s protection

Underwriting Checklist: Covenant Design and Ongoing Monitoring

At Origination

1. ☐ Build independent normalized EBITDA, document every add-back decision and challenge any item without 12 months of audited support
2. ☐ Run three EBITDA stress scenarios (base, –20%, –35%) and calculate leverage and coverage at each level
3. ☐ Set maintenance covenant levels using a 25–35% cushion to the lender’s, not the sponsor’s, EBITDA model
4. ☐ Define restricted group perimeter: which subsidiaries are included, which are excluded, and what cash earnings can flow in from excluded entities
5. ☐ Document cure right mechanics: maximum number of permitted cures, cure contribution limits, and whether cured EBITDA can be double-counted in future periods
6. ☐ Confirm quarterly testing dates, compliance certificate delivery deadlines, and officer certification requirements

Ongoing Portfolio Monitoring

1. ☐ Review compliance certificates within five business days of receipt, flag any new add-back categories not present at close
2. ☐ Track covenant headroom on a trailing-four-quarter basis and flag when headroom narrows below 15%
3. ☐ Monitor revolver utilization trends as an early warning indicator, sustained draws above 50% warrant a lender call
4. ☐ Review management accounts monthly for revenue run-rate, gross margin, and working capital changes that may not yet appear in covenant calculations
5. ☐ Evaluate any unrestricted subsidiary designation requests with the same rigor applied to an incremental loan amendment

Key Documentation Requirements: Executed credit agreement with covenant schedules, quarterly compliance certificates with supporting EBITDA build, audited annual financials, management accounts (monthly), lender update calls (at minimum semi-annually), and any amendment or waiver history.

Key Takeaways

• 🔑 Calibration is everything: A covenant set too loosely, cushion above 35% to the lender’s model, provides false comfort and may be functionally useless by the time it triggers.
• 🔑 Stress test before you set: Model a 30–35% EBITDA compression scenario and confirm the covenant triggers with time and liquidity remaining to remediate, not after both are exhausted.
• 🔑 Cov-lite is a structural concession, not a market standard to accept passively: When cov-lite is unavoidable in competitive processes, compensate with tighter incurrence covenants, lower advance rates, and enhanced monitoring obligations.
• 🔑 The compliance certificate is a monitoring tool, not a formality: Consistent use of maximum add-backs, new add-back categories, or rising revolver draws are behavioral signals that precede financial deterioration, often by two to four quarters.
• 🔑 Covenant quality is a regulatory credit signal: Per OCC guidance, structural covenant weaknesses can justify Special Mention classifications independent of payment performance, which means weak covenant packages carry direct implications for credit ratings and reserve requirements.