EBITDA Normalization in Credit Underwriting: How Lenders Dissect Adjusted Earnings Before Committing Capital

The Underwriting Challenge: What Number Are You Actually Lending Against?

A sponsor submits a Confidential Information Memorandum. The cover page reads: “$10.2M Adjusted EBITDA, 5-Year CAGR of 22%.” Three pages later, the footnotes reveal $3.1M in add-backs: acquired cost savings not yet realized, one-time legal settlements that appeared in two of the last three fiscal years, and owner compensation normalized to a market rate that conveniently triples the executive’s stated salary.

This is the central challenge of credit underwriting. The seller’s Adjusted EBITDA is a starting point, not a conclusion. Every dollar of inflated or unjustified adjusted EBITDA has a disproportionate impact on leverage ratios, debt capacity, and ultimately, loss severity in a default. For a lender underwriting at 5.0x Debt/EBITDA, a $500K reduction in lender-accepted EBITDA eliminates $2.5M of supportable debt. For a $50M loan, the arithmetic of bad add-backs can mean the difference between a performing credit and a restructuring.

The market environment adds urgency to this discipline. Average closing leverage among private credit deals increased to 5.1x in 2025, up from 4.8x in 2024, while average permitted debt capacity reached 6.3x, leaving a razor-thin cushion if earnings disappoint post-close. In this environment, rigorous EBITDA normalization is not a formality. It is the first line of credit defense.

The Analytical Framework: Three Layers of Lender-Side Normalization

Institutional lenders do not simply review the seller’s adjustment schedule; they reconstruct it independently. The lender-side normalization process operates across three sequential layers, each with its own analytical standard and documentation threshold.

Layer 1: Definitional Adjustments (EBITDA Base)

The foundation: reported net income plus interest expense, taxes, depreciation, and amortization. This baseline is straightforward but requires verification against audited financials, not management accounts. Confirm that GAAP revenue recognition policies are applied consistently, that no revenue has been pulled forward, and that period costs have not been deferred to improve trailing margins.

Layer 2: Quality-of-Earnings Adjustments (Non-Recurring Items)

These remove genuinely one-time items: a settled litigation charge, restructuring severance from a completed headcount reduction, or a one-time acquisition transaction fee. The critical test is the prior-year recurrence test: if an item labeled “one-time” appears in the general ledger in two or more of the last three fiscal years, it is recurring. Full stop. Lenders applying proper underwriting standards will treat it as a permanent cost.

According to Pepperdine’s 2025 Private Capital Markets Report, adjusted EBITDA is the most-used multiples method in lower middle-market valuation, applied in roughly 76% of cases, which means QoE scrutiny of add-backs is where most deal-level value is won or lost.

Layer 3: Normalization Adjustments (Market-Rate Economics)

This layer corrects items that are recurring but economically mispriced: owner compensation above or below market replacement cost, related-party rent that diverges from arm’s-length rates, family members on payroll without documented operating roles, or management fees paid to the sponsor that will cease post-close. Critically, normalization adjustments work in both directions. If an owner is paying himself below-market compensation, a lender must normalize upward, adding a phantom cost that will exist under new ownership, which reduces EBITDA. A credible normalization schedule includes both add-backs and add-downs.

Working Example, $10M EBITDA Manufacturing Company

Seller presents: Reported EBITDA $7.8M + $2.2M in add-backs = $10.0M Adjusted EBITDA.

Lender QoE analysis:

• $800K “one-time” IT consulting costs: appeared in FY2023, FY2024, FY2025 → Rejected. Recurring operational cost.
• $600K excess owner compensation vs. $180K market replacement GM salary → Accepted only for $420K delta.
• $400K “anticipated” cost savings from vendor renegotiation not yet implemented → Rejected. Post-close action, not in run-rate.
• $400K settled legal claim (no ongoing exposure, single event) → Accepted in full.

Lender-Underwritten EBITDA: $7.8M + $0.82M accepted add-backs = $8.62M. At 5.0x leverage, this is a $43.1M loan, not $50.0M. The EBITDA gap costs the sponsor $6.9M of financing capacity before the first term sheet is signed.

The Analysis: Key Metrics, Coverage Calculations, and Stress Testing

Once lender-underwritten EBITDA is established, the underwriter calculates the credit’s core ratios using that number exclusively, never the seller’s adjusted figure.

Primary Credit Metrics (using lender-normalized EBITDA)

• Total Leverage (Debt / EBITDA): Benchmark threshold for typical direct lending deals is 4.0x–5.5x. At $43.1M debt on $8.62M EBITDA, this credit sits at 5.0x, acceptable but with limited headroom.
• Interest Coverage (EBITDA / Cash Interest): At SOFR + 550bps (approximately 10.3% all-in at current rates), annual cash interest on a $43.1M loan is ~$4.4M. Coverage = 8.62M / 4.4M = 1.96x: below many lenders’ minimum 2.0x–2.5x threshold. This triggers a pricing or structural conversation.
• Fixed Charge Coverage Ratio (FCCR): EBITDA minus capex minus taxes, divided by total debt service. If capex is $1.2M and taxes are $600K: ($8.62M − $1.2M − $0.6M) / $5.6M (P+I) = 1.22x FCCR: marginal. Most institutional lenders require 1.10x–1.25x minimum covenant with 10%–15% headroom to the covenant level.

Stress Scenario: 20% EBITDA Decline

Applying a 20% revenue-driven EBITDA stress (a reasonable test in a macro slowdown scenario, given Moody’s baseline GDP growth forecast of approximately 1.5%, near historical “stall speed”): stressed EBITDA = $6.9M. Stressed interest coverage drops to 1.57x, stressed FCCR falls to 0.91x, a covenant breach in most structures. This stress test identifies whether the credit can absorb a moderate revenue contraction without triggering a default event, and it informs covenant headroom design.

The math reinforces a foundational principle: every dollar of accepted add-back has 5x leverage amplification on debt capacity, and every dollar of rejected add-back should produce a corresponding reduction in committed financing, not a negotiated work-around.

Industry Standards: What Institutional Lenders Look For, and What They Reject

Market Data on EBITDA Quality and Credit Performance

The data paints a clear picture of what happens when normalization discipline breaks down:

• Private Credit Default Rates (2025–2026): Moody’s analysis indicates the private credit default rate in 2025 likely ranged between 1.6% and 4.7%, depending on whether distressed exchanges are included. Distressed restructurings accounted for approximately 65% of all defaults in 2025, meaning most “defaults” were deferred stress, not resolved losses. Fitch, applying a broader definition, reported the U.S. private credit default rate reaching a record 6.0% in April 2026.
• Proskauer Private Credit Default Index (Q2 2026): Across 716 senior secured and unitranche loans representing $195.6 billion in original principal, the default rate registered 2.51% in Q2 2026, a more conservative measure focused on payment defaults in established direct lending portfolios.
• Cliffwater CDLI-S Loss Rate: The CDLI-S index, which isolates senior direct loans, posts an annualized loss rate of just 0.29% from September 2010 through September 2025, compared to 1.49% for high-yield bonds over the same period. This performance advantage is, in large part, a product of first-lien position and rigorous underwriting, including disciplined EBITDA normalization.
• Recovery Rates Are Declining: First-lien senior secured loans have historically recovered roughly 70–80 cents on the dollar on an ultimate recovery basis. However, expected recoveries have drifted down toward 60–70% because modern capital structures carry more first-lien debt and fewer junior cushions, and because distressed exchanges often deliver weaker outcomes than traditional bankruptcies. This declining recovery cushion makes entry-point EBITDA accuracy even more critical.
• Add-Back Inflation Is Structural: EBITDA add-backs averaged 28% of reported EBITDA across covenant-lite leveraged buyout structures, with 92% of companies subsequently missing their projected earnings, a direct consequence of accepting management-case adjustments without independent normalization discipline.

Common Red Flags in EBITDA Presentations

• The “One-Time” That Isn’t: Any add-back appearing in the same cost category across multiple years. Recurring expenses dressed up as non-recurring are the most common rejected category in QoE reviews.
• Run-Rate Revenue for Contracts Not Yet Signed: Pro forma revenue for contracts “in negotiation” or “expected to close” has no place in underwriting EBITDA. Full stop.
• Uncapped Add-Back Schedules: Sophisticated lenders cap permitted add-backs at a percentage of non-adjusted EBITDA. One precedent credit agreement caps add-backs for companies below $50M EBITDA at 40% of non-adjusted EBITDA, a hard structural limit on earnings inflation.
• Missing Add-Downs: A normalization schedule that only adds back, never reduces, is a signal of motivated reasoning, not objective analysis. Market-rate replacement costs for underpaid owner-operators must be modeled explicitly.
• Synergy or Post-Close Savings: Anticipated cost savings dependent on post-closing operational changes are buyer-created benefits, not seller-demonstrated earnings. These belong in the acquirer’s upside model, never in the underwriting EBITDA base.

The Underwriting Checklist: EBITDA Normalization Due Diligence

Use this step-by-step framework before accepting any adjusted EBITDA figure as the basis for leverage or coverage calculations:

Phase 1: Source Document Collection

• ☐ Three years of audited or reviewed financial statements (GAAP basis)
• ☐ Trailing twelve months (TTM/LTM) management accounts with month-by-month breakout
• ☐ General ledger detail for all accounts cited in the add-back schedule
• ☐ Seller-provided QoE report (if available), treat as a starting point, not a conclusion
• ☐ Independent Big-4 or nationally recognized QoE report (preferred for transactions >$20M EBITDA)

Phase 2: Add-Back Challenge Process

• ☐ Apply the prior-year recurrence test to every non-recurring add-back (3-year look-back minimum)
• ☐ Obtain invoice-level documentation for all add-backs exceeding $50K
• ☐ Verify owner compensation vs. published market salary data for comparable roles and geography
• ☐ Confirm related-party lease rates against independent appraisal or comparable market rents
• ☐ Reject all pro forma add-backs for unsigned contracts, unimplemented price increases, or post-close cost saves
• ☐ Reconcile adjusted EBITDA to actual cash collected (bank statement reconciliation)

Phase 3: Lender-Underwritten EBITDA Build

• ☐ Construct an independent earnings bridge from reported net income to lender-underwritten EBITDA
• ☐ Apply add-downs for any underpaid owner compensation or below-market related-party economics
• ☐ Document every accepted and rejected add-back with rationale in the credit memo
• ☐ Cap total accepted add-backs as a percentage of reported EBITDA per internal policy

Phase 4: Ratio and Stress Testing

• ☐ Calculate Debt/EBITDA, Interest Coverage, and FCCR using lender-underwritten EBITDA exclusively
• ☐ Run a base, downside (−15%), and stress (−25%) scenario on EBITDA to test covenant headroom
• ☐ Confirm that minimum FCCR covenant (typically 1.10x–1.25x) provides adequate early-warning buffer
• ☐ Validate that stressed leverage stays within lender’s maximum leverage covenant with margin

Key Takeaways: Five Risk Disciplines for EBITDA Underwriting

1. Lender-Underwritten EBITDA is the Only Number That Matters. Seller Adjusted EBITDA initiates the diligence process; it never concludes it. Every leverage ratio, coverage test, and debt capacity calculation must use the independently reconstructed figure.
2. The Prior-Year Recurrence Test Is Non-Negotiable. If an add-back appears in the same cost category in two or more of the trailing three fiscal years, it is a recurring cost, regardless of how it is labeled. Accepting a recurring cost as non-recurring is the most common source of underwriting error in middle-market lending.
3. Add-Downs Are as Important as Add-Backs. A normalization schedule that only moves in one direction is not independent analysis; it is advocacy. Market-rate replacements for underpaid management, below-market leases, and missing overhead must reduce EBITDA wherever they apply.
4. Declining Recovery Rates Raise the Cost of EBITDA Errors. With first-lien expected ultimate recoveries drifting toward 68% versus a historical average in the high 70s, the loss severity from an over-levered, over-adjusted credit is meaningfully higher than in prior credit cycles. Entry-point underwriting discipline is the primary lever a lender controls.
5. Maintenance Covenants Are the Alarm System; EBITDA Normalization Is the Architecture. Lenders who negotiate maintenance covenants recover, on average, approximately 10 percentage points more than lenders who discover problems only at payment default. But covenants calibrated to an inflated EBITDA base provide false protection, the covenant trigger is set too high to provide early warning. Accurate normalization at origination is what makes every downstream risk management tool work as intended.