Covenant Design and Monitoring in Middle Market Direct Lending: How to Structure Maintenance Tests That Actually Work

The Underwriting Problem: Covenants That Do Not Protect

Here is the scenario every credit officer should fear: a portfolio company’s EBITDA falls 30% over six quarters. Leverage quietly moves from 4.5x to 7.8x. The CFO misses two internal budget targets. And your maintenance covenant — set at a maximum 8.25x total leverage — hasn’t been breached once.

This is not a hypothetical. According to an April 2025 analysis by S&P Global Ratings, a leveraged buyout transaction with debt-to-EBITDA of 5.0x at closing may carry a maximum total leverage ratio threshold of approximately 8.25x as its maintenance covenant, tested quarterly. S&P Global Ratings further found that more than a third of credit-estimated borrowers subject to actively tested leverage-based maintenance covenants appear to have loose thresholds — with current headroom above 40% — and that in the majority of cases, these large cushions reflect covenants set very wide at closing rather than meaningful deleveraging progress.

Covenant design is not a documentation formality. It is a credit control mechanism. Done poorly, it gives lenders the illusion of protection while stripping them of any real intervention right. Done well, it provides a calibrated early-warning trigger that forces a borrower to the negotiating table before cash flow deteriorates into a default.

This guide walks through how to build maintenance covenants that function as intended.

The Analytical Framework: What Maintenance Covenants Are Actually Doing

Before designing a covenant package, a credit underwriter must be precise about the instrument’s purpose. Maintenance covenants and incurrence covenants serve fundamentally different functions. As explained in institutional leveraged finance practice, maintenance covenants require borrowers to continuously satisfy financial tests at regular intervals — typically quarterly — regardless of any specific action, while incurrence covenants only apply when the borrower takes a specific action such as incurring new debt or paying a dividend. The practical consequence: maintenance covenants give lenders earlier intervention rights during performance deterioration, while incurrence covenants preserve borrower flexibility until a specific trigger is attempted. ibinterviewquestions.com

In the middle market, maintenance covenants remain standard — but their effectiveness is eroding. As Sidley Austin noted in a March 2026 analysis, a typical leverage covenant in a direct lending transaction may be set with a 25–35% cushion to the EBITDA projected in a sponsor or borrower model that is prepared and delivered to the lender prior to closing — and cushions above this level create a so-called “cov-loose” structure. sidley.com

The three-covenant architecture that experienced credit officers use in the lower and core middle market combines:

• Maximum Total Net Leverage Ratio — the primary test, comparing total debt net of unrestricted cash to trailing twelve-month (TTM) EBITDA, tested quarterly with a step-down schedule.
• Minimum Fixed Charge Coverage Ratio (FCCR) — measuring EBITDA less unfinanced capex divided by total fixed charges (cash interest plus scheduled amortization). A floor of 1.10x–1.20x is standard in lower middle market transactions.
• Minimum Liquidity — a cash-plus-revolver-availability floor that catches liquidity stress independently of earnings-based tests, particularly important for businesses with working capital volatility or high fixed-cost structures.

The Federal Reserve’s Interagency Guidance on Leveraged Lending is explicit on this point: well-defined underwriting standards must define acceptable leverage levels and describe amortization expectations for senior and subordinate debt, with projections stress-tested for downside scenarios including a covenant breach. federalreserve.gov

Worked Example: A $10M EBITDA Specialty Manufacturer

Consider a sponsor-backed specialty manufacturer with $10.0M TTM EBITDA, $40M total debt (4.0x leverage at close), $1.5M mandatory amortization, and $2.0M normalized maintenance capex. The sponsor model projects $10.5M EBITDA in Year 1 and $11.2M in Year 2. Below is how a properly calibrated covenant package should be structured versus a cov-loose alternative:

The properly calibrated structure gives the lender an actionable event of default if EBITDA deteriorates to $7.6M — a 24% decline from the closing baseline. The cov-loose structure does not trigger until EBITDA falls nearly 47%. In a material operational stress, those 23 percentage points of EBITDA degradation represent months of value destruction before the lender has any seat at the negotiating table.

The Analysis: Building Covenants from the Ground Up

The most common underwriting error in covenant construction is working backward from a desired answer — selecting a threshold that will not be breached under the sponsor’s optimistic model — rather than forward from a risk question: at what level of EBITDA deterioration do I need intervention rights?

Step 1 — Anchor to Your Independent Base Case, Not the Sponsor Model. As the OCC’s credit risk rating guidance makes clear, there is no substitute for rigorous analysis of a borrower’s financial statements, with quantitative analysis of revenues, profit margins, income and cash flow, leverage, liquidity, and capitalization sufficiently detailed to identify trends and anomalies. occ.gov The starting EBITDA for covenant calibration must be your model — audited, normalized, and stripped of non-recurring items — not the sponsor’s projected case.

Step 2 — Set the Leverage Covenant at 25–35% Headroom to Your Base Case. This is the market standard in the lower and core middle market. Headroom beyond 35% begins entering cov-loose territory that provides negligible practical protection. On the $10M EBITDA manufacturer, 30% headroom to a lender-constructed $10.0M base case anchors the initial covenant at approximately 5.25x on $40M of debt — not 8.25x.

Step 3 — Model the Covenant Breach Trigger Against Your Downside Scenario. The Federal Reserve guidance is explicit: projections should be stress-tested for one or more downside scenarios, including a covenant breach. federalreserve.gov If your downside scenario (e.g., −10% revenue, −200 bps margin compression) does not trigger the leverage covenant, the covenant has been set too loosely. The breach threshold must sit within the range of a plausible but non-catastrophic deterioration.

Step 4 — Include a Step-Down Schedule. Covenants that remain static throughout a five-to-seven year loan term do not reflect the lender’s legitimate expectation of de-leveraging. A quarterly step-down — for example, 5.25x declining 0.25x per year — requires the borrower to actually pay down debt or grow EBITDA to maintain compliance. The Interagency Guidance notes that supervisors commonly evaluate the ability to fully amortize senior secured debt or repay at least 50% of total debt over a five-to-seven year period as evidence of adequate repayment capacity. federalreserve.gov

Step 5 — Govern the EBITDA Definition. Arguably the most consequential covenant design decision is the definition of EBITDA used in the compliance calculation. S&P Global noted that EBITDA calculations for covenant compliance purposes usually include various add-backs that can obfuscate a borrower’s true earnings power — a risk that is amplified when headroom is already wide. S&P Global Ratings Credit underwriters should scrutinize: (a) run-rate synergy add-backs with no documentation trail, (b) pro forma EBITDA from acquisitions not yet integrated, and (c) management consulting fees characterized as non-recurring. Each of these inflates compliance EBITDA without reflecting actual cash generation.

Industry Standards, Market Data & Red Flags

The Covenant Erosion Problem in 2025–2026

The erosion of covenant quality in private credit is not a theoretical concern — it is a documented structural shift with measurable default cycle implications. Key data points from current sources:

Red Flags in Covenant Documentation

Experienced lenders watch for specific structural weaknesses at the term sheet and credit agreement stage:

• Unlimited or uncapped EBITDA add-backs: Borrower-friendly credit agreements that allow projected synergies, cost savings, and operational improvements to be added to EBITDA without a cap or sunset date can effectively double the compliance EBITDA relative to actual earnings.
• Unrestricted subsidiary carve-outs: As Sidley Austin noted, credit agreements in the competitive sponsor-backed market may permit designation of “unrestricted subsidiaries” that fall outside the covenant framework — subsidiaries whose debt and EBITDA contribution are excluded from the leverage calculation. sidley.com
• Auto-reset provisions: Covenant levels that automatically reset to a wider threshold if certain conditions are met eliminate the step-down protection that gives lenders leverage over time.
• Revolving-lender-only covenants: In some cov-lite structures, financial maintenance covenants benefit only revolving lenders, with term lenders receiving only cross-acceleration rights — and quarterly testing is triggered only if revolving credit usage exceeds a threshold of 40% or higher. sidley.com This means term loan holders may have no independent intervention right.
• Cash-netting provisions without limits: Allowing unrestricted cash to net against gross debt in the leverage calculation can mechanically suppress the reported ratio even as operational performance deteriorates.

The Basel Committee on Banking Supervision’s 2025 updated Principles for the Management of Credit Risk emphasize that banks must have comprehensive policies, processes and methodologies for grading, classifying and monitoring all credit exposures, with processes that define criteria for identifying and reporting potential problem exposures in a timely manner to ensure they are subject to more frequent monitoring and possible corrective action. bis.org A covenant with 40%+ headroom does not satisfy this standard.

Covenant Design & Monitoring Checklist

The following checklist applies to any new direct lending origination in the middle market. Each item should be completed at underwriting and revisited at every quarterly monitoring cycle.

At Origination

1. ✅ Build your own EBITDA model — normalize TTM EBITDA from audited financials; do not accept the sponsor model as the baseline for covenant calibration.
2. ✅ Set maximum leverage at 25–35% headroom to your base case — document the headroom calculation explicitly in the credit memo; any cushion above 35% requires committee-level justification.
3. ✅ Include a fixed charge coverage ratio covenant — minimum 1.10x–1.15x FCCR captures debt service capacity independently of the leverage ratio and is particularly important in businesses with high amortization schedules.
4. ✅ Add a minimum liquidity covenant — particularly for asset-intensive, seasonal, or acquisition-oriented borrowers. Set the floor at a meaningful multiple of peak monthly cash burn.
5. ✅ Negotiate a quarterly step-down schedule — covenant levels should decline over the loan term to reflect expected de-leveraging; a borrower that refuses step-downs is signaling limited confidence in its own projections.
6. ✅ Define EBITDA with precision — cap total add-backs (synergies, cost saves, non-recurring items) at a fixed percentage of unadjusted EBITDA (e.g., 20–25%) with an 18-month sunset on any forward-looking add-back.
7. ✅ Restrict unrestricted subsidiaries — require that any subsidiary holding material assets or generating material revenue be a restricted subsidiary subject to the covenant package.
8. ✅ Limit equity cure mechanics — equity cures should be limited in frequency (e.g., no more than two in any rolling four-quarter period), capped in amount, and applied first to reduce debt rather than inflate EBITDA.

Quarterly Monitoring Protocol

1. ✅ Receive compliance certificates within 45 days of quarter-end — require CFO certification with full EBITDA bridge from GAAP net income to covenant EBITDA, supported by a working paper trail.
2. ✅ Calculate headroom independently — do not rely solely on the borrower’s compliance certificate; reconstruct the calculation from financial statements to verify add-back legitimacy.
3. ✅ Track variance from original underwriting model — the Federal Reserve Interagency Guidance requires that transactions be reviewed at least quarterly to determine variance from plan, the related risk implications, and the accuracy of risk ratings. federalreserve.gov
4. ✅ Flag headroom compression below 20% — any deterioration in covenant headroom to below 20% should trigger an internal watch-list review, management call, and updated forward projection regardless of whether a breach has occurred.
5. ✅ Review add-back sustainability each quarter — confirm that add-backs used in the prior period compliance calculation have either been realized or are on track to be realized; sunset expired items immediately.

Key Takeaways

• 🔶 Maintenance covenants are early-warning systems — not backstops. Their value is not in providing recourse after default; it is in giving lenders intervention rights while the borrower still has operational runway to restructure. A covenant that only triggers at 47% EBITDA decline provides no meaningful early warning.
• 🔶 Headroom above 35% is a structural risk, not conservatism. S&P Global data shows that over a third of private credit borrowers carry leverage covenant headroom above 40% — and that most of this reflects lax documentation rather than strong performance. Lenders should treat excessive headroom as a red flag, not a safety margin.
• 🔶 EBITDA definition discipline is the covenant’s integrity test. An uncapped add-back regime can render a 5.25x leverage covenant functionally equivalent to a 7.5x one. Every covenant negotiation must include a precise, negotiated EBITDA definition with explicit carve-out caps and sunset dates.
• 🔶 The cov-lite migration is a credit cycle risk amplifier. With distressed debt exchanges comprising 64% of all defaults in H1 2025 and leveraged loan default rates projected to peak near 7.9%, the absence of maintenance covenant triggers means lenders in cov-lite structures are reaching distressed outcomes with fewer negotiating levers and less time.
• 🔶 Quarterly monitoring must be active, not administrative. Receiving a compliance certificate is not the same as monitoring a credit. Independent reconstruction of covenant calculations, active tracking of add-back legitimacy, and variance-from-plan analysis are the disciplines that convert a covenant from a paper protection into a functioning risk management instrument.